® .Geomorphology 41 2001 77­91 www.elsevier.comrlocatergeomorph Yangtze River of China: historical analysis of discharge variability and sediment flux Zhongyuan Chen a,) , Jiufa Li a , Huanting Shen a , Wang Zhanghuab a State Key Laboratory for Estuarine and Coastal Research, East China Normal Uniersity, Shanghai, 200062, China b Department of Geography, East China Normal Uniersity, Shanghai, 200062, China Received 14 February 2000; received in revised form 17 August 2000; accepted 26 May 2001 Abstract ® .Hydrological records covering a 100-year period from the upper, middle and lower Yangtze River were collected to examine the temporal and spatial distribution of discharge and sediment load in the drainage basin. The Yangtze discharge, as expected, increases from the upper drainage basin downstream. Only an estimated 50% of the discharge is derived from the upper Yangtze, with the rest being derived from the numerous tributaries of the middle and lower course. However, the distribution of sediment load along the Yangtze is the reverse of that observed for discharge, with most of the sediment ® 8 .being derived from the upper basin. A dramatic reduction in sediment load by ;0.8=10 tonsryear occurs in the middle Yangtze because of a marked decrease in slope and the change to a meandering pattern from the upper Yangtze rock sections. Considerable siltation also occurs in the middle Yangtze drainage basin as the river cuts through a large interior Dongting Lake system. Sediment load in the lower Yangtze, while significantly less than that of the upper river, is somewhat higher than the middle Yangtze because of additional load contributed by adjacent tributaries. A strong correlation exists between the discharge and sediment load along the Yangtze drainage basin during the dry season as lower flows carry lower sediment concentration. During the wet season, a strong correlation is also present in the upper Yangtze owing to the high flow velocity that suspends sand on the bed. However, a negative to poor correlation occurs in the middle and lower Yangtze because the flow velocity in these reaches is unable to keep sand in suspension, transporting only fine-grained particles downstream. ® .Hydrological data are treated for 30 years 1950­1980 , when numerous dams were constructed in the upper Yangtze drainage basin. At Yichang and Hankou hydrological stations, records revealed a decreasing trend in annual sediment load, along with slightly reduced annual discharge at the same stations. This can be interpreted as the result of water diversion primarily for agriculture. Sediment load at Datong further downstream is quite stable, and not influenced by slightly reduced discharge. Furthermore, sediment concentration at the three hydrological stations increased, which can be attributed to sediment loss in association with intensifying human activity, especially in the upper drainage basin, such as deforestation and construction of numerous dams. Mean monthly sediment load of these 30 years pulses about 2 months behind discharge, implying dam-released sediment transport along the entire river basin during the high water stage. q 2001 Elsevier Science B.V. All rights reserved. Keywords: Human impact; Discharge; Seasonal fluctuation; Sediment load; Yangtze drainage basin ) Corresponding author. Department of Geography, Yangtze Delta Program, East China Normal University, Shanghai, 200062, China. Tel.: q86-21-62232706; fax: q86-21-62232416. ® .E-mail address: Z.Chen@gislab.ecnu.edu.cn Z. Chen . 0169-555Xr01r$ - see front matter q 2001 Elsevier Science B.V. All rights reserved. ® .PII: S0169-555X 01 00106-4 ( )Z. Chen et al.rGeomorphology 41 2001 77­9178 1. Introduction Runoff and sediment load play an important role in modifying river morphology and patterns through time and space, and supporting riverine ecosystems ®along adjoining fluvial surfaces Hupp and Os- terkamp, 1996; Gupta et al., 1999; Lu and Higgitt, .1998, 1999; Miller and Gupta, 1999 . The physical processes driving these two important components result from the interplay among the source of water and sediment, sediment flux, fluvial geomorphology ®and atmospheric circulation Milliman and Syvitski, .1992; Gupta and Asher, 1998 . Over its geological history, China's Yangtze River, which flows from west to east to debouch into the East China Sea, has served as a link between nature and people. Its abundant fluvial resources, operating under the eastern Asian monsoon were vital for early Chinese Neolithic civilization ) ® .10,000 years B.P. Chang, 1986 . The huge Yangtze drainage basin, which is more than 6300 km in length and has a catchment area of 1.94=106 km2 , can be divided into the upper, middle and lower Yangtze reaches, primarily on the basis of geology and climate, and secondarily on the basis of the resulting geomorphology of the river. The upper Yangtze is more than 4300 km long from the source to Yichang, and has a total drainage 4 2 ® .area of about 100=10 km Tong and Han, 1982 . ® .The mainstem the Jingshajiang is jointed by four major tributaries in this river section: the Ya- ®longjiang, Mingjiang, Jialingjiang and Wujiang Fig. .1, . These rivers originate on the Tibetan plateau where the elevation is generally over 4000 m, and the rivers are deeply incised into rocky canyons with )1000 m in elevation difference between the ® .riverbeds and mountain peaks Fig. 2a . The chan- nels are about 0.5­1.5 km wide, 5­20 m deep, and vary in slope from 10­40=10y5 , reaching a maxi- y5 ® .mum of 450=10 Fig. 3 . The middle Yangtze is 950 km in length, and has 4 2 ®a total drainage area of about 68=10 km Tong .and Han, 1982 . Three large inputs join the main stream in this section: the Dongting Lake drainage basin, Hanjiang River and Poyang Lake drainage ® .basin Fig. 1; . This section of the river starts from Yichang, at the end of the Three-Gorges reach, to Hukuo, an outlet of the Yangtze on Poyang Lake ® .Fig. 1 . In the middle Yangtze, the slope decreases dramatically to 2­3=10y5 , but at some locations slope can be as high as 6­8=10y5 . Locally nega- tive bed slope occurs in the middle Yangtze where the two interior Dongting and Poyang Lakes are ® .located Figs. 1 and 3 . The channel of middle Yangtze is wider and deeper than the upper course, with the width between 1 and 2 km and the depth between 6 and 15 m. A typical meandering river pattern with many cutoffs prevails in this river reach, where the river exits from the upper rock-confined ® .valley into the Jingjiang fluvial plain Fig. 2b . The well-known Jingjiang dike has been constructed along the river and elevated over time to 12­16 m in ® .different places above the ground surface Fig. 2c . This is the most vulnerable region for flood hazards ®Changjiang Hydrological Committee of Hydrology .Ministry, 1997; Li et al., 1999 . The Three-Gorges Dam will be completed in 2009 in the Yichang reach ® .to serve as the flood control Fig. 2d . Several major meander cutoffs on the Jingjiang plain occurred dur- ing the early to mid-20th century. In the 1960s and the 1970s, a set of cutoff projects were used to ® .manage the fluvial environment You, 1987 . This stabilized the migration of the river channel although the course of the river was shortened by approxi- mately 78 km. This straightening of Jingjiang has resulted in tremendous siltation downstream in the ®river channel, bypasses, and interior lakes You, .1987 . Consequently, river and lake beds were raised ®to increase flood hazard potentials Yin and Li, this .volume . Below the Hukuo, the final 930 km constitutes the lower Yangtze River with a total drainage area of 12=104 km2 . Several large interior lakes, such as ® .the Chaohu and Taihu Fig. 1; and in associa- tion with many tributaries, drain into the lower Yangtze. This segment of the river wanders among plains and hills, on which a high-stage water level about 8 m above mean water level, is clearly marked ® .Fig. 2e . The slope of riverbed decreases to 0.5­1.0 =10y5 , and the channel widens to 2­4 km and deepens to 10­20 m. The river channel can however, be wider than )15 km and as shallow as ;6 m in ® .its estuarine region Fig. 2f . The lower Yangtze drainage basin, including its large estuarine system, benefits largely from upstream discharge and sedi- ® .ment accumulation Chen et al., 1988 . These are ( )Z. Chen et al.rGeomorphology 41 2001 77­91 79 ® .Fig. 1. The location of the Yangtze drainage basin, the major tributaries and interior lakes are indicated. Arrows A, B, C show the plan of the South-North Transfer Water Project. vital natural resources for the intensive agriculture, fisheries, waterfowl production, vegetable irrigation and domestic consumption in the lower basin. Large ®estuarine islands up to 40 km wide and 100 km .long and vast coastal wetlands, primarily due to rapid sediment progradation seaward during the last ® .2000 years B.P. Chen et al., 2000 have been used for agriculture, housing and industry. On average, Southwest Pacific typhoons effect the lower Yangtze basin directly or indirectly one to three times each summer. On some occasions, typhoon-generated storms have met the annual Yangtze flood, raising ®sea level along the estuary by 2­3 m Shao et al., .1991 . 2. Methods and observations In the study reported here, records of discharge and suspended sediment concentration in the Yangtze drainage basin were collected to examine their spa- tial and temporal distributions. Suspended sediment ( )Z. Chen et al.rGeomorphology 41 2001 77­9180 ( )Z. Chen et al.rGeomorphology 41 2001 77­91 81 Fig. 3. The Yangtze fluvial variables, represented by slope, width and depth, along the entire river basin. concentration was measured two or three times per month and averaged from four to six different water depths. Suspended sediment load is the product of discharge and concentration. Measurements were carried out along transects at hydrological stations ®located along the Yangtze River Yangtze Water .Conservancy Committee, 1875­1985 . Three major hydrological stations were chosen for this study at Yichang, Hankou and Datong, representing the up- per, middle and lower Yangtze River, respectively, ® .Fig. 1 . Data from Loushan station positioned near Chengliji, east of the Dongting Lake in the middle ® .Fig. 2. Fluvial landscape of the Yangtze drainage basin. a Upper Yangtze topography showing high relief between the water level and the ® . ® . ® . ® .mountain peaks )1000 m ; b meandering river pattern and wide river channel )1000 m of the middle Yangtze; c Jingjiang dike along the middle Yangtze River, which is now )12 m above the adjacent ground surface, this dike was inundated during the big flood in ® . ® .1998, and later was raised to another 1.0 m; d Three Gorges dam under construction, to be ready in 2009; e lower Yangtze plain reaches ® .with local rock outcrops with about 8 m high flood marker that is almost 1.0 m higher than adjacent fluvial plain; f Yangtze estuary, )10 km wide. ( )Z. Chen et al.rGeomorphology 41 2001 77­9182 Fig. 4. Relation between discharge and time. ( )Z. Chen et al.rGeomorphology 41 2001 77­91 83 Fig. 5. Relation between sediment load and time. ( )Z. Chen et al.rGeomorphology 41 2001 77­9184 Fig. 6. Correlation of sediment load to discharge. ( )Z. Chen et al.rGeomorphology 41 2001 77­91 85 ® .Yangtze Fig. 1 were also used. The period record of these three stations did not start simultaneously. ®Hankou station has the longest record 1875­1980 .for discharge and 1950­1980 for sediment load , ®Yichang has the second longest 1875­1980 for .discharge and 1953­1980 for sediment load , and ®Datong has the shortest 1950­1985 for both dis- .charge and sediment load . All the data are docu- mented in an internal report AAnnual Report on ®Yangtze Water and SedimentB Yangtze Water Con- .servancy Committee, 1875­1985 . Data for dis- ®charge and sediment load after 1980 Yichang and . ® .Hankou and 1985 Datong , respectively, have not been released to the public and thus have not incor- porated into the present study. Discharge and sediment load were plotted against time to help understand the variations in water and sediment flux and their impact on the fluvial mor- phology in the Yangtze drainage basin over the last ® .century Figs. 4 and 5 . Mean monthly data were calculated for July, August and January for the three sites to represent wet and dry seasons in the study ®area. Mean annual discharge and sediment load sum .of monthly values divided by 12 for the same period are also indicated on Figs. 4 and 5. There is a marked difference in discharge between ® .the wet and dry seasons Fig. 4 . At Yichang, dis- charge ranges approximately from 2.0 to 4.0=104 m3 rs in the wet season, in contrast to 0.4=104 m3 rs during the dry season. Downstream at Hankou ® .about 1000 km east of Yichang discharge increases to ;3.0­5.0=104 m3 rs in the wet season and about 0.6­1.1=104 m3 rs during the dry season. ®Further downstream, at Datong about 700 km east . 4 Hankou discharge ranges from about 3.3­6.0=10 m3 rs during the wet season to 0.9­1.2=104 m3 rs ® .in the dry season Fig. 4 . It is also obvious that the frequency of major discharge peaks along the Yangtze drainage basin increases downstream. There was almost no single major discharge peak exceeding 5.5=104 m3 rs in the upper Yangtze drainage basin in the 20th century ® .Fig. 4 . In contrast, five major peaks of this size occurred in the middle Yangtze drainage basin within ®the same time period, and even more peaks at least .nine times from 1950 to 1985 occurred in the lower ® .Yangtze drainage basin Fig. 4 . The spatial variations in sediment load along the Yangtze show a trend opposite to that displayed by ® .discharge Fig. 5 . Sediment load in the upper Yangtze, as recorded by Yichang hydrological sta- tion, varies during the wet season from 1.0 to 2.0= 8 ® 8 .10 tonsrmonth about 0.5=10 tons mean annual . The sediment load is considerably lower in the dry season and is too low to be shown in Fig. 5. Down- stream at Hankou, sediment load during the wet season decreases to about 0.6­1.2=108 tonsrmonth ® 8 .about 0.35=10 tons mean annual , while the dry season load is again very low, being only about 0.1=108 tonsrmonth. In the lower Yangtze, sedi- ment load during the wet season, as recorded by the Datong hydrological station ranges from about 0.7­ 1.4=108 tonsrmonth, less than the upper Yangtze, but nearly the same as that from the middle Yangtze. Mean annual sediment load in the lower Yangtze ® 8 .about 0.45=10 tons is slightly higher than the ® .middle Yangtze Fig. 5 . The dry season sediment load in the lower Yangtze is also similar to that in ® 8 the middle reach of the order of ;0.1=10 .tonsrm , but still higher than that in the upper Yangtze. We analyzed the correlation between discharge and sediment load for the past decades. The results demonstrate that along the upper, middle and lower Table 1 ® .Variables of suspended sediment Yangtze Water Conservancy Committee, 1875­1985 Station Annual Maximum Annual Maximum Minimum sediment sedimentConcen- Date Annual Year Annual Year concentration loadtration load load 3 63 6 6® . ® .kgrm =10 tons® . ® . ® .kgrm =10 tons =10 tons Yichang 1.175 10.5 1959.7.26 516 754 1954 321 1992 Luoshan 0.677 5.66 1975.8.12 435 615 1981 307 1992 Hankou 0.594 4.42 1975.8.14 421 579 1964 267 1954 Datong 0.504 3.24 1959.8.06 451 678 1964 309 1992 ( )Z. Chen et al.rGeomorphology 41 2001 77­9186 Table 2 ® .Average monthly precipitation along the upper, middle and lower Yangtze drainage basin modified after You, 1987 . Shaded box represents the wet season Yangtze drainage basin a strong positive correlation ® .significant level as0.001 exists between dis- ®charge and sediment load during the dry season Fig. .6 . The middle Yangtze has the highest correlation ® .coefficient Rs0.94 for January and February , ®which is followed by the lower Yangtze Rs0.90 .for January, Rs0.85 for February and then fol- ®lowed by the upper Yangtze Rs0.79 for January, .Rs0.57 for February . In contrast, our analysis reveals generally no significant correlation between ®discharge and sediment load in the middle Rs .y0.12 for July and August and lower Yangtze ® .Rs0.03 for July, Rsy0.02 for August; Fig. 6 during the wet season. However, a strong correlation ®Rs0.79 for July, Rs0.80 for August, significant .level as0.001 still exists during the wet season in the upper Yangtze. Also, calculated are the variables of suspended ® .sediment Table 1 , average monthly and annual ® .precipitation Table 2 , and average monthly distri- ®bution of suspended sediment concentration Table . ®3 of the Yangtze River data after Yangtze Water .Conservancy Committee, 1875­1985; You, 1987 . ® .These data all indicate: 1 decreased sediment con- Table 3 ® 3.Average monthly distribution of the suspended sediment concentration kgrm along the upper, mid and lower Yangtze drainage basin ® .Yangtze Water Conservancy Committee 1875­1985 . Shaded box represents the wet season ( )Z. Chen et al.rGeomorphology 41 2001 77­91 87 ®centration and sediment load downstream Tables 1 . ® .and 3 ; and 2 concentrated precipitation from March to September, which increases from the upper ® .drainage basin downstream Table 2 . Mean annual sediment load and concentration, and mean annual discharge over the past 30 years ® .1950­1980 for Yichang, Hankou and Datong hy- drological stations are shown on Fig. 7, respectively. Three 7­8-year cycles of sediment and discharge fluctuations are recognizable during that time period. Regression analysis indicates decreasing trends in sediment load and discharge at the Yichang and ® .Hankou stations Fig. 7a,c , except for Datong's sediment load, which increases slightly with time ® .Fig. 7a . Sediment concentration at the three sta- ® .tions also increases slightly Fig. 7b over that time Fig. 7. Fluctuations of sediment distribution and discharge with ® . ® . ® .time 1950­1980 ; a sediment load, b sediment concentration, ® .c discharge. Fig. 8. Monthly runoff discharge and sediment load averaged for ® .30 years 1950­1980 . period. The distribution of average monthly dis- charge and sediment load over the past 30 years ® .1950­1980 is also depicted to show the non-coeval fluctuations between the discharge and sediment load ® .pattern at the three hydrological stations Fig. 8 . 3. Discussion 3.1. Discharge ® .The annual Yangtze discharge Fig. 4 increases as the drainage basin area increases downstream. The recorded data indicate increases in mean discharge, 1.4=104 m3 rs at Yichang to 2.3=104 m3 rs at Hankou and then to 2.8=104 m3 rs at Datong ®Yangtze Water Conservancy Committee, 1875­ .1985 . On the basis of these data, it seems that only about 50% of the annual runoff discharge in the lower Yangtze drainage basin is derived from the upper basin. This is due probably to both increased drainage basin area and higher precipitation down- ®stream China Academy Science and Ministry of .Electrohydrology, 1963 . Precipitation increases gradually from 400 to 1600 mmryear from the ® .upper to lower Yangtze Table 2 . The network of ( )Z. Chen et al.rGeomorphology 41 2001 77­9188 tributaries that are located along the mid and lower Yangtze drainage basin also deliver a high regional ® .runoff volume into the Yangtze Fig. 1 . Although the annual runoff that discharges to the Yangtze estuary can be as high as 8.94=1011 m3 , this is however, much less than the volume of tidal water incoming from the sea. The annual tidal prism into the river mouth has a total volume of 83.98= 1011 m3 , which is an order of magnitude greater than that of the annual Yangtze runoff discharge to the ® .sea Chen et al., 1988 . The water chemistry in the estuary therefore will be sensitive to the impacts of ASouth to North Water Transfer ProjectB, which has ® .been planned upstream Fig. 1 . Saltwater will in- trude further into the estuary, especially in the winter season. This could affect irrigation, and domestic and industrial water supplies in the estuarine region, where the metropolitan city Shanghai, with )12 million people located. The ecosystem of the large Yangtze estuary is nursed by the upper stream discharge regarding the growth of large estuarine islands, freshwater water- table maintenance and aquaculture. The role of high-river stage, in associated with tidal oscillations along the river mouth area, has been long deliberated for various industrial and agricultural purposes. Ex- amples include maintaining a smooth channel, partic- ularly through the estuarine turbidity maximum zone where the water depth is only 7 m, and the reclama- tion of tidal flats which has enlarged the Yangtze coastal land by nearly 50% during the last century ® .Chen et al., 1988 . Any significant change of the Yangtze discharge from upper stream will alter the ecological balance in such an important habitat base. 3.2. Sediment load Suspended sediment load dominates over the bed- load in the Yangtze River, representing about 95% of ®the total Yangtze sediment flux Yangtze Water Conservancy Committee, 1875­1985; Chen et al., .1988 . The decrease in sediment load downstream from the upper Yangtze is directly responsible for siltation within several large-scale interior lakes con- nected with the mid and lower Yangtze basin. Data collected by the Yangtze Water Conservancy Com- ® .mittee 1875­1985 reveal that nearly 27% of the ® 8 sediment load 5.16=10 tonsryear measured to- .tally at Yichang station is deposited in Dongting Lake, before the flow leaves the lake. About 0.59= 108 tonsryear suspended sediment load are carried out of the lake through the Chenglinji outlet to the ® .east Fig. 1 to merge with the remaining sediment in the Yangtze trunk stream. The total volume mea- sured at the hydrological station of Loushan is 4.35 8 ® .=10 tonsryear Table 1 . The Hanjiang is a major tributary that meets the ® .middle Yangtze at Hankou Fig. 1 . The Dan- jiangkou dam constructed in the upper Hanjiang basin in the late 1960s has been responsible for a 70% reduction in the suspended sediment load for the Hangjiang from a pre-dam level of 0.85=108 tonsryear to a present level of 0.25=108 tonsryear ® .Tong and Han, 1982 . Ordinarily, the sediment load in the Hankou station would be expected to increase due to the input from the Hangjiang, but the sus- pended sediment load measured at Hankou actually decreases to 4.21=108 tonsryear when compared ® .with upstream Loushan Table 1 . This indicates that siltation occurs obviously along the middle Yangtze River between Chenglinji and Hankou section. Below Hankou station, the Poyang Lake drainage ® .system Fig. 1 carries a low suspended sediment ® 8 load 0.1=10 tonsryear; Yangtze Water Conser- .vancy Committee, 1875­1985 to the main Yangtze River. However, data measured at Datong station, located downstream, demonstrate that suspended sed- iment in the Yangtze increases to 4.51=108 tonsr ® .year Table 1 . We propose that this increase results from both net erosion andror re-mobilization of numerous longitudinal bars and the carving of soft riverbanks along the lower Yangtze, together with significant inputs of sediment from the regional trib- ® .utaries Fig. 1 , especially during the wet season. 3.3. The correlation between discharge and sediment load The relationship between discharge and sediment load along the Yangtze is affected by a number of river channel variables, including slope, flow veloc- ®ity, nature of sediment and river pattern Higgitt and .Lu, 1996 , plus the fact that discharge is used to calculate sediment load. The strong correlation be- tween discharge and sediment load observed during the dry season along the entire Yangtze arises par- ( )Z. Chen et al.rGeomorphology 41 2001 77­91 89 tially because the low discharges during this season ® .are matched by very low sediment loads Fig. 6 . The same high level of correlation in the upper Yangtze during the flood season can be interpreted as reflecting the fact that high flow velocities and stream power during high stages are able to mobilize and transport commensurately high sediment loads. The poor correlation that occurs in the middle Yangtze during the flood season results from the ® .abruptly diminished slope Fig. 3 and lower veloci- ties which are unable to transport as much sand as the upper reach. Concentrations of fine sediment are related to supply and not to flow. Sediment load is largely reduced in the Yichang area, where the river undergoes a transition from the upper Yangtze rock- cut channel to the middle Yangtze meandering course ® .Fig. 2 . A large amount of sediment settles out in Dongting Lake, where the slope changes to 1­2= y5 ® .10 or even becomes negative Fig. 3 . These river variables in the middle Yangtze greatly alter the balance between discharge and sediment concentra- tion. The Danjiangkou Dam along the upper Han- jiang basin in the middle Yangtze, which cuts off a tremendous amount of sediment but does not alter ® .the mean discharge greatly Tong and Han, 1982 , also contributes to the poor relationship. The poor correlation between discharge and sediment load still ® .occurs in the lower Yangtze River Fig. 6 . Low flow velocity and further reduced riverbed slope accounts for this low correlation. 3.4. Sediment and runoff discharge fluctuations oer the last 30 years The 7­8-year cycles of sediment and discharge during 1950­1980 along the Yangtze drainage basin ® .Fig. 7a may be related to variations in precipitation ®driven by the El Nino southern oscillation Neelin~ .and Latif, 1998 although more research is required on this possibility. The decreasing trend in sediment load for Yichang and Hankou is due to reduced ® .discharge over the last 30 years Fig. 7a,c . This reduction in discharge over that time period is possi- bly caused by water diversion for intensifying agri- culture along the upper and middle Yangtze stream. In contrast, the slight increase in sediment load at Datong on the lower Yangtze is associated with the increase in sediment concentration at all three sta- ® .tions, especially Yichang Fig. 7b . The overall in- crease in sediment concentration along the river probably reflects the recent intensive human impact, such as widespread deforestation in the upper ® .Yangtze drainage basin Higgitt and Lu, 1996 . We would expect sediment changes due to human activ- ity in the upper Yangtze to cause greater increase in sediment concentration downstream than the low rates that we have observed, particularly at Hankou ® .and Datong Fig. 7b . This low increase in sediment concentration possibly results from numerous small to medium size river dams constructed in the upper Yangtze drainage basin, mostly during 1950s to 1970s. This has reduced the large volume of sedi- ®ment discharging downstream cf. Lu and Higgitt, .1998 . 3.5. Seasonal ariation of runoff and sediment flux Fluctuation of the Yangtze discharge and sedi- ment load varies seasonally, primarily reflecting ® .runoff from monsoonal precipitation Table 2 . Vari- ability of discharge and sediment load can be charac- terised by two extremes: maximum peak and mini- ® .mum low events Figs. 4 and 8 . For instance, at Yichang, the maximum discharge reaches 5.1=104 m3 rsrmonth, and minimum can be as low as 0.45 =104 m3 rsrmonth. At Datong, the maximum peaks to 8.5=104 m3 rsrmonth and the minimum is only 0.89=104 m3 rsrmonth. Rainfall along the Yangtze drainage basin directly controls the pattern of annual ® .discharge distribution Fig. 8; Table 2 . The rainy season along the entire Yangtze drainage basin usu- ally occurs from May to September, but varies in the upper, middle and lower Yangtze basins. Precipita- tion along the upper basin is concentrated from May to September, which amounts to 79.1% of the year total. Precipitation along the middle and lower basin is primarily between March and August, amounting to 70.9% and 72.8% of the annual total, respectively. ®Precipitation in the winter season January and De- .cember usually represents -1%. ® .During the wet season May to October , the sediment load measured at Yichang makes up 93% ® .of the annual total Shen et al., 1992 . Over the same period, the sediment load at Datong station approxi- mates 85% of the total, and July alone represents 22.7% of the annual load. This may be contrasted to ( )Z. Chen et al.rGeomorphology 41 2001 77­9190 the month of February, when sediment load is 0.66% of the annual total. At Yichang station, mean sus- pended sediment concentration is only 0.205 kgrm3 ® .in the dry season November to April; Table 3 , and can be as high as 1.310 kgrm3 in the rainy season ® .May to October . Similar patterns occur in Hankou and Datong stations, showing 0.253 and 0.221 kgrm3 in the dry season, and 0.701 and 0.647 kgrm3 in the ® .rainy season Table 3 . ® .Based on the 30-year database Fig. 8 , distribu- tion pattern of mean monthly discharge and sediment load through the entire Yangtze drainage basin did not fluctuate coevally in the wet season. Apparently, there is a time lag between the two. A generally symmetrical annual precipitation pattern results in a symmetrical increase in mean monthly hydrograph ® .starting from early April . In contrast, the sediment load pattern is asymmetrical, indicating a nearly 2-month lag in the sediment pulse over the same time scale. This time-lagged sediment load pattern ® .based on 30-year 1950­1980 observation may also record the human-impacted sediment transport along the river channel. A large volume of sediment is primarily trapped behind the dams in wet season flood. As subsequent high-river stage occurs, numer- ous dams from all tributaries in the upper drainage basin release highly concentrated sediment down- stream, causing this asymmetrical sediment load dis- tribution pattern. 4. Summary This study demonstrates that the Yangtze dis- charge in the middle and lower reaches tends to increase by about 50% over that in the upper reach above Yichang. This results from the contribution of numerous tributaries that join the middle and lower Yangtze and from an increase in precipitation from the upper drainage basin towards the lower Yangtze. On the basis of a century long database, annual sediment load has decreased by 0.8=108 tons in the middle Yangtze, which reflects significant siltation along the main river channel and in the large interior Dongting Lake of central China, next to the Yangtze River. This can be explained primarily by the strik- ing decrease in channel slope along the middle Yangtze downstream of the Three Gorges. The sedi- ment load in the lower Yangtze is much less than that of the upper Yangtze, but slightly higher than that in the middle Yangtze. The additional sediment sources are probably from in-channel erosion of sand bars and soft banks, coupled with input from the tributaries that join the main channel along its lower course. A statistically significant linear relationship exists between the discharge and sediment load in the upper, middle and lower Yangtze during the dry season. This relationship reflects that low discharges during the dry season carry very low sediment loads. In the wet season, a strong correlation also occurs in the upper Yangtze, where high flow velocities and stream power during high stages are able to mobilize and transport commensurately high sediment loads. Conversely, linear correlation coefficients are weak or even negative for the middle and lower Yangtze in the wet season. This may indicate a high increase in suspended sediment concentration as wash load processes, in association with decreased flow veloc- ity downstream to the estuary. ® .Over the 30-year time scale 1950­1980 , sedi- ment load in the upper and middle Yangtze had been decreasing along with discharge, which is possibly due to water diversion for intensive agriculture of the region. In contrast, an overall, but very slight in- crease in sediment concentration had occurred along the entire Yangtze and caused an increase in sedi- ment load in the lower Yangtze. This phenomenon can be attributed to human activity, such as defor- estation and dam construction in the upper Yangtze drainage basin. Mean monthly discharge and sediment load along the Yangtze are highly affected by seasonal precipi- tation. During the wet season, sediment load peaks almost 2 months after the discharge throughout the Yangtze drainage basin, probably recording dam-re- leased sediment load during the high river water stage. Acknowledgements Authors would like to thank Drs. Gordon E. Grant, Ray A. Kostaschuk, Colin R. Thorne and Adrian M. Harvey, who kindly reviewed this ( )Z. Chen et al.rGeomorphology 41 2001 77­91 91 manuscript. Misses Li, X.P., Yang, M. and Mr. Song, B.P. helped to complete the diagrams. China ®National and Natural Science Foundation Grant no. .49971011 , TCTPF-China and US National Geo- ® .graphical Society Grant no. 6693-00 funded this study. References Chang, K.C., 1986. The archaeology of ancient China. 4th edn. Yale Univ. Press, New Haven, CT. Changjiang Hydrological Committee of Hydrology Ministry, 1997. Report on River Channel Management of Mid-and Lower ® .Yangtze Interior Report, unpublished, in Chinese . Chen, J.Y., Shan, H.T., Yun, C.X., 1988. Processes of Dynamics and Geomorphology of the Changjiang Estuary. Shanghai ®Scientific and Technical Publishers, Shanghai in Chinese with .English summary . Chen, Z., Song, B.P., Wang, Z.H., Cai, Y.L., 2000. The Late Quaternary evolution of the subaqueous Yangtze delta: stratig- raphy, sedimentation, palynology and deformation. Marine Geology 162, 423­441. China Academy Science and Ministry of Electrohydrology, 1963. ® .Atlas of China Hydrology, 136 pp. unpublished, in Chinese . Gupta, A., Asher, M.G., 1998. Environment and the Developing World: Principles, Policies and Management. Wiley, Chich- ester. Gupta, A., Kale, V.S., Rajaguru, S.N., 1999. The Narmada River, India, through space and time. In: Miller, A.J., Gupta, A. ® .Eds. , Varieties of Fluvial Form. Wiley, Chichester, pp. 114­ 143. Higgitt, D.L., Lu, X.X., 1996. Patterns of sediment yield in the Upper Yangtze Basin, China. In: Walling, D.E., Webb, R.W. ® .Eds. , Erosion and Sediment Yield: Global and Regional Perspectives. Proceedings of the Exeter Symposium, July 1996. IAHS Publication 236, Walllingford, pp. 205­214. Hupp, C.R., Osterkamp, W.R., 1996. Riparian vegetation and fluvial geomorphic processes. Geomorphology 14, 277­295. Li, C.A., Yin, H.F., Chen, D.X., Wang, B., 1999. Problems and strategies for flood control of middle reaches of Yangtze River: reflection on catastrophic flood of Yangtze River in 1998. Earth Science--Journal of China University of Geo- ® . ® .sciences 24 4 , 329­334 in Chinese, with English Summary . Lu, X.X., Higgitt, D.L., 1998. Recent changes of sediment yield in the Upper Yangtze China. Environmental Management 22, 697­709. Lu, X.X., Higgitt, D.L., 1999. Sediment Yield Variability in the Upper Yangtze China. Earth Surface Processes and Landforms 24, 1077­1093. ® .Miller, A.J., Gupta, A. Eds. , Varieties of Fluvial Form. Wiley, Chichester. Milliman, J.D., Syvitski, J.P.M., 1992. Geomorphicrtectonic con- trol of sediment discharge to the ocean: the importance of small mountainous rivers. Journal of Geology 100, 525­544. Neelin, J.D., Latif, M., 1998. El Nino Dynamics. Physics Today December 1998, 32­36. Shao, X.S., Yan, Q.S., Xu, S.Y., Chen, Z.Y., 1991. Storm de- posits in the coastal region of Shanghai, the Yangtze Delta China. Geologie en Mijinbouw 70, 45­58. Shen, H.T., Li, J.F., Han, M.B., 1992. Transport of the Suspended Sediment in the Changjiang estuary. International Journal of ® .Sediment Research 7 3 , 45­63. Tong, Z.J., Han, Q.W., 1982. River Channel Morphology and Evolution Below Danjiangkou Dam, Hanjiang River. Hydro- logical Bureau of Changjiang River Planning Office, Wuhan, ® .pp. 23­36 Interior report in Chinese . Yin, H.F., Li, C. Human impact on floods and flood disasters on ® .the Yangtze River this volume . You, L.Y., 1987. A study of the formation and evolution of braided channels with stable islands: the middle and lower ® .reaches of the Yangtze River. In: Gardiner, V. Ed. , Interna- tional Geomorphology 1986, Part I. Wiley, Chichester, pp. 649­662. ® .Yangtze Water Conservancy Committee 1875­1985 . Annual ®Report of Changjinag Water and Sediment interior report, .unpublished, in Chinese .