Journal of Hydrology (2008) 349, 318– 332
available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/jhydrol
Reconstruction of sediment flux from the Changjiang
(Yangtze River) to the sea since the 1860s
Houjie Wang
J. Paul Liu d
a,b,*
, Zuosheng Yang
a,b
, Yan Wang
a,b
, Yoshiki Saito c,
a
College of Marine Geosciences, Ocean University of China, Qingdao 266100, China
Key Laboratory of Submarine Geosciences and Prospecting Technology, Ministry of Education, 238 Songling Rd.,
Qingdao 266100, China
c
Geological Survey of Japan, AIST, Tsukuba 305-8567, Japan
d
Department of Marine, Earth and Atmospheric Sciences, North Carolina State University, Raleigh, NC 27695, USA
b
Received 1 June 2007; received in revised form 5 November 2007; accepted 8 November 2007
KEYWORDS
Changjiang (Yangtze
River);
Sediment flux;
Rating curves;
Human activities
The Changjiang (Yangtze River) has been effectively gauged since the 1950s
and demonstrates the transformation of a river system due to intensified human activities
in its drainage basin over the past 50 yr. However, the 50-yr measurements of water and
sediment are inadequate to show the long-term trend of sediment flux from the river to
the sea or to capture the transition from natural to human dominance over the sediment
flux. In this study we used the existing water discharge and sediment load records (1950s–
2005) at the Hankou gauging station, together with water discharge recorded since 1865 at
the same station, to reconstruct the changes of sediment flux to the sea since the 1860s.
We established rating curves between stream discharge and suspended sediment concentration from the recent 50-yr data sets, which show that human disturbances have had a
substantial impact on rating parameters. The commissioning of dams and undertaking of
soil-conservation works have decreased sediment supply, leading to a decrease in the rating coefficient a of the rating curve equation Cs = aQb. The decreases in suspended sediment concentration have increased the erosive power of the river, and hence increased
the rating exponent b. In particular, the commissioning of the Three Gorges Reservoir
in 2003 resulted in a further increase of b, and channel scour in the middle and lower
reaches has increased sediment flux to the sea to a level higher than sediment supply from
the upper reaches. Our results suggest that the rating curves derived from 1954 to 1968
data are appropriate for estimating sediment loads for the period from 1865 to 1953, since
both were periods of minimal human disturbance. This approach provides a time series of
Summary
* Corresponding author. Address: Key Laboratory of Submarine Geosciences and Prospecting Technology, Ministry of Education, 238
Songling Rd., Qingdao 266100, China. Tel.: +86 532 66782950; fax: +86 532 66782062.
E-mail address: [email protected] (H. Wang).
0022-1694/$ - see front matter ª 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.jhydrol.2007.11.005
Reconstruction of sediment flux from the Changjiang (Yangtze River) to the sea since the 1860s
319
sediment loads from 1865 to 2005 at Hankou gauging station, which yields a time series of
sediment flux from the Changjiang to the sea over the past 140 yr. The estimated mean
annual sediment flux to the sea between 1865 and 1968 was 488 Mt/yr, a comparable
result to the previously published estimate from Milliman and Syvitski [Milliman, J.D.,
Syvitski, J.P.M., 1992. Geomorphic/tectonic control of sediment discharge to the ocean:
the importance of small mountainous rivers. Journal of Geology 100, 525–544] and to that
from an equation proposed by Syvitski and Morehead [Syvitski, J.P.M, Morehead, M.D.,
1999. Estimating river-sediment discharge to the ocean: application to the Eel margin,
northern California. Marine Geology 154, 13–28]. The long-term variation of annual sediment flux from the Changjiang to the sea shows a transition from a river system mostly
dominated by nature (the monsoon-dominated period, 1865–1950s) to one strongly
affected by human activities (the human-impacted period, 1950s–present).
ª 2007 Elsevier B.V. All rights reserved.
Introduction
The Changjiang (Yangtze River) is the largest river in China
and the main source of terrigenous sediment delivered to
the continental shelf of the East China Sea. Before human
activities dominated the Changjiang, it carried sediment
loads of 480 million tons per year (Mt/yr) to the sea
(Milliman and Meade, 1983; Milliman and Syvitski, 1992; Wang
et al., 1997; Yang et al., 2006). The Changjiang, together
with the Huanghe (Yellow River), is a major link between
the world’s largest continent (Asia) and its largest ocean
(the Pacific), and plays a critical role in delivering both particulate and dissolved terrestrial material from the Chinese
mainland to the western Pacific Ocean (e.g., DeMaster
et al., 1985; Milliman et al., 1985; Liu et al., 2004, 2006;
Wang et al., 2007). Estimates from numerous boreholes
and seismic profiles suggest that during the middle Holocene
the Changjiang distributed 1,700,000 Mt of sediment to its
huge delta plain and proximal subaqueous delta on the inner
shelf of the East China Sea—an average of approximately
240 Mt/yr (Saito et al., 2001; Liu et al., 2006, 2007). Over
the past several decades there has been growing interest
in the Changjiang sediment flux to the sea (e.g., Chen
et al., 2001; Shen, 2001; Yang et al., 2002, 2006). The reasons for this are numerous and diverse, and include interest
in issues such as delta morphology, estuarine processes, and
mud deposits on the continental shelf of the East China Sea,
and the implications of sediment flux for engineering projects. The high water and sediment discharges of the Changjiang also account for a considerable amount of nutrients,
particularly carbon, from the river basin to the coastal
ocean (Martin and Meybeck, 1979; Ludwig et al., 1996;
Smith et al., 2001). The materials delivered from this river
play an important role in biogeochemical cycles in the East
China Sea. Consequently, effective studies of the Changjiang sediment flux to the sea have been beyond the scope
of individual disciplines such as marine sedimentology,
hydrology, or marine geochemistry.
Many recent publications have documented the variations of sediment flux from the Changjiang to the East China
Sea over the latter half of the twentieth century on the
basis of continuous records at Datong gauging station since
1950 (e.g., Chen et al., 2001; Yang et al., 2002, 2006; Xu
et al., 2006). Together, these studies present a detailed
view of the rapid decrease of Changjiang sediment flux as
a result of extensive human activities (e.g., dam construction, soil-conservation works), and provide excellent case
studies on the transformation of world river systems by human activities over the past 50 yr (e.g., Syvitski, 2003; Meybeck and Vörösmarty, 2005; Nilsson et al., 2005; Syvitski
et al., 2005). However, longer time series of sediment flux
data, such as centennial-scale data sets, is unavailable but
would be beneficial to geoscientists, since they would (1)
illustrate the long-term variations and trends of Changjiang
sediment flux to the East China Sea, and show the transition
to the presently human-dominated period (1950–present);
and (2) provide critically important data for the study of
environmental changes since the Industrial Revolution. Furthermore, because the delivery of pollutants to the sea is
mostly associated with sediment transport by rivers (Huh
and Chen, 1999), a longer-term data set would allow more
definitive assessment of basin-wide pollutants and contaminants released by human activities and delivered from rivers to the sea. Fortunately, there is a long time series of
monthly water discharge records (since 1865) at Hankou
gauging station in the middle reaches of the Changjiang
(Fig. 1); however, the sediment load was not monitored
there until the 1950s. Despite the absence of sediment load
records, water discharge data from Hankou station are valuable and essential, not only for the estimation of long-term
water discharge to the sea (e.g., Yang et al., 2005) but also
for the reconstruction of time series of annual sediment flux
from the Changjiang to the sea over the past century and a
half.
Due to the absence of continuous measurements of suspended sediment concentration for rivers, hydrologists have
made considerable efforts on the rating curves that can
estimate or predict the suspended sediment concentration
from the stream flow records (e.g., Campbell and Bauder,
1940; Gregory and Walling, 1973; Walling, 1977; Syvitski
et al., 1987; Syvitski and Alcott, 1995; Phillips et al.,
1999; Asselman, 2000). The rating curve generally takes
the form of a power regression that relates suspended sediment concentration (or load) to stream discharge:
Cs ¼ aQ b
ð1Þ
or, alternatively, a linearized equation based on natural logarithm transformation:
ln Cs ¼ ln a þ b ln Q
ð2Þ
320
H. Wang et al.
o
115 E
o
110 E
Bohai Sea
Datong
Jia
414 Mt/yr
3
an
v.
Ri
R
iv
.
Tai Lake
Danjiangkou
Reservoir
Hankou
Riv.
a
a Se
Chin
g
Minjian
Yichang
East
g
n
jia
ng
Jingjiang
Fluvial plain
Datong
ng
ya ke
Po La
ing
ngt
Do Lake
Lake
90 oE
o
29 N
iv.
gR
an
uji
W
Tributary
24 oN
o
34 N
jia
Riv.
Gauging station
iv.
jiang R
Reservoir
Yalong
N
Jinsha
29 o
Upper reaches
Middle reaches
Lower reaches
Yellow Sea
H
g
lin
Latitude
714 km /yr
34 oN
Hankou
470 Mt/yr
3
436 km /yr
Yichang
3
105oE
903 km /yr
100oE
o
125 E o
39 N
o
120 E
Huk
ou
95 oE
384 Mt/yr
90 oE
39 oN
o
Three Gorges Dam
95 oE
100oE
o
o
110 E
105 E
o
115 E
o
120 E
24 N
o
125 E
Longitude
Figure 1 Map of the Changjiang drainage basin, major tributaries, major gauging stations, and reservoirs, and mean annual water
discharge (km3/yr) and sediment load (Mt/yr) at major stations along the main stream (averages from the 1950s to 2005).
where Cs is the suspended sediment concentration (kg/m3),
Q is the stream discharge (m3/s), and a and b are the rating
coefficient and rating exponent, respectively. Since the suspended sediment of a river is essentially a non-capacity load
(Walling and Moorehead, 1989), the rating curves usually
demonstrate considerable scatter and consequently underestimate or overestimate the sediment concentrations
(Horowitz, 2003). The errors are primarily ascribed to the
fact that the suspended sediment delivery is impacted not
only by the hydraulic properties of stream flow, but also
by the sediment supply from the catchment and intensity
and spatial distribution of rainfall. Despite these potential
deficits, the rating curves, however, present a simple way
that allows empirical conversion of stream discharge hydrographs into sediment load estimates, provided that the time
series of stream discharge are known (Syvitski and Alcott,
1995). The rating curves have been widely used for a variety
of scientific and engineering purposes, such as for prediction of the life span of a dam on a river, or for estimation
of the sediment load of an ungauged river (Syvitski et al.,
2000). However, it is important to note that the rating
parameters can vary greatly both geographically and over
time.
The extensive records since the 1950s from major gauging stations on the Changjiang provide a basis for detailed
analysis of stream flow and suspended sediment load. For
example, rating curves established from existing records
can be used to quantitatively estimate sediment deficits
during catastrophic flooding events (Xu et al., 2005). Yang
et al. (2007) used data from the 1950s to the 1980s to calculate rating parameters and discussed their implications
for hydrological processes and their relationship to the East
Asian monsoon. However, their time-limited data set
(1950s–1980s) made it difficult to reach a convincing con-
clusion on the relationships between channel morphology,
hydrological processes, and the East Asian monsoon. Therefore, in this study we reconstructed the annual sediment
load delivered from the Changjiang to the sea since 1865
by using the historical monthly stream-flow data from the
Hankou station. Our first step in this approach was to
approximate the rating parameters for the long-term estimation by using the records of stream discharges and suspended sediment concentrations at Hankou station from
1954 to 2005; we then used the derived rating parameters
to estimate the annual sediment load during the period from
1865 to 1953; finally, we combined these two data sets to
reconstruct the time series of sediment load for the period
from 1865 to 2005, thus determining an estimated time series of sediment flux from the Changjiang to the sea over the
past 140 yr. We believe that this estimate will provide an
important centennial-scale reference for investigations into
the evolution of the Changjiang estuary and deltaic depositional system, and also for studies of sedimentary processes
on the inner shelf of the East China Sea.
Regional setting
The Changjiang rises on the Qinghai–Tibetan Plateau at an
elevation of 6600 m and flows eastward into the eastern China for a total length of 6300 km and a drainage area of
1.94 · 106 km2 before debouching into the East China Sea
(Liu et al., 2007). It is divided into three segments: (1)
the upper reaches extend 4500 km from the river source
to Yichang, flowing through a mountainous region where a
considerable number of tributaries join the main stream;
(2) the middle reaches flow 950 km through the flat Jingjiang fluvial plain with very gentle slope from Yichang to
Reconstruction of sediment flux from the Changjiang (Yangtze River) to the sea since the 1860s
b
800
Sediment
1200
600
800
400
Water
200
400
Yichang
Hankou
120
60
Water
40
Sediment
80
40
20
Datong
0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
c
100
80
80
60
40
40
20
0
0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
160
Datong station
80
120
60
80
40
40
20
Monthly sediment load (Mt)
Hankou station
Monthly water discharge (km3)
d
120
Monthly sediment load (Mt)
Monthly water discharge (km 3)
Yichang station
0
0
0
160
80
Monthly water discharge (km3)
1600
Annual sediment load (Mt/yr)
Annual water discharge (km3/yr)
a
mean annual water discharge at Datong is approximately
twice that coming from the upper reaches (records at
Yichang station) (Fig. 2a). The sediment loads at the three
major gauging stations along the main stream indicate that
the Changjiang sediment is predominately derived from the
upper reaches, where severe soil erosion and frequent landslides have a significant effect, but the water from the
upper reaches accounts for only 50% of that discharging
into the East China Sea.
The Changjiang river basin experiences a subtropical
monsoon climate with monsoons initiated in the southeastern Pacific Ocean and Indian Ocean, which is closely related
to the spatial and temporal variability of precipitation in the
river basin. The summer monsoon normally starts to influence the river basin in April and retreats in October,
accounting for more than 50% of annual total rainfall. The
rainy season, when water discharge is high, is typically from
July to September, whereas the dry season, when water discharge is low, is from December to March (Fig. 2b–d). The
annual precipitation increases eastwards geographically
from 300 mm/yr in the upper reaches to more than
1000 mm/yr in the middle and lower reaches. In the upper
reaches, the total water discharge from July to September
accounts for 50% of annual discharge, whereas discharge
in the middle and lower reaches during the same period contributes only 40% to annual discharge. This indicates that
water discharge from the upper reaches is dominantly influenced by monsoonal conditions, whereas water discharge
from the middle and lower reaches is not. The seasonal distributions of sediment load at the three gauging stations
show a strong association with the patterns of water
discharge, as characterized by the coincidence of high
Monthly sediment load (Mt)
Hukou; and (3) the lower reaches flow from Hukou to the
river mouth through a tidally influenced segment downstream from Datong gauging station (Fig. 1).
The annual water discharge from the upper reaches recorded at Yichang gauging station is approximately
463 km3/yr, and the annual sediment load amounts to
470 Mt/yr, which constitutes the main source of sediment
delivered to the middle and lower reaches (Fig. 2a). Several
major tributaries, including the Jinshajiang, Jialingjiang,
Minjiang, and Wujiang contribute 86% of the sediment
load at Yichang station (Fig. 1, Yang et al., 2006). In the
middle reaches, the river meanders across the almost flat
Jingjiang fluvial plain, forming several large lowland lakes,
including Dongting Lake and Poyang Lake. In addition, a major tributary, the Hanjiang, joins the Changjiang at Hankou
station (Fig. 1). The Hanjiang annually contributed 73 Mt/
yr of sediment load to the Changjiang before construction of
the Danjiangkou Reservoir in 1968 (Yang et al., 2006;
Fig. 1). The middle reaches, particularly over the Jingjiang
fluvial plain, act as an important flood-diversion area,
accounting for at least 100 Mt of sediment loss due to overbank deposition (Yang et al., 2006) during years of catastrophic flooding (e.g., 1954 and 1998). As a result, the
annual sediment load at Hankou averages 383 Mt/yr,
accounting for 80% of the sediment input from the upper
reaches, whereas the mean annual water discharge at Hankou increases to 711 km3/yr (Fig. 2a). A large proportion of
the lower reaches between Datong station and the river
mouth (600 km in length) is influenced by tides of the East
China Sea. The mean annual sediment load from the Changjiang to the sea, as recorded at Datong station, is 413 Mt/
yr, slightly higher than that at Hankou station, whereas the
321
0
0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Figure 2 Water discharge and sediment load at three major gauging stations (locations shown in Fig. 1). (a) Whisker-box plots of
annual water discharge and sediment load at Yichang, Hankou, and Datong gauging stations. Monthly water discharge and sediment
load at (b) Yichang station (1950–2005), (c) Hankou station (1954–2005), and (d) Datong station (1951–2005).
322
sediment load and high water discharge during the rainy
season (Fig. 2b–d).
Since 1950, human activities have led to a notable reduction of sediment flux from the Changjiang to the sea,
whereas water discharge to the sea has remained almost unchanged (Yang et al., 2002, 2006). The commissioning of
several major reservoirs and soil-conservation works in the
upper reaches has resulted in distinct stepwise decreases
in sediment load delivered to the sea (Yang et al., 2006).
For example, construction of the Danjiangkou Reservoir in
1968 produced the first downward step at Datong station,
where the mean annual sediment load from 1969 to 1985 accounted for only 90% of the reference level (the average
from 1950 to 1968); the second downward step in the mid1980s is ascribed to the construction of numerous hydropower stations and soil-conservation works in the upper
reaches, which led to a further decrease of 20% from
the reference level during the period from 1986 to 2002;
and in 2003, the Three Gorges Dam began trapping sediment
and storing water from the upper reaches, resulting in the
third downward step, with the mean annual sediment load
to the sea during the period from 2003 to 2005 decreasing
to 189 Mt/yr, accounting for only 39% of the reference level.
The stepwise decreases of sediment load during a period
when water discharge remained unchanged illustrates a significant loss of sediment from the river and reflects the
extensive human impacts on this large river system. Therefore, it is noteworthy that the drastic decrease of sediment
load will have a definite impact on the use of the rating
curve method to examine the relationship between water
discharge and suspended sediment concentration.
Data and methodology
The monthly records of water discharge at Hankou gauging
station from 1865 to 1953 and those of both monthly and annual water discharge and sediment load from 1954 to 2005
were supplied by the Changjiang Water Conservancy Commission (CWCC). Records of the monthly and annual water
discharge and sediment load from Datong station came
mostly from the CWCC, but partly from Bulletins of Chinese
River Sediment IRTCES (2001–2005) and a recent publication of the Ministry of Water Resources of China (E.J.,
2006). The monthly suspended sediment concentrations
(Cs) are of discharge-weighted values that are derived as
the monthly suspended sediment loads divided by the
monthly water discharges. We chose to use the data from
Hankou station for our analysis for three reasons: (1) Hankou station, located at the mouth of the Hanjiang, records
the water discharge and sediment load delivered from the
upper and middle reaches of the Changjiang (Fig. 1); (2)
Hankou station provides the longest records of water discharge (1865–2005) as well as recent records of both water
discharge and sediment load (1954–2005); and (3) the water
discharge and sediment load records at Hankou station
approximate those at Datong station, because there is little
additional input of water or sediment in the lower reaches
(Fig. 1). River sediment load was measured according to
the Chinese national standard criteria. Water samples were
taken from a consistent gauging section across the full
water column, and all flow discharge records were taken
H. Wang et al.
at the same time of day; the daily, monthly, and annual suspended sediment load and water discharge were derived
from these data (Yang et al., 2006).
We used the uninterrupted records of monthly water discharge and suspended sediment concentration from Hankou
gauging station from the 1950s to the present to determine
the rating parameters for reconstruction of sediment load
from 1865 to 1953. Because of the close similarity of sediment loads recorded at Hankou and Datong gauging stations, our analysis of data from Hankou station provides,
for the first time, a data set that represents the sediment
load delivered by the Changjiang to the sea during the past
century and a half.
Results
Rating parameters and their implications
Given stream flows (Q) are known, the suspended sediment
loads (Qs) can be approximated from the rating curve
equation:
Q s ¼ Cs Q ¼ aQ bþ1
ð3Þ
where Qs is the estimated suspended sediment load (kg/s)
that can be converted to monthly or annual load at different
time scales. Clearly, the determination of rating parameters
(a and b) is critical to the estimation of sediment loads. The
rating parameters are an index of erosion severity in the river channel (Morgan, 1995; Asselman, 2000). The coefficient
a primarily reflects the grain size of sediment yielded from
the basin and represents the erodibility of soil, or an index
of sediment supply from the source region. High values of a
generally indicate high sediment supply, such as in areas
characterized by intensively weathered materials that can
easily be eroded and transported. Although the coefficient
a largely varies for different sizes of basins, the variations
of a with time for a given basin are expected to have implications of human disturbances on the hydrological process.
The exponent b represents the erosive power of the river
and the influence on the sediment supply from the entire
catchment surface; large values of b indicate an increase
in sediment transport capacity of river stream (Asselman,
2000). And the values of b can also be affected by the
grain-size distribution of the material available for transport in the river basin (Walling and Moorehead, 1989). A
number of factors affect the suspended sediment transport
in a river, including long-term climate, short-term weather
events and discharge type (Syvitski et al., 2000), and these
factors have characteristic impacts on the rating parameters. For example, rating coefficient a for rivers in areas
of arid climate is high (100–80,000) and rating exponent b
is low (0.2–0.7). Conversely, for rivers in areas of temperate and humid climates,a ranges from 0.004 to 40 and b
from 1.4 to 2.5 (Reid and Frostick, 1987; Syvitski et al.,
2000). The values of ln(a) and b are generally inversely correlated and reflect the soil erodibility and stream erosivity,
respectively (Thomas, 1988; Asselman, 2000). In addition to
the above natural factors, human activities within the river
basin (e.g., dam construction, soil-conservation works) create significant disturbances to the relationship between
stream discharge and suspended sediment load. The marked
Reconstruction of sediment flux from the Changjiang (Yangtze River) to the sea since the 1860s
68
-19
54
19
a
0.5
85
-19
69
19
1954-1968
1969-1985
0.0
1986-2002
2003-2005
1954.7
-1.5
s
ea
ds
on
o
Flo
-2.0
1954.9
1954.8
05
-20
03
20
Ln (Cs)
-1.0
02
-20
86
19
-0.5
-2.5
Dr
n
so
ea
ys
-3.0
-3.5
8.0
8.5
9.0
9.5
10.0
10.5
11.0
11.5
12.0
Ln (Q)
b 1.6
1955-1968
1969-1985
1986-2002
2003-2005
1.4
1.0
0.8
Increase of erosive power
(decrease in SSC)
1.2
b
Decrease of sediment supply
(impacts from dams and soil-conservation works)
0.6
-16
-14
-12
-10
-8
-6
Ln (a)
Figure 3 (a) Rating curves derived from water discharge and
suspended sediment concentration (SSC) data from Hankou
station between 1954 and 2005. Both seasonal transitions along
the regression lines and shifts of data perpendicular to the
regression lines as a result of human disturbances are clearly
demonstrated. (b) Rating parameters obtained from intraannual data over different time periods.
reduction of sediment load delivered by the Changjiang to
the sea, with almost unchanged levels of water discharge,
is a good example of such an effect that human activities
can have on a river system (e.g., Yang et al., 2002, 2006).
323
At Hankou station, the stream flow and suspended sediment concentration data from 1954 to 2005 (data unavailable in 2000) exhibit an approximately linear natural log–
transformed relationship with a coefficient of determination r2 = 0.61 (Fig. 3a). The corresponding rating parameters
are ln(a) = 10.14 and b = 0.92, which are similar to those
of Yang et al. (2007). However, the relationship shows a distinct variation with time (Fig. 3a). For example, the data for
1954–1968 lie on the upper part of the plot, except for
three anomalous data points in July–September of 1954,
when there was catastrophic flooding in the middle and lower Changjiang (Fig. 3a). The mean stream flow during this
period was 22,143 m3/s with a mean suspended sediment
concentration of 0.52 kg/m3. The corresponding rating
parameters are ln(a) = 9.6 and b = 0.89, respectively
(Table 1). For the period 1969–1985, the data points move
downward in response to the construction and commissioning of the Danjiangkou Reservoir on the Hanjiang tributary
(Fig. 1, Yang et al., 2006). The suspended sediment concentration decreased to 88% of the previous level, and the rating coefficient decreased to ln(a) = 10.69, whereas the
rating exponent b increased slightly (Table 1). After 1985,
the commissioning of numerous hydropower stations and
soil-conservation works in the upper reaches caused a significant decrease in sediment load of the Changjiang (Yang
et al., 2006), and the data points shift distinctly downward
to the middle part of the scatter plot (Fig. 3a). From 1986 to
2002, the suspended sediment concentration and rating
coefficient decreased continuously while the rating exponent increased gradually (Table 1). After commencement
of filling of the Three Gorges Dam in June 2003, the mean
suspended sediment concentration over the next 3 yr decreased to only 33% of the reference level, leading to a
low rating coefficient of ln(a) = 12.45 and high rating
exponent of b = 1.05 (Table 1). Consequently, the data
points for the period 2003–2005 move further downward
to the lowest part of the plot (Fig. 3a). All of the above
changes in recorded suspended sediment concentrations at
Hankou station, as well as the changes in rating parameters,
are the effect of human disturbances on the relationship between stream flow and suspended sediment concentration.
The results from Yang et al. (2007) also illustrate such
trends; however, no reasonable explanation was presented.
The rating parameters obtained from annual rating
curves also illustrate high variability at different stages during the total period from 1955 to 2005 (Fig. 3b). Data from
the catastrophic flood year of 1954 was excluded. During
the period 1955–1968, the values of ln(a) ranged from
Table 1 Mean stream discharge (Q), mean suspended sediment concentration (Cs), rating parameters (ln(a) and b), and
coefficient of determination (r2) for different periods at Hankou gauging station
(m3/s)
s (kg/m3)
Periods
Q
C
ln(a)
b
r2
1955–1968
1969–1985
1986–2002a
2003–2005
1954–2005
a
22,143
22,139
22,563
22,765
22,503
(100%)
(99.9%)
(102%)
(102%)
(102%)
The data sets of 2000 are unavailable.
0.52
0.46
0.35
0.17
0.42
(100%)
(88%)
(67%)
(33%)
(81%)
9.6
10.69
11.27
12.45
10.14
0.89
0.98
1.01
1.05
0.92
0.78
0.80
0.76
0.78
0.61
324
H. Wang et al.
7.6 to 12.3, and the values of b varied from 0.67 to 1.18.
During the period 1969–1985, ln(a) ranged from 8.0 to
13.8, suggesting a decreasing trend from past times, while
the values of b increased slightly within the range 0.72–1.3.
This pattern of variation of rating parameters was maintained during the period 1986–2002, before filling of the
Three Gorges Dam commenced, and the values of ln(a) decreased significantly to be within the range from 9.2 to
15.7, while b increased correspondingly to 0.8–1.5. During the period 2003–2005, this trend was no longer clearly
evident, possibly because there was insufficient data to capture the long-term variation (Fig. 3b). The decreases in the
values of rating coefficient a imply that the sediment available from the source regions (e.g., the upper reaches and
the sub-basin of Hanjiang tributary) was progressively reduced as a result of human disturbances (e.g., the Danjiangkou Reservoir, numerous dams in the upper reaches,
and the Three Gorges Dam) (see Yang et al., 2006). A decrease in sediment load during a period when water discharge is constant clearly produces a decrease in
suspended sediment concentration, which increases the
erosive power of the river, and therefore increases the value of b (Table 1). The variations of rating parameters we
observed suggest that the influence of human activities on
hydrological processes is substantial (Fig. 3) and must be
considered when sediment loads are estimated from rating
curves.
a
200
We estimated monthly sediment loads at Hankou station
from 1954 to 2005 (Fig. 4a) from the annual rating parameters. These estimates agree well with the measured seasonal and interannual variations of monthly sediment load
over the past 56 yr, apart from some underestimation during
flood seasons. Thus, we were able to reproduce the longterm decreasing trend of sediment loads at Hankou station
as a result of human activities within the drainage basin
from 1954 to 2005 (Fig. 4a).
We assumed that the rating curves derived from the
1954–1968 data would be the most appropriate curves for
estimating the unknown sediment loads from 1865 to
1953, since the effect of human disturbances on rating
curves during 1954–1968 was much less than that of subsequent years. However, to select the best-fit annual rating
parameters for the period since 1865 was proved to be difficult, because the annual rating parameters from 1954 to
2005, showed distinct annual variability (Fig. 3b). In order
to rationally determine rating parameters for estimating
sediment loads from 1865 to 1953, we established rating
curves based on consideration of differences in stream flow
during the period from 1954 to 1968, when the effects of human disturbance were nearly negligible. We grouped the
data in three categories: Q < 6000 m3/s (3 pairs of data),
6000 m3/s 6 Q 6 44,000 m3/s (164 pairs of data), and
Q > 44,000 m3/s (13 pairs of data). We further divided the
second category into 30 subcategories, 23 of which were
Measured
Monthly sediment load (Mt)
Estimated
160
120
80
40
NA
0
Monthly sediment load (Mt)
b
1955
1960
1965
1970
1975
1980
1985
1990
1995
2000
2005
200
160
120
80
40
0
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
Year
Figure 4 (a) Estimated monthly sediment load at Hankou station compared with observed sediment load (1954–2005, monthly
data unavailable in 2000). The estimates were derived from the annual rating curves shown in Fig. 3b; (b) Measured monthly
sediment load at Hankou station (1954–1968) compared with estimates from rating curves based on statistical analysis of different
levels of stream discharges.
Reconstruction of sediment flux from the Changjiang (Yangtze River) to the sea since the 1860s
Estimated monthly sediment load, QsE (Mt)
200
QsE=0.80QsM+5.7
r2=0.83, n=180
RMSE=14.8
l
ve
160
ce
i
nf
%
co
le
n
de
95
120
80
l
ve
40
e
nc
le
e
fid
n
%
co
95
0
40
80
120
160
200
Measured monthly sediment load, QsM (Mt)
Figure 5 Estimated monthly sediment load at Hankou station
(1954–1968) compared with recorded data. The comparison
indicates that the estimates explain 83% of the data variance
with 95% confidence.
7
Chance of occurrence (%)
between 6000 and 30,000 m3/s at intervals of
DQ = 1000 m3/s, and 7 between 30,000 and 44,000 m3/s at
intervals of DQ = 2000 m3/s. We calculated rating curves
for each of these categories and subcategories of stream
discharge, and then used the least-squares method to determine the rating parameters. Comparison of the measured
sediment load from 1954 to 1968 and the estimates from
these rating curves showed that this approach is effective
(Fig. 4b). There are a few large discrepancies between the
measured and estimated loads during the catastrophic
floods of 1954 owing to a large flood-derived sediment deficit in the middle reaches (Xu et al., 2005; Yang et al.,
2006). The mean differences between the measured and
estimated sediment loads from 1955 to 1968 are 20% for
flood seasons (July–September) and 16% for non-flood
seasons. The estimates from the rating curves explain
approximately 83% of the data variance within the 95% confidence limits, with a residual mean squared error of 14.8
(Fig. 5). The estimates agree with more than 80% of the observed sediment loads during the period 1954–1968, despite
some scatter around the regression line. It is worthwhile to
note that clear relationships are hard to obtain for some
categories of limited data sets, which may be an additional
cause for the differences between the measured and estimated sediment loads.
Since there are only 180 pairs of monthly data of stream
discharge and suspended sediment concentration (1954–
1968), it is quite difficult to definitely conclude that these
rating parameters derived from limited data are suitable
for estimating the unknown suspended sediment concentration over the period of 1865–1953. However, the monthly
stream discharges for both periods of 1865–1953 and
1954–1968 demonstrate very similar mode (Fig. 6), suggesting that the stream discharge varying from 6000 m3/s to
40,000 m3/s covers approximately 88% and 92% of the data
sets of 1865–1953 and 1954–1968, respectively. Therefore,
the rating parameters derived from the categorizing levels
of stream discharge appear to be valid to estimate the unknown suspended sediment concentrations at Hankou station for the period 1865–1953.
325
1865-1953
1954-1968
6
5
4
3
2
1
0
0
10000
20000
30000
40000
50000
60000
70000
Q (m3/s)
Figure 6 Occurrences of monthly stream discharges for the
periods of 1865–1953 and 1954–1968, illustrating a very similar
mode.
Time series of sediment load at Hankou station
since 1865
Using the rating curves derived from the categorized stream
discharges in the period from 1954 to 1968, together with
the monthly records of stream discharge at Hankou station
(1865–1953), we estimated the monthly sediment loads at
Hankou station from 1865 to 1953 (Fig. 7a). These estimates
and the records from 1954 to 2005 illustrate the long-term
variations of monthly sediment load at Hankou station at
centennial scale (Fig. 7a). They suggest that before construction of the Danjiangkou Reservoir on the Hanjiang tributary in 1968 (Fig. 1), the monthly sediment loads at Hankou
station were relatively stable for a period of more than
100 yr, when there was little human disturbance on the river
hydrological processes. Since 1968, the sediment loads at
Hankou station have been strongly affected by human activities. The human influences are indicated by distinct stepwise decreases in sediment load that correspond to major
events with hydrological implications; for example, the
commissioning of the Danjiangkou Reservoir in 1968, numerous dams and soil-conservation works in the upper reaches
since the mid-1980s, and the recent (2003) commencement
of operations at the Three Gorges Dam (Fig. 7a; Yang et al.,
2006).
Comparisons of the annual time series of sediment load
with annual water discharge at Hankou station from 1865
to 2005 (Fig. 7b) shows that they fluctuated synchronously
during the 100 yr preceding 1965 and have diverged only
since the early 1990s, when the effects of human activities
became more influential (Fig. 7b). We estimated the 100-yr
mean annual sediment load from 1865 to 1968 to be
455 Mt/yr. After the divergence of the trends of annual
sediment load and water discharge, the mean annual sediment load decreased to 286 Mt/yr, which is only 63% of
the level before the effects of human disturbance took hold
(Fig. 7b). These variations of sediment load clearly reflect
the transition from a river system mostly dominated by nature to one strongly affected by human activities.
Time series of sediment load delivered to the sea
since 1865
Datong gauging station, the last station on the mainstream
of Changjiang, is located at the upstream limit of the tidally
326
H. Wang et al.
Monthly sediment load (Mt)
a
200
1865-1953/Estimated
1954-2005/Measured
160
120
80
40
0
Annual sediment load (Mt)
b
1875
1885
1895
1905
1915
1925
1935
1945
1955
1965
1975
1985
1995
2005
2015
1200
700
Annual water discharge
600
Annual sediment load
1000
500
800
400
600
300
400
200
200
100
Annual water discharge (km3)
1865
0
0
1865
1875
1885
1895
1905
1915
1925
1935
1945
1955
1965
1975
1985
1995
2005
2015
Time (year)
Figure 7 (a) Time series of monthly sediment load at Hankou station from 1865 to 2005, including both estimates from rating
curves (1865–1953) and recorded data (1954–2005), and (b) time series of annual water discharge and sediment load at Hankou
station from 1865 to 2005.
influenced river segment (Fig. 1). Water discharge and sediment load recorded at Datong station are representative of
the discharge from the Changjiang to the sea, even though
the station is approximately 600 km away from the river
mouth. Downstream from Hankou station, there is a little
sediment input from channel erosion; therefore, the sediment load at Datong is only slightly higher than that at Hankou station. For example, the mean annual sediment load at
Hankou (1950–2005) was 391 Mt/yr, while that at Datong
station was 417 Mt/yr. A significant relationship between
the annual sediment loads at Hankou (QSH) and Datong
(QSD) station is obtained as follows:
Q SD ¼ 1:07 Q SH
ðr2 ¼ 0:84; n ¼ 52Þ
ð4Þ
This relationship (see Fig. 8) allowed us to estimate sediment load at Datong (which represents sediment flux from
the Changjiang to the sea) from the data recorded at Hankou station. Thus, we established the centennial-scale time
series of annual sediment flux from the Changjiang to the
sea (1865–2005), including estimates derived from rating
curves for 1865–1950 and from gauging data for 1951–
2005 (Fig. 9a). According to our calculations, over the
100 yr prior to the completion of Danjiangkou Reservoir in
1968, the annual sediment load at Datong station was
488 Mt/yr in average, which is comparable to the estimate
presented by Milliman and Syvitski (1992).
Syvitski and Morehead (1999) proposed an equation to
estimate sediment load from maximum basin relief and basin area according to the balance of gravity-driven sediment
yield and potential energy, such that
Q s ¼ aqg1=2 H3=2 A1=2 ¼ 2 108 H3=2 A1=2
ð5Þ
where a is a constant of proportionality; q is grain density,
usually taken as 2650 kg/m3; g is gravitational acceleration
(9.8 m/s2); H is maximum basin relief (m), and A is basin
area (m2). Eq. (5) was developed from data of 230 rivers
worldwide (Syvitski et al., 2000). The maximum relief of
the Changjiang basin (H) is 6600 m, and its area (A) is
1.94 · 106 km2 (Liu et al., 2007); these values in Eq. (5) give
an annual sediment load of approximately 470 Mt/yr (equivalent to a sediment transport rate of 14,936 kg/s), which is
close to the estimate of Milliman and Syvitski (1992) and the
present study (488 Mt/yr). The estimate from Eq. (5) is valid for the period with little human disturbances on hydrological processes, and is therefore comparable to our
estimate using the rating curve method.
The marked decrease in sediment flux from the Changjiang to the sea over the past 56 yr is well documented, demonstrating the transformation of the Changjiang river
system under intensive human impacts (e.g., Yang et al.,
2002, 2006). The reconstruction of annual sediment flux
from the Changjiang to the sea over the past 140 yr is of
great significance because it provides useful data for studies
of climatic and human impacts on the river system, geomorphologic evolution of the Changjiang estuary and mega-delta, and related biogeochemical cycles in the East China Sea.
Discussion
Shifting dominance over the Changjiang sediment
flux to the sea: 1865 to the present
Yang et al. (2005) discussed the link between water discharge from the Changjiang to the sea and basin-wide
Reconstruction of sediment flux from the Changjiang (Yangtze River) to the sea since the 1860s
327
800
QsD=1.07QsH
r2=0.84, n=52
700
Qs at Datong (Mt/yr)
600
500
400
300
200
100
0
0
100
200
300
400
500
600
700
800
Qs at Hankou (Mt/yr)
Figure 8 Relationship between annual sediment load at Hankou and Datong stations demonstrating a linear relationship that
explains approximately 84% of the data variance.
1 Completion of the Danjiangkou Reservoir in 1968
2 Soil conservation works in the upper reaches since 1980s
3 Operation of the Three Gorges Dam since 2003
Monsoon-dominated period
Human-impacted period
700
1
a
2
600
500
3
400
300
200
200
100
100
0
b
0
East Asian Summer Monsoon Index
-100
4
-200
c
-300
2
Anomaly of Indian Monsoon Rainfall (mm)
Annual sediment flux (Mt/yr)
800
0
-2
-4
-6
1865
1875
1885
1895
1905
1915
1925
1935
1945
1955
1965
1975
1985
1995
2005
2015
Figure 9 (a) Time series of annual sediment flux from the Changjiang to the sea (1865–2005) derived from annual sediment load
at Hankou (Fig. 6b) and Eq. (4); (b) anomalies of Indian Monsoon Rainfall since 1870 (data available at http://www.iges.org/india/
allindia.html). The anomalies are defined as annual deviations from the mean annual rainfall between 1871 and 1994 (853 mm); (c)
Index of East Asian Summer Monsoon, from IPCC (2007b). Solid lines in (a), (b), and (c) are 5-yr running averages.
328
precipitation. Basin-wide precipitation is closely associated
with global climate episodes such as El Niño Southern Oscillation (ENSO) events and related flood-drought cycles (Jiang
et al., 2006; Zhang et al., 2007). Global climate episodes
through teleconnections have affected river hydrological
processes and sediment fluxes to the sea. For example,
ENSO events are usually accompanied by lower regional precipitation in the Huanghe (Yellow River, China) river basin,
and account for a 50% decrease in water discharge to the
sea and 30% of the total decrease of sediment load over
the past 56 yr (Wang et al., 2006, 2007). Weather conditions
and soil erodibility are critical factors that influence both
sediment yield and sediment flux to the sea. Most of the
sediment delivered by the Changjiang is derived from the
upper reaches, from the Jinshajiang and Jialingjiang tributaries in particular (Fig. 1; Yang et al., 2006). Therefore,
precipitation in the upper reaches has an important influence on sediment yield, but is itself substantially affected
by climate change.
During the past 140 yr, there has been a decreasing trend
in maximum annual stream flow in the upper reaches of the
Changjiang but an increasing trend in the middle and lower
reaches (Zhang et al., 2007). This is because different monsoons dominate precipitation in these segments of the river.
For example, the Indian monsoon is dominant in the upper
reaches, whereas the East Asian monsoon is dominant in
the middle and lower reaches (Zhang et al., 2005). These
two climatic systems are spatially and temporally asynchronous, and have different effects on trends and changes of
stream flow. The Indian monsoon influences sediment yield
in the upper reaches where most of the Changjiang sediment load is derived, and it is therefore of primary importance to the Changjiang sediment flux. Comparison of the
time series of annual sediment flux to the sea with variations in Indian monsoon rainfall from the late 1860s to the
present suggests that the Indian monsoon has an important
influence on the Changjiang sediment flux to the sea
(Fig. 9). It appears that during the period of little human
disturbance (1865–1950s), the Changjiang sediment flux
to the sea was dominantly influenced by the Indian monsoon, whereas the influence of the East Asian summer monsoon is unclear (Fig. 9).
Rainfall during the Indian monsoon wet season, which
runs from June to September, accounts for about 70% of annual rainfall in the upper reaches of the Changjiang. Links
between monsoon-related events (e.g., rainfall over East
Asia) weakened between 1890 and 1930 but strengthened
during the period from 1930 to 1970 (Kripalani and Kulkarni,
2001). A strong inverse relationship between Indian monsoon rainfalls and El Niño events has been recognized in several studies (Kumar et al., 1999; Krishnamurthy and
Goswami, 2000; Sarkar et al., 2004) and implies that before
the period of significant human disturbances, annual sediment flux from the Changjiang to the sea was closely associated with global climate change (Fig. 9). The East Asian
monsoon has the dominant influence on precipitation in
the middle and lower reaches of the Changjiang but has little impact in the upper reaches (the source region of Changjiang sediment). In addition, the long-term changes of the
East Asian monsoon are consistent with a tendency for a
southward shift of the summer rain belt over eastern China
(Zhai et al., 2005). Therefore, the monsoons were the dom-
H. Wang et al.
inant influence on sediment flux to the sea during the period
from the 1860s to the 1950s.
Since the 1950s, human activities have become the dominant influence on sediment flux to the sea, leading to stepwise decreases that correspond to major hydrologic events
in the recent history of the river basin (Fig. 9a; Yang
et al., 2006). Thus, the dominant influence on sediment flux
to the sea has shifted from that of monsoons from the 1860s
to the 1950s, to human activities from the 1950s to the present. Although global warming has increased the frequencies
of heavy rainfall and severe droughts since the 1950s (IPCC,
2007a), leading to higher peaks of monthly sediment loads
(see Fig. 7a and compare the 1860s and 1980s), human
activities have been of primary importance to the sediment
flux from the Changjiang to the sea since the 1950s, particularly since the 1980s. At present, the mechanisms by which
the monsoons affect sediment flux to the sea are far from
being totally understood. It is clear that since 1865 monsoons have initially been the dominant influence on sediment flux from the Changjiang to the sea, and that there
is a transition to a period dominated by humans, but the
boundary is not clearly identified. We have used data from
the period of 1954–1968 to determine rating parameters
and then applied those parameters to estimate sediment
loads between 1865 and 1953. The period 1954–1968 may
have been influenced by human activities to some extent.
If so, it would then have an influence on the rating curves
and on the sediment loads estimated from them. Impacts
of human activities will likely continue into the foreseeable
future. Although future global warming is expected to increase the frequency of heavy rainfall and severe drought
(IPCC, 2007a), the effects of human activities on the Changjiang sediment flux will be more direct and serious than
those of global warming.
Human impacts on the rating curves
Few studies on the use of rating curves have considered the
effect of human disturbances on the relationship between
stream discharge and suspended sediment concentration
1.10
?
2003-2005
1.05
1986-2002
1.00
1969-1985
b 0.95
0.90
1955-1968
0.85
1865-1953
0.80
-14
-13
-12
-11
-10
-9
-8
Ln (a)
Figure 10 Changes of the rating parameters due to human
activities in the river basin since 1865, illustrating progressive
decreases in sediment supply and increases in the erosive
power of the river.
Reconstruction of sediment flux from the Changjiang (Yangtze River) to the sea since the 1860s
Deposition
329
Erosion
Annual sediment load (Mt/yr)
800
Yichang
Datong
700
1.2
600
500
1.1
400
1.05
1.0
b
300
200
0.9
0.89
100
0.98
1.01
0.8
0
1950
1955
1960
1965
1970
1975
1980
1985
1990
1995
2000
2005
Figure 11 Comparison of annual sediment load at Yichang and Datong stations from 1950 to 2005 and selected values of rating
exponent b (see also Table 1) showing the transition from a depositional to erosive river environment.
(e.g., Syvitski and Alcott, 1995; Syvitski and Morehead,
1999; Asselman, 2000; Syvitski et al., 2000; Yang et al.,
2007). However, acknowledging the effect of human disturbances on rating parameters is extremely important since
most rivers worldwide have been substantially affected by
human activities (Vörösmarty and Sahagian, 2000; Vörösmarty et al., 2000; Walling and Fang, 2003; Nilsson et al.,
2005; Syvitski et al., 2005; Syvitski and Milliman, 2007).
The effectively gauged Changjiang makes itself an excellent
candidate for study of this topic since it discharges a huge
amount of water and sediment to the East China Sea and
has been extensively affected by human activities over
the past half century.
Over thousands of year human activities have modified
the landscape of Changjiang river basin (e.g., extensive
agricultural reclamations). And the ancient Dujiangyan Irrigation System, constructed in the Minjiang tributary 2000 yr
ago but still working well now, essentially changed the input
of water and sediment from the Minjiang to the mainstream
of Changjiang (Li and Xu, 2006). Meanwhile, the constructions of flood diversions in the middle and lower reaches
also impacted the river geomorphological features. By far
the levees and embankments have been well developed in
the middle and lower reaches to control the floods and prevent from breaching. However, the early human activities
had been of minor importance to the Changjiang sediment
flux to the sea compared to the present situation. Observations at gauging stations on the Changjiang over the past
56 yr indicate that human activities, such as the operation
of dams and undertaking of soil-conservation works, have
decreased the sediment supply from the source region,
and therefore decreased coefficient a of the rating curves
(Figs. 3 and 10, Table 1). Conversely, rating coefficient a
would be expected to increase as a result of human activities such as extensive deforestation or agricultural reclamation. Exponent b of the rating curves increases when
suspended sediment concentrations in streams decrease as
a result of sediment retention within reservoirs, and the
downstream channel becomes subject to scour (Table 1,
Fig. 10). If the recorded stream discharges and estimated
sediment concentrations for the period from 1865 to 1953
are used to establish rating curves, the derived rating
parameters are: ln(a) = 9.1 and b = 0.84. The value of a
is higher than that calculated for 1955–1968, and b is lower
than that calculated for the same period; this indicates that
human disturbances had negligible effect during the first
100 yr of the study period (Fig. 10, Table 1). The Three
Gorges Dam is designed to trap 70% of the sediment supplied from the upper reaches by the time of its planned
completion in 2009 (Yang et al., 2006). This will cause further decrease in coefficient a and increase in exponent b,
because as more sediment is trapped within the Three
Gorges Reservoir, the erosive power of the river will increase substantially in the middle and lower reaches
(Fig. 10).
It is noteworthy that this higher value of b due to human
activities will increase the erosive power of stream discharge such that the river channel will be scoured. When
b increased to a value of 1.05 after commissioning of the
Three Gorges Dam in June 2003, a level much higher than
that for the period 1955–1968 (Fig. 10, Table 1), the river
channel in the middle and lower reaches became subject
to riverbed scour. Annual sediment loads at Yichang station
prior to 2002 were higher than those at Datong station (except for 4 abnormal years, see Fig. 11). We estimated that
the mean annual sediment load at Yichang station (1950–
2002) was approximately 491 Mt/yr, much higher than that
at Datong (426 Mt/yr). This suggests that deposition was
then dominant in the middle and lower reaches (Fig. 11).
The values of b have gradually increased in parallel with
the decrease of sediment load at Datong. However, this situation was reversed after commissioning of the Three
Gorges Dam. The mean annual sediment flux to the sea at
Datong (189 Mt/yr, 2003–2005) then became higher than
the mean annual sediment supply from the upper reaches
(125 Mt/yr at Yichang station), which indicates that there
has been severe scouring in the middle and lower reaches.
This is corroborated by the continuous increase in the calculated value of b and the implied increase in erosive power of
the river (Fig. 11).
Conclusions
The Changjiang is the major source of terrestrial materials
to the East China Sea and to the western Pacific Ocean. It
has been effectively gauged since the 1950s; however, a
long-term time series of sediment flux to the sea (centennial scale) is essentially needed for the studies of geomorphology of the estuary and delta, sedimentology of the
continental shelf, and biogeochemistry in the marginal
seas.
330
In this study, we reconstructed the centennial-scale time
series of annual sediment flux from the Changjiang to the
East China Sea since 1865 based on the records of water discharge at Hankou gauging station. We established rating
curves from the monthly records of water discharge and suspended sediment concentration at Hankou station during
the period 1954–2005. Our results show that human activities in the river basin have substantially affected the rating
curves and their parameters. The commissioning of reservoirs (e.g., the Danjiangkou Reservoir in 1968 and the Three
Gorges Reservoir in 2003) together with soil-conservation
works undertaken in the upper reaches substantially reduced sediment supply to the middle and lower reaches.
This caused a decrease in the value of rating coefficient a.
A marked decrease in suspended sediment concentration increased the erosive power of the river, as indicated by an
increase in rating exponent b. Consequently, the river channel has been scoured in the middle and lower reaches in recent years when the annual sediment flux to the sea became
higher than the sediment supply from the upper reaches.
We used data recorded at Hankou station from 1954 to
1968 to establish rating curves that allow the estimation
of sediment load for the period of 1865–1953 when sediment loads were not monitored. Our estimates for this period combined with data recorded from 1954 to 2005
constitute a long-term time series of sediment loads at Hankou station and a centennial-scale time series of annual sediment flux to the sea that is derived based on linear
regression analysis of sediment loads between Hankou and
Datong. We estimated that the mean annual sediment flux
to the sea during the period from 1865 to 1968, when the
influence of human activities was negligible, was 488 Mt/
yr, a result comparable to the estimate of Milliman and
Syvitski (1992) and to that from an equation proposed by
Syvitski and Morehead (1999). The annual sediment flux
from the Changjiang to the sea from 1865 to the present
can be considered in two periods: a monsoon-dominated
period (1865–1950s) and a human-impacted period
(1950s–present), although the mechanism of monsoon domination has not been well understood yet. During the
human-impacted period, the rating curves changed markedly, leading to significant morphological responses in the
middle and lower reaches and delta system of the
Changjiang.
Acknowledgements
Two anonymous reviewers are appreciated for their constructive suggestions on improving the scientific quality of
our manuscript. We are grateful to the Changjiang Water
Conservancy Committee for the access to valuable data
sets. This work was financially supported by the National
Science Foundation of China (NSFC, Grant Nos. 90211022
and 40676036) and Program for New Century Excellent Talents in University (NCET-06-0598).
References
Asselman, N.E.M., 2000. Fitting and interpretation of sediment
rating curves. Journal of Hydrology 234, 228–248.
H. Wang et al.
Campbell, F.B., Bauder, H.A., 1940. A rating-curve method for
determining silt discharge of streams. EOS Transactions, AGU
21, 603–607.
Chen, Z.Y., Li, J.F., Shen, H.T., Wang, Z.H., 2001. Yangtze River of
China: historical analysis of discharge variability and sediment
flux. Geomorphology 41 (2–3), 77–91.
DeMaster, D.J., McKee, B.A., Nittrouer, C.A., Qian, J., Cheng, G.,
1985. Rates of sediment accumulation and particle reworking
based on radiochemical measurements from continental shelf
deposits in the East China Sea. Continental Shelf Research 4 (1–
2), 143–158.
E, J.P., 2006. Bulletin of Chinese River Sediment. Water Resources
and Hydropower Publishing House, Beijing, China (in Chinese).
Gregory, K.J., Walling, D.E., 1973. Drainage Basin Form and
Process: A Geomorphological Approach. Edward Arnold, London,
p. 456.
Horowitz, A.J., 2003. An evaluation of sediment rating curves for
estimating suspended sediment concentrations for subsequent
flux calculations. Hydrological Processes 17, 3387–3409.
doi:10.1002/hyp.1299.
Huh, C.-A., Chen, H.Y., 1999. History of lead pollution recorded in
East China Sea sediments. Marine Pollution Bulletin 38 (7), 545–
549.
IPCC, 2007a. Climate change 2007: the physical science basis. In:
Solomon, S., Qin, D., Manning, M., Marquis, M., Averyt, K.,
Tignor, M.M.B., Miller, H.L., JrJr, Chen, Z. (Eds.), Contribution
of Working Group I to the Fourth Assessment Report of
Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, New York, pp. 7–8.
IPCC, 2007b. Climate change 2007: the physical science basis. In:
Solomon, S., Qin, D., Manning, M., Marquis, M., Averyt, K.,
Tignor, M.M.B., Miller, H.L., JrJr, Chen, Z. (Eds.), Contribution
of Working Group I to the Fourth Assessment Report of
Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge New York, p. 297.
IRTCES (International Research and Training Center on Erosion and
Sedimentation), 2001. Bulletin of Chinese River Sediment (2000).
Ministry of Water Resources Conservancy, Beijing, China. Available from: <http://www.irtces.org/database.asp> (in Chinese).
IRTCES (International Research and Training Center on Erosion and
Sedimentation), 2002. Bulletin of Chinese River Sediment (2001).
Ministry of Water Resources Conservancy, Beijing, China. Available from: <http://www.irtces.org/database.asp> (in Chinese).
IRTCES (International Research and Training Center on Erosion and
Sedimentation), 2003. Bulletin of Chinese River Sediment
(2002). Ministry of Water Resources Conservancy, Beijing, China.
Available from: <http://www.irtces.org/database.asp> (in
Chinese).
IRTCES (International Research and Training Center on Erosion and
Sedimentation), 2004. Bulletin of Chinese River Sediment
(2003). Ministry of Water Resources Conservancy, Beijing, China.
Available from: <http://www.irtces.org/database.asp> (in
Chinese).
IRTCES (International Research and Training Center on Erosion and
Sedimentation), 2005. Bulletin of Chinese River Sediment
(2004). Ministry of Water Resources Conservancy, Beijing, China.
Available from: <http://www.irtces.org/database.asp> (in
Chinese).
Jiang, T., Zhang, Q., Zhu, D., Wu, Y., 2006. Yangtze floods and
droughts (China) and teleconnections with ENSO activities
(1470–2003). Quaternary International 144 (1), 29–37.
doi:10.1016/j.quaint.2005.05.010.
Kripalani, R.H., Kulkarni, A., 2001. Monsoon rainfall variations and
teleconnections over South and East Asia. International Journal
of Climatology 21, 603–616.
Krishnamurthy, V., Goswami, B.N., 2000. Indian monsoon–ENSO
relationship on interdecadal timescale. Journal of Climate 13,
579–595.
Reconstruction of sediment flux from the Changjiang (Yangtze River) to the sea since the 1860s
Kumar, K.K., Rajagopalan, B., Cane, A.M., 1999. On the weakening
relationship between the Indian monsoon and ENSO. Science
284, 2156–2159.
Li, K., Xu, Z., 2006. Overview of Dujiangyan Irrigation scheme of
ancient China with current theory. Irrigation and Drainage 55
(3), 291–298. doi:10.1002/ird.234.
Liu, J.P., Milliman, J.D., Gao, S., Cheng, P., 2004. Holocene
development of the Yellow River’s subaqueous delta, North
Yellow Sea. Marine Geology 209, 45–67. doi:10.1016/
j.margeo.2004.06.009.
Liu, J.P., Li, A.C., Xu, K.H., Velozzi, D.M., Yang, Z.S., Milliman,
J.D., DeMaster, D.J., 2006. Sedimentary features of the Yangtze
River-derived along-shelf clinoform deposit in the East China
Sea. Continental Shelf Research 26, 2141–2156. doi:10.1016/
j.csr.2006.07.013.
Liu, J.P., Xu, K.H., Li, A.C., Milliman, J.D., Velozzi, D.M., Xiao,
S.B., Yang, Z.S., 2007. Flux and fate of Yangtze River sediment
delivered to the East China Sea. Geomorphology 85, 208–224.
doi:10.1016/j.geomorph.2006.03.023.
Ludwig, W., Probst, J.L., Kempe, S., 1996. Predicting the oceanic
input of organic carbon by continental erosion. Global Biogeochemical Cycles 10, 23–41.
Martin, J.M., Meybeck, M., 1979. Elemental mass-balance of
material carried by major world rivers. Marine Chemistry 7,
173–206.
Meybeck, M., Vörösmarty, C., 2005. Fluvial filtering of land-toocean fluxes: from natural Holocene variations to Anthropocene.
Comptes Rendus Geoscience 337, 107–123. doi:10.1016/
j.crte.2004.09.016.
Milliman, J.D., Meade, R.H., 1983. World-wide delivery of river
sediment to the oceans. Journal of Geology 91, 1–21.
Milliman, J.D., Shen, H.T., Yang, Z.S., Meade, R.H., 1985. Transport and deposition of river sediment in the Changjiang estuary
and adjacent continental shelf. Continental Shelf Research 4 (1–
2), 37–45.
Milliman, J.D., Syvitski, J.P.M., 1992. Geomorphic/tectonic control
of sediment discharge to the ocean: the importance of small
mountainous rivers. Journal of Geology 100, 525–544.
Morgan, R.P.C., 1995. Soil Erosion and Conservation, second ed.
Longman, London, p. 198.
Nilsson, C., Reidy, C.A., Dynesius, M., Revenga, C., 2005. Fragmentation and flow regulation of the world’s large river systems.
Science 308, 405–408. doi:10.1126/science.1107887.
Phillips, J.M., Webb, B.W., Walling, D.E., Leeks, G.J.L., 1999.
Estimating the suspended sediment loads of rivers in the LOIS
study area using infrequent samples. Hydrological Processes 13,
1035–1050.
Reid, I., Frostick, L.E., 1987. Discussion of conceptual models of
sediment transport in streams. In: Thorne, C.R., Bathurst, J.C.,
Hey, R.D. (Eds.), Sediment Transport in Gravel-Bed Rivers. John
Wiley, New York, pp. 410–411.
Saito, Y., Yang, Z., Hori, K., 2001. The Huanghe (Yellow River) and
Changjiang (Yangtze River) deltas: a review on their characteristics, evolution and sediment discharge during the Holocene.
Geomorphology 41 (2–3), 219–231.
Sarkar, S., Singh, R.P., Kafatos, M., 2004. Further evidences for the
weakening relationship of Indian rainfall and ENSO over India.
Geophysical Research Letters 31, L13209. doi:10.1029/
2004GL020259.
Shen, H.T., 2001. Material Flux of the Changjiang Estuary. China
Ocean Press, Beijing, p. 176 (in Chinese).
Smith, S.V., Renwick, W.H., Buddemeier, R.W., Crossland, C.J.,
2001. Budget of soil erosion and deposition for sediments
and sedimentary organic carbon across the conterminous
United States. Global Biogeochemical Cycles 15 (3), 696–
707.
Syvitski, J.P.M., 2003. Supply and flux of sediment along hydrological pathways: research for the 21st century. Global and
331
Planetary Change 39, 1–11. doi:10.1016/S0921-8181(03)000080.
Syvitski, J.P.M., Alcott, J.M., 1995. RIVER3: simulation of water and
sediment river discharge from climate and drainage basin
variables. Computers and Geoscience 21, 89–151.
Syvitski, J.P.M., Morehead, M.D., 1999. Estimating river-sediment
discharge to the ocean: application to the Eel margin, northern
California. Marine Geology 154, 13–28.
Syvitski, J.P.M., Milliman, J.D., 2007. Geology, geography, and
humans battle for dominance over the delivery of fluvial
sediment to the coastal ocean. Journal of Geology 115, 1–19.
Syvitski, J.P.M., Burrell, D.C., Skei, J.M., 1987. Fjords: Processes
and Products. Springer-Verlag, New York, p. 379.
Syvitski, J.P.M., Morehead, M.D., Bahr, D.B., Mulder, T., 2000.
Estimating fluvial sediment transport: the rating parameters.
Water Resources Research 36 (9), 2747–2760.
Syvitski, J.P.M., Vörösmarty, C.J., Kettner, A.J., Green, P., 2005.
Impact of humans on the flux of terrestrial sediment to the
global coastal ocean. Science 308, 376–380. doi:10.1126/
science.1109454.
Thomas, R.B., 1988. Monitoring baseline suspended sediment in
forested basins: the effects of sampling on suspended sediment
rating curves. Hydrological Sciences Journal 33, 499–514.
Vörösmarty, C.J., Sahagian, D., 2000. Anthropogenic disturbance of
the terrestrial water cycle. Bioscience 50, 753–765.
Vörösmarty, C.J., Green, P., Salisbury, J., Lammers, R., 2000.
Global water resources: vulnerability from climate change and
population growth. Science 289, 284–288.
Walling, D.E., 1977. Assessing the accuracy of suspended sediment
rating curves for a small basin. Water Resources Research 13,
531–538.
Walling, D.E., Moorehead, P.W., 1989. The particle size characteristics of fluvial suspended sediment: an overview. Hydrobiologia
176–177 (1), 125–149. doi:10.1007/BF00026549.
Walling, D.E., Fang, D., 2003. Recent trends in the suspended
sediment loads of the world’s rivers. Global and Planetary
Change 39, 111–126. doi:10.1016/S0921-8181(03)00020-1.
Wang, Y., Ren, M.-E., Syvitski, J.P.M., 1997. Sediment transport
and terrigenous flux. In: The Sea. In: Robinson, A.R., Brink, K.H.
(Eds.), The Global Coastal Ocean: Processes and Methods, vol.
10. Wiley, New York, pp. 346–401.
Wang, H.J., Yang, Z.S., Saito, Y., Liu, J.P., Sun, X., 2006. Interannual
and seasonal variation of the Huanghe (Yellow River) water
discharge over the past 50 years: connections to impacts from
ENSO events and dams. Global and Planetary Change 50 (3–4),
212–225. doi:10.1016/j.gloplacha.2006.01.005.
Wang, H.J., Yang, Z.S., Saito, Y., Liu, J.P., Sun, X., Wang, Y., 2007.
Stepwise decreases of the Huanghe (Yellow River) sediment load
(1950–2005): impacts of climate changes and human activities.
Global and Planetary Change 57 (3–4), 331–354. doi:10.1016/
j.gloplacha.2007.01.003.
Xu, K.Q., Chen, Z., Zhao, Y., Wang, Z., Zhang, J., 2005. Simulated
sediment flux during 1998 big-flood of the Yangtze (Changjiang)
River, China. Journal of Hydrology 313, 211–233. doi:10.1016/
j.jhydrol.2005.03.006.
Xu, K., Milliman, J.D., Yang, Z.S., Wang, H., 2006. Yangtze
sediment decline partly from Three Gorges Dam. EOS, Transactions, American Geophysical Union 87 (19), 185, 190.
Yang, G.F., Chen, Z., Yu, F., Wang, Z., 2007. Sediment rating
parameters and their implications: Yangtze River, China. Geomorphology 85 (3–4), 166–175. doi:10.1016/j.geomorph.2006.
03.016.
Yang, S.L., Zhao, Q.Y., Belkin, I.M., 2002. Temporal variation in the
sediment load of the Yangtze River and the influences of human
activities. Journal of Hydrology 263 (1–4), 56–71.
Yang, S.L., Gao, A., Hotz, H.M., Zhu, J., Dai, S.B., Li, M., 2005.
Trends in annual discharge from the Yangtze River to the sea
(1865–2004). Hydrological Sciences 50 (5), 825–836.
332
Yang, Z., Wang, H., Saito, Y., Milliman, J.D., Xu, K., Qiao, S., Shi,
S., 2006. Dam impacts on the Changjiang (Yangtze) River
sediment discharge to the sea: the past 55 years and after the
Three Gorges Dam. Water Resources Research 42, W04407.
doi:10.1029/2005WR003970.
Zhai, P.-M., Zhang, X.-B., Wan, H., Pan, X.-H., 2005. Trends in total
precipitation and frequency of daily precipitation extremes over
China. Journal of Climate 18, 1096–1108.
H. Wang et al.
Zhang, Q., Tong, J., Gemmer, M., Becker, S., 2005. Precipitation,
temperature and runoff analysis from 1950 to 2002 in the
Yangtze basin, China. Hydrological Sciences Journal 50 (1), 65–
80. doi:10.1623/hysj.50.1.65.56338.
Zhang, Q., Xu, C., Jiang, T., Wu, Y., 2007. Possible influence of
ENSO on annual maximum streamflow of the Yangtze River,
China. Journal of Hydrology 333, 265–274. doi:10.1016/
j.jhydrol.2006.08.010.