basin denudation and sedimentation - School of Earth Sciences

gayoldMechanics

Feb 21, 2014 (3 years and 8 months ago)

84 views

The Ganges and Brahmaputra rivers in Bangladesh:
basin denudation and sedimentation
Mohammad Rezwanul Islam,* Syeda Fahliza Begum,Yasushi Yamaguchi
and Katsuro Ogawa
Department of Earth and Planetary Sciences,Graduate School of Sciences,Nagoya University,
Furo Cho,Chikusa Ku,Nagoya 464-8602,Japan
Abstract:
Every year the Ganges and Brahmaputra rivers in Bangladesh transport 316 and 721 million tonnes of
sediment,respectively.These high loads of suspended sediment re¯ect the very high rate of denudation in their
drainage basins.The average mechanical denudation rate for the Ganges and Brahmaputra basins together is
365 mm10
3
yr
ÿ1
.However,the rate is higher in the Brahmaputra Basin than that in the Ganges Basin.Several
factors,including mean trunk channel gradient,relief ratio,runo￿,basin lithology and recurring earthquakes
are responsible for these high denudation rates.Of the total suspended sediment load (i.e.1037 million tonnes)
transported by these rivers,only 525 million tonnes (c.51%of the total load) are delivered to the coastal area of
Bangladesh and the remaining 512 million tonnes are deposited within the lower basin,o￿setting the
subsidence.Of the deposited load,about 289 million tonnes (about 28%of the total load) are deposited on the
¯oodplains of these rivers.The remaining 223 million tonnes (about 21%of the total load) are deposited within
the river channels,resulting in aggradation of the channel bed at an average rate of about 3
.
9 cmyr
ÿ1
.
Although the Brahmaputra transports a higher sediment load than the Ganges,the channel bed aggradation
rate is much higher for the Ganges.This study also documents a wide range of interannual,seasonal and daily
variation in suspended sediment transport and water discharge.Interannual variation in sediment deposition
within the basin is also suggested.Copyright
#
1999 John Wiley & Sons,Ltd.
KEY WORDS
Ganges;Brahmaputra;suspended sediment load;basin subsidence;denudation;sedimentation
INTRODUCTION
The two Himalayan rivers,the Ganges and Brahmaputra,are among the most sediment-laden rivers in the
world (Milliman and Meade,1983).Every year these two rivers carry a large sediment load from the
Himalayan Mountains to the Bengal Basin and have formed the Ganges±Brahmaputra Delta,which is the
largest in the world.Based on a comparison between Rennell's (1781) map and a modern map,Morgan and
McIntire (1959) concluded that the delta front had advanced little over the last 200 years.They suggested
that most of the sediment deposited within the subsiding delta simply balanced the subsidence.Coleman
(1969) suggested that strong tidal currents at the mouths of the river systemprevented sediment fromsettling,
so that it was transported through a deep submarine canyon known as the`Swatch of No Ground'to the
deeper ocean.On the other hand,studies undertaken by Curray and Moore (1974) and Emmel and Curray
(1985) found no evidence of recent sediment accumulation on the Bengal deep sea fan.They found that most
of the sediment was being trapped on the lower deltaic plain and on the shelf.
CCC 0885±6087/99/172907±17$1750
Received 30 October 1998
Copyright
#
1999 John Wiley & Sons,Ltd.
Accepted 29 March 1999
HYDROLOGICAL PROCESSES
Hydrol.Process.13,2907±2923 (1999)
*Correspondence to:Dr M.R.Islam,Department of Earth and Planetary Sciences,Nagoya University,Nagoya,464-8602,Japan.
E-mail:islam@eps.nagoya.u.ac.jp
In this study,an attempt is made to quantify the sediment load transported by the Ganges and the
Brahmaputra rivers.Attempts are also made to quantify sediment accumulation on the land and the delivery
of sediment to the coastal sea of the Bay of Bengal through the combined Ganges±Brahmaputra River
system.Denudation rates in the basins of these rivers are also discussed.This information is important for
simulating sedimentation processes between the foot of the Himalayan Mountains and the Bengal Deep Sea
Fan.This work is now being undertaken.
An understanding of water discharge and sediment transport in Bangladesh extends beyond academic
interest to embrace several important environmental problems.As a consequence of global warming,this
country could face a serious problem due to sea level rise.With higher sea levels,additional areas will be
inundated by ¯oods and a￿ected by tidal surges.Inland freshwater lakes,ponds and aquifers are likely to be
a￿ected by saline and brackish water intrusion.The present limit of tidal in¯uence could extend further north.
Moreover,as a ¯at and low-lying country,Bangladesh could lose a signi®cant portion of its valuable land to
the sea.However,delivery of sediment to the coastal sea might,at least partially,reduce these problems.
THE GANGES RIVER
The Ganges River (Figure 1) originates from the Gangotri glaciers at an elevation of 7010 m near the
Tibet±Indian border in the high Himalayas.It ¯ows south-east across India and Bangladesh and eventually
Figure 1.The drainage basins of the Ganges and Brahmaputra rivers,with the locations of the gauging stations (modi®ed fromMorgan
and McIntire,1959)
Copyright
#
1999 John Wiley & Sons,Ltd.HYDROLOGICAL PROCESSES,VOL.13,2907±2923 (1999)
2908
M.R.ISLAM ET AL.
discharges into the Bay of Bengal.Before entering Bangladesh,it bifurcates into two tributaries.One,called
the Hoogly River,¯ows south through Calcutta into the Bay of Bengal.The other,called the Padma River,
¯ows into Bangladesh,joins the Brahmaputra (Jamuna) at Goalondo and drains into the Bay of Bengal.The
combined ¯ow of the Ganges and the Brahmaputra is also known as the Padma in Bangladesh.For
convenience the combined ¯owof these two rivers will be referred to as the combined Ganges±Brahmaputra
River system in this study.From source to sea,the Ganges is about 2500 km long with a drainage area of
about 980 000 km
2
.Only 34188 km
2
(3
.
6%) of this drainage area lies within Bangladesh.The Ganges within
the plain of Bangladesh is primarily a meandering river system.
THE BRAHMAPUTRA RIVER
The Brahmaputra River (Figure 1) rises froma large glacier in the southern slopes of the high Himalayas and
its course extends over 2896 km to the sea.It ¯ows for 1600 kmthrough China (Tibet),for 900 kmthrough
eastern India and for 400 km through Bangladesh.It has a drainage area of about 640 000 km
2
.Only
50505 km
2
(8
.
7%) of the drainage area lies within Bangladesh.By the beginning of 19th century this river had
shifted westwards to its present course,which is known as the Jamuna in Bangladesh.Its former principal
course,known as the Old Brahmaputra,has nowturned into a small distributary of the Jamuna and is ¯owing
through Maymonsing district.This former principal course is supposed to have formed only the eastern part
of the Delta (Figure 1).The present principle course is much younger than that of the Ganges and has made
a much smaller contribution to building the Ganges±Brahmaputra Delta (Kahn,1991).The Brahmaputra
can be classi®ed as a braided river system.In China,it is known as the Yarlung/Tsang Po River.
THE SUSPENDED SEDIMENT LOADS OF THE GANGES AND BRAHMAPUTRA RIVERS
There are several published estimates of the total suspended sediment loads of the Ganges and Brahmaputra
rivers (Tables I and II).However,to obtain a clearer understanding of the sediment load and its deposition,
the suspended sediment loads of the Ganges and Brahmaputra rivers in Bangladesh are estimated again in
this study.These estimates are based on the daily mean water discharge (m
3
s
ÿ1
) and the available sediment
discharge data fromthe di￿erent gauging stations (1,2 and 3 in Figure 1) provided by the Bangladesh Water
Development Board (BWDB).In most cases suspended sediment samples are collected during the wet season
(May/June to October,also known as the high discharge period).However,in some years the data cover both
the wet and dry (November to April/May) seasons.The sampling intervals are mostly irregular,with the
Table I.Estimates of the sediment load of the Ganges River cited in di￿erent studies
Suspended sediment
(million tonnes yr
ÿ1
)
Reference Gauging stations/
sampling locations
Period of
measurement
328*,{ Abbas and Subramanian (1984) Calcutta 1981
729{,{ Abbas and Subramanian (1984) Farakka 1981
403{ Singh (1988) Not available Not available
485 Coleman (1969) Hardinge Bridge 1958±1962
520 BWDB} (1972) Hardinge Bridge 1966±1969
375 NEDECO} (1967) Not available 1960?
680 Milliman and Meade (1983) Hardinge Bridge 1966±1967
1600 Holeman (1968) Bengal Delta Not available
350±600 Hossain (1982) Hardinge Bridge 1980±1986
316 This study Hardinge Bridge 1979±1995
*This estimate is for the Hoglee tributaries of the Ganges River.
{Estimated for Indian reach.Rest of the estimates stand for Bangladesh reach.
{Estimate before bifurcation of the Ganges River.
} For the abbreviations,please see the text.
Copyright
#
1999 John Wiley & Sons,Ltd.HYDROLOGICAL PROCESSES,VOL.13,2907±2923 (1999)
BASIN DENUNDATION AND SEDIMENTATION
2909
shortest being once a week.This irregularity is more prominent in the case of the Bahadurabad gauging
station of the Brahmaputra River.In this work,we have tried to select the years which provide representative
discharge scenarios for both the wet and dry seasons.The suspended sediment loads have been estimated
using sediment rating curves involving empirical correlations between sediment discharge and water dis-
charge for the individual river.Discrete values of the two variables are plotted after logarithmic trans-
formation.The general form of the sediment rating curve is:
S  aQ
b
where S is the sediment discharge (kg s
ÿ1
),Q is the water discharge (m
3
s
ÿ1
),a is a constant and b is an
exponent.In this study`b'is 1
.
51 for the Ganges,1
.
56 for the Brahmaputra and 1
.
23 for the combined
Ganges±Brahmaputra rivers (Figures 2±4).Goswami (1985) found that values of this exponent ranged from
Table II.Sediment load of the Brahmaputra River estimated in di￿erent studies
Suspended sediment load
(10
6
t yr
ÿ1
)
Reference Gauging stations/
sampling location
Period of
measurement
402* Goswami (1985) Pandu,Asam 1955±1979
710* Subramanian (1987) Not available Not available
650 Hossain (1992) Bahadurabad 1982±1988
750 NEDECO (1967) Bangladesh 1960?
617 Coleman (1969) Bahadurabad 1958±1962
541 BWDB (1972) Bahadurabad 1967±1969
1157 Milliman and Meade (1983) Bahadurabad 1966±1967
800 Holeman (1968) Bengal Delta Not available
721 This study Bahadurabad 1989±1994
*Estimated for Indian reach.Rest of the estimates stand for Bangladesh reach.
Figure 2.Sediment rating curve for the Ganges River
Copyright
#
1999 John Wiley & Sons,Ltd.HYDROLOGICAL PROCESSES,VOL.13,2907±2923 (1999)
2910
M.R.ISLAM ET AL.
1
.
77 to 2
.
28 for the load/Qrelationship for the Brahmaputra River in India.Several studies (e.g.Walling and
Webb,1981) have reported that loads estimated using sediment rating curves will underestimate the actual
load.But use of ¯ood event data reduced the degree of underestimation (Walling and Webb,1988).It is
important to note that suspended sediment concentrations in both rivers are quite high (on average,
1300 mgl
ÿ1
for the Ganges and 1006 mgl
ÿ1
for the Brahmaputra;after Coleman,1969).
The mean annual suspended sediment load of each river was estimated using a single rating curve for the
whole period of record.The average suspended load of the Ganges River was estimated using 13 years (1980,
1983±1986 and 1988±1995) of data from the Hardinge Bridge gauging station (Figure 2).The mean annual
suspended sediment load was estimated to be about 316 million tonnes yr
ÿ1
.This value is lower than other
estimates produced in the recent past.Although not conclusive,it seems likely that dams on the Ganges have
decreased the sediment discharge.However,more detailed analysis is needed to substantiate any such trend.
The mean annual suspended sediment load of the Brahmaputra River was estimated using data from the
Bahadurabad gauging station.Our ®ndings indicate that the mean annual suspended sediment load of
the Brahmaputra is about 721 million tonnes.This load estimate was based on six years of data extend-
ing from1989 to 1994.This estimate shows that the Brahmaputra transports more than twice the suspended
sediment load carried by the Ganges River in Bangladesh.Together they transport a mean annual suspended
sediment load of 1037 million tonnes into Bangladesh.Although Figures 2±4 suggest a trend of increasing
sediment discharge with increasing water discharge,signi®cant variation is seen in the sediment discharge
measured for similar levels of water discharge (particularly in Figures 2 and 4).This will be discussed in
detail in a later section.
The sediment loads of the Ganges and Brahmaputra rivers,estimated by di￿erent studies are cited in
Tables I and II,respectively.These estimated loads cover a wide range.Such variation can be partly
Figure 3.Sediment rating curve for the Brahmaputra River
Copyright
#
1999 John Wiley & Sons,Ltd.HYDROLOGICAL PROCESSES,VOL.13,2907±2923 (1999)
BASIN DENUNDATION AND SEDIMENTATION
2911
attributed to seasonal and interannual variations in the suspended sediment load (Subramanian and
Ramanathan,1996).Another reason might be spatial variability of suspended sediment transport.
For the sediment loads reported for the Ganges River (Table I),the estimate provided by Holeman (1968)
is signi®cantly higher than other estimates and di￿ers signi®cantly from other contemporary or subsequent
estimates.Although,he did not indicate the methodology employed to derive this estimate,it appears that it
was based on estimates of basin erosion.Coleman's (1969) estimates are based on sampling that he carried
out from 1958 to 1962.He estimated the mean annual sediment loads for both rivers.The sediment load
reported by Goswami (1985) was derived for the period 1955±1979 (disregarding the gap in the data from
1961±1965 and in 1970).The sediment load of the Ganges (Table I) obtained from this work is in close
agreement with the values reported by the Netherlands Engineering Consultants Ltd.(NEDECO,1967 in
Khalil et al.,1995),Singh (1988) and Abbas and Subramanian (1984).The suspended sediment load for the
Brahmaputra (Table II) estimated in this study conforms closely to the values reported by NEDECO(1967)
and Subramanian (1987) and is also close to the estimates provided by Hossain (1992),Coleman (1969) and
Holeman (1968).
THE SUSPENDED SEDIMENT LOAD OF THE GANGES±BRAHMAPUTRA RIVER
The suspended sediment load of the combined Ganges±Brahmaputra River was estimated using data from
the Mawa gauging station (Figure 1) for the 13-year period (1979,1982±1983,1985±1986 and 1988±1995).
The Mawa gauging station is about 45 km upstream from the Brahmaputra±Ganges±Meghna con¯uence
and about 160 km and 235 km downstream from the Hardinge Bridge and Bahadurabad Ghat gauging
stations,respectively.This estimate suggests that the combined Ganges±Brahmaputra transports a mean
annual suspended sediment load of about 525 million tonnes (Figure 4).This load represents about 51%of
Figure 4.Sediment rating curve for the Ganges±Brahmaputra River
Copyright
#
1999 John Wiley & Sons,Ltd.HYDROLOGICAL PROCESSES,VOL.13,2907±2923 (1999)
2912
M.R.ISLAM ET AL.
the total suspended load (i.e.1037 million tonnes yr
ÿ1
) transported by the Ganges and the Brahmaputra into
Bangladesh.As the Mawa station is close to the mouth of the Ganges±Brahmaputra±Meghna river system
(Figure 1),the value of 525 million tonnes represents the mean annual input of sediment to the coastal sea of
the Bay of Bengal.This value ignores the contribution fromthe Meghna,since it carries a signi®cantly lower
suspended sediment load than the Ganges and Brahmaputra rivers (after Coleman,1969).Sedimentation
between Mawa and the mouth of these rivers was also ignored.However,this proposed load is signi®cantly
lower than some previous estimates.For example,Coleman (1969) suggested that about 1
.
13 billion tonnes
yr
ÿ1
are transported to the coastal sea of Bangladesh.Alam (1996) also speculated that this load was
1
.
27 billion tonnes yr
ÿ1
.
The main reason for the lower value reported in this study could be the location of the Mawa gauging
station,which is much closer to the river mouth.Coleman (1969) estimated the sediment load delivered to
the sea by summing up the sediment loads of the Ganges and Brahmaputra rivers at Hardinge Bridge and
Bahadurabad Ghat,respectively,which are further inland than Mawa.Thus,sediment deposition in the
downstream reach was ignored.It should also be noted that,besides the Ganges±Brahmaputra River in
Bangladesh,a signi®cant portion of the suspended load of the Ganges River is transported to the coastal sea
through the Hoogly distributary.Abbas and Subramanian (1984) estimated that the suspended load of this
distributary at Calcutta in India (Figure 1),was 328 million tonnes yr
ÿ1
.Taking into account the sediment
load discharged through the Hoogly distributary (Figure 1),the mean annual suspended sediment load
transported to the coastal sea by the Ganges and Brahmaputra rivers can be estimated to be about
850 million tonnes.
SEASONAL AND INTERANNUAL VARIATIONS IN SUSPENDED SEDIMENT TRANSPORT
Both the Ganges and Brahmaputra exhibit signi®cant seasonal and interannual variations in suspended
sediment transport.Figure 5 shows interannual variations in the suspended sediment load of the Ganges,
Brahmaputra and of the combined Ganges±Brahmaputra.The annual loads were estimated using the rating
curve based on that year's data.The sediment load of the Ganges River varied from 155 to 863 million
tonnes yr
ÿ1
during the period 1979±1980,1983±1986 and 1988±1995.The suspended load of the Brahma-
putra varied from455 to 992 million tonnes yr
ÿ1
over the period 1989±1994.A similar degree of variability
was evident in the sediment load of the combined Ganges±Brahmaputra River during the period 1979,
1982±1983,1985±1986 and 1988±1995.
Figure 5.Interannual ¯uctuations in suspended sediment load of the Ganges,Brahmaputra and combined Ganges±Brahmaputra rivers
Copyright
#
1999 John Wiley & Sons,Ltd.HYDROLOGICAL PROCESSES,VOL.13,2907±2923 (1999)
BASIN DENUNDATION AND SEDIMENTATION
2913
Seasonal (monthly) variations in the suspended sediment discharge of the Brahmaputra at Bahadurabad
over the three-year period (1990,1992±1993) are shown in Figure 6.While the suspended sediment discharge
in the monsoon wet season (June to October) increases abruptly compared with the dry season in all three
years,the peak monthly suspended sediment discharge occurs in di￿erent months of the monsoon season.
Data presented by Coleman (1969) showed that about 87%of the total sediment load and 75%of the total
runo￿ of the Brahmaputra River were discharged during the monsoon period.The same work suggested that
about 96% of the total sediment load and 83% of the total runo￿ of the Ganges River were discharged
during the monsoon period.
WATER DISCHARGE
In terms of water and sediment discharge,the Ganges and Brahmaputra rivers rank among the
principal rivers of the world.The average discharge of the Ganges River is 11600 m
3
s
ÿ1
and that of the
Brahmaputra is 19 200 m
3
s
ÿ1
.The mean annual water discharge of the combined Ganges±Brahmaputra
River is 31 000 m
3
s
ÿ1
and 971 10
9
m
3
yr
ÿ1
(Coleman,1969).The mean annual precipitation over the
Ganges and Brahmaputra basins is 1516 mm (Ludwig and Probst,1998).The mean annual runo￿ of the
Ganges and Brahmaputra are 373 and 951 mm,respectively (Maybeck,1976).
Interannual ¯uctuations of water discharge of the Ganges,Brahmaputra and of the combined Ganges±
Brahmaputra rivers are illustrated in Figure 7.As with the sediment discharge scenario presented in Figure 5,
the water discharge of all of these rivers also ¯uctuates signi®cantly.Comparison between the two sets of
®gures suggests that they are not always ¯uctuating in parallel.Higher water discharges (humid years) are not
always marked by higher sediment discharges.The monthly discharge hydrographs of the Brahmaputra for
1993 and 1994 and the hydrograph for the Ganges fromJune to December 1994 are shown in Figure 8.Both
the Ganges and Brahmaputra experience very high water discharges in the monsoon season.These high
discharges are driven by a combination of snowmelt runo￿ from the high Himalayas together with local
monsoon rainfall in north and north-eastern India and the Bangladesh.The Himalayan sector of the
Brahmaputra drainage basin in particular receives some of the highest precipitation (5000 mmyr
ÿ1
) in the
world.The monsoon rains from the June to September account for 60±70% of the total annual rainfall
(Goswami,1985).
Figure 8 also suggests that the Brahmaputra reaches its maximumdischarge between June and August and
its minimumdischarge between January and March.The Ganges reaches its maximumin August.There are
currently insucient data to study the variation of discharge in the Ganges during the dry season.However,
Figure 6.Average sediment discharge of the Brahmaputra River in the months from January to October of 1990,1992 and 1993
Copyright
#
1999 John Wiley & Sons,Ltd.HYDROLOGICAL PROCESSES,VOL.13,2907±2923 (1999)
2914
M.R.ISLAM ET AL.
data presented by Coleman (1969) indicate that the Ganges reaches its lowest discharge between February
and May.
THE RELATIONSHIP BETWEEN WATER AND SEDIMENT DISCHARGE
The patterns of water and sediment discharge of the combined Ganges±Brahmaputra River during the
monsoon period of 1995 are illustrated in Figure 9.Although both the sediment and water discharge increase
during the monsoon period,they do not follow the same trend.As in the case of Figures 5 and 7,Figure 9
also suggests a weak relationship between water discharge and sediment discharge.The lack of accordance in
Figure 7.Interannual ¯uctuations in water discharge of the Ganges,Brahmaputra and combined Ganges±Brahmaputra rivers
Figure 8.Average water discharge of the Ganges and Brahmaputra rivers
Copyright
#
1999 John Wiley & Sons,Ltd.HYDROLOGICAL PROCESSES,VOL.13,2907±2923 (1999)
BASIN DENUNDATION AND SEDIMENTATION
2915
the relationship between these two parameters suggests that sediment discharge may be controlled by both
¯ow hydraulics and sediment availability.The latter will be in¯uenced by geological,hydrological,vegeta-
tion cover and anthropogenic activities.Values of daily ¯ow and the associated suspended sediment
concentration (SSC) are plotted in Figure 10.Again the lack of a strong relationship between water discharge
and SSC con®rms the previous conclusion.It is interesting to note that large ¯uctuations in water,sediment
discharge and in SSC occur even during the monsoon season (Figures 9 and 10).
MECHANICAL DENUDATION OF THE BASIN
Many researchers have estimated basin denudation rates from river load data (Holeman,1968;Curray and
Moore,1971;Gosswami,1985;Summer®eld and Hulton,1994).Such estimates are derived by assuming
uniformremoval of mass over the basin equivalent to the average annual load measured at the basin outlet.
Despite the potential error caused by the overgeneralization of uniformremoval of sediment fromthe entire
Figure 9.Water and suspended sediment discharge for 1995 at Mawa gauging station of the combined Ganges±Brahmaputra River
Figure 10.Water discharge and respective suspended sediment concentration for 1995 at the Mawa gauging station for the combined
Ganges±Brahmaputra River
Copyright
#
1999 John Wiley & Sons,Ltd.HYDROLOGICAL PROCESSES,VOL.13,2907±2923 (1999)
2916
M.R.ISLAM ET AL.
watershed,`the values of sediment yield per unit drainage area'thus obtained provide the most commonly
used index of basin denudation (Goswami,1985).
In this work,mechanical denudation rates were estimated using information on the suspended sediment
load and bed load carried by the Ganges and Brahmaputra rivers.For the Brahmaputra,we used the
estimate of suspended sediment load obtained fromthis study (i.e.721 million tonnes yr
ÿ1
).However,for the
load of the Ganges we employed the sediment load of 729 million tonnes yr
ÿ1
estimated for Farakka in
India (Table I) by Abbas and Subramanian (1984).As the Ganges bifurcates at Farakka,it is important to
consider the sediment load measured at this gauging station or one should sum up the loads of these two
branches estimated at Hardinge Bridge and Calcutta,respectively.Unfortunately previous researchers have
ignored this requirement.
Estimates of bed load are commonly unavailable for large rivers,since it is dicult to measure accurately.
For a fewlarge rivers,where bed load has been measured or calculated,it has been shown to be 10%or less of
the total suspended sediment discharge (e.g.Emmett,1981).The bed load of the Brahmaputra River involves
considerable controversy.Curray and Moore (1971) suggested that the bed load was 25±50% of the
suspended sediment load based on the observations of Coleman (1969).On the other hand,Goswami (1985)
suggested that it was 5±15%of the suspended load.However,in this study,we have estimated the bed load as
10% of the suspended sediment load for both the Ganges and Brahmaputra rivers.After considering the
suspended load and bed load,the mechanical denudation rate for the Ganges and Brahmaputra basin stands
at 303 mm and 458 mm10
3
yr
ÿ1
,respectively.Excluding the bed load this estimate stands at 276 mm and
417 mm10
3
yr
ÿ1
,respectively.These calculations assume a bulk density of 2700 kg m
ÿ3
.Assuming the same
bulk density,Summer®eld and Hulton (1994) estimated these rates as 257 mm and 670 mm10
3
yr
ÿ1
,
respectively.If we assume a bulk density of 1000 kg m
ÿ3
,the mechanical denudation rates become much
higher Ð818 mm and 1240 mm10
3
yr
ÿ1
for the Ganges and Brahmaputra basins,respectively.
The contemporary mechanical denudation rate for the whole Ganges±Brahmaputra River Basin,as
estimated in this study,is 365 mm and 985 mm10
3
yr
ÿ1
,assuming a bulk density of 2700 kg m
ÿ3
and
1000 kg m
ÿ3
,respectively.These values are summarized in Table III.Compared with the average denuda-
tion rate of 60 mm10
3
yr
ÿ1
reported for North America by Judson and Ritter (1964) and a global average of
30 mm10
3
yr
ÿ1
,the denudation rate of the Ganges±Brahmaputra basin is very high.
Goswami (1985) suggested that the present denudation rate of the Brahmaputra Valley in Assamis higher
(1150 mm10
3
yr
ÿ1
) than the rate (30 mm10
3
yr
ÿ1
) over the last 2±3 million years.However,Einsele (1992)
suggested a mean denudation rate of 435 mm10
3
yr
ÿ1
for the period 20 to 7
.
7 million years ago,and a value
of 300 mm10
3
yr
ÿ1
for the period between 7
.
7 million years ago and the present for the Ganges and
Brahmaputra basins.Summer®eld and Hulton (1994) content that estimates of basin denudation derived
from mass ¯ux estimates are not signi®cantly di￿erent from estimates of long-term rates based on sediment
volume and thermochronological data.Although a direct comparison is not possible,the present denudation
rate estimated from this work is consistent with the denudation scenario reported by Einsele (1992) and
supports Summer®eld and Hulton's (1994) conclusion.
Table III.Sediment load and denudation of the Ganges and Brahmaputra basins
River Suspended load
(10
6
t yr
ÿ1
)
Bed load
(10
6
t yr
ÿ1
)
Denudation
(mm10
3
yr
ÿ1
)
Chemical load
(10
6
t yr
ÿ1
)
Chemical load
(t km
ÿ2
yr
ÿ1
)
Ganges 729* 73{ 303{ 75} 77
Brahmaputra 720 72{ 458{ 51} 80
Ganges±Brahmaputra 1450 145{ 365{ 126 78
*Suspended load at Farakka (Abbas and Subramanian,1984).
{Considering bed load is 10%of the suspended load.
{Assuming rock density is 2700 kg m
ÿ3
.
} Summer®eld and Hulton (1994).
}Sarin and Krishnaswami (1984).
Copyright
#
1999 John Wiley & Sons,Ltd.HYDROLOGICAL PROCESSES,VOL.13,2907±2923 (1999)
BASIN DENUNDATION AND SEDIMENTATION
2917
Compared with many other large rivers,the mean speci®c suspended sediment yields of the Ganges
and Brahmaputra are very high (Table IV).For the Brahmaputra,it is 1126 t km
ÿ2
yr
ÿ1
.However,Walling
and Webb (1996) showed that the sediment yield of many rivers with relatively small drainage areas exceed
10000 t km
ÿ2
yr
ÿ1
.The highest value is reported as 53500 t km
ÿ2
yr
ÿ1
for the Haungfuachan River
(3199 km
ÿ2
) in China,which is a tributary of the Yellow River and drains the loess region.
FACTORS AFFECTING BASIN DENUDATION
Several researchers have studied the factors controlling basin denudation rates.Summer®eld and Hulton
(1994) estimated the contemporary denudation rates for the externally drained basis of 33 rivers exceeding
5 10
5
km
2
in area.They concluded that variables re¯ecting basin relief characteristics (e.g.basin relief,
relief ratio,mean trunk channel gradient,etc.) and runo￿ are strongly associated with basin denudation
rates.Mean annual precipitation has a lesser in¯uence.Basin area,runo￿ variability and mean temperature
were seen to be weakly associated with basin denudation rates.In their list of denudation rates the three
Himalayan rivers,the Brahmaputra,the Ganges and the Indus,showed the three highest rates.These rivers
are also characterized by very high channel gradients,relief ratios and runo￿ and by high mean annual
precipitation.In contrast,based on an analysis of 280 rivers,Milliman and Syvitski (1992) suggested that
sediment loads/yields are a log-linear function of basin area and maximumbasin elevation and that mean net
precipitation and runo￿ generally a￿ect the sediment yield to a lesser extent.They concluded that the high
erosion rates found throughout much of southern Asia partly re¯ect poor soil conservation,deforestation
and the increasing intensity of agricultural activities.
The high denudation rates of Himalayan rivers could also be attributed to the rapid uplift of the
Himalayas (Sharma et al.,1978),the presence of poorly compacted and easily erodible recent alluviumover
a vast area of the drainage basin and seismic activity (Morgan and McIntire,1959;Goswami,1985;
Subramanian and Ramanathan,1996).In a recent study,Ludwig and Probst (1998) investigated a set of
60 major world rivers.They suggested that present-day river sediment ¯uxes are mainly controlled by a
combination of several factors such as runo￿,slope of the drainage basin,seasonal variability of rainfall and
an index characterizing the softness of the lithology.
DEPOSITION OF SUSPENDED SEDIMENT
It has been shown that every year the Ganges and Brahmaputra together transport about 1037 million
tonnes of suspended sediment into Bangladesh.However,at the Mawa gauging station downstream,the
estimated suspended sediment load of the combined Ganges±Brahmaputra is only 525 million tonnes yr
ÿ1
.
The remaining 512 million tonnes (about 49%of the total suspended sediment load) can be assumed to have
been deposited within the area of the deltaic basin.However,some recent studies have led to speculation that
40±80%of the total suspended load might be stored landward of the river mouth (Milliman and Syvitski,
Table IV.Sediment yield of some major world rivers
River Drainage area
(10
3
km
2
)
Sediment yield
(t km
ÿ2
yr
ÿ1
)
River Drainage area
(10
3
km
2
)
Sediment yield
(t km
ÿ2
yr
ÿ1
)
Ganges 980 774* Amazon 5903 203
Brahmaputra 640 1126* Zaire 3704 13
Ganges±Brahmaputra 1620 895* Mississippi 3246 123
Ganges±Brahmaputra 1656 640 Nile 1874 64
Indus 912 274 Yensi 2567 5
Haunghe 823 1338
*Data after this study.Rest of the values are reported in Ludwig and Probst (1998).
Copyright
#
1999 John Wiley & Sons,Ltd.HYDROLOGICAL PROCESSES,VOL.13,2907±2923 (1999)
2918
M.R.ISLAM ET AL.
1992;Allison et al.,1998).The deposition of these huge amounts of sediment is causing the channel bed to
rise,with the formation of channel bars,and is o￿setting basin subsidence.
Many researchers have referred to subsidence in the Bengal Delta/Ganges±Brahmaputra Deltaic Basin
(e.g.Morgan and McIntire,1959;Alam,1996).Subsidence in the basin is generally considered to be of
tectonic origin and can be attributed to two major causes.One is related to isostatic adjustment of the crust
(deposition of sediment and uplift of the Himalayas) and the other is related to dewatering and compaction
of the shales and muds of the Proto-Bengal Fan (Alam,1989).There are no systematic quantitative data on
subsidence rates in the basin.However,radiocarbon dating of wood,peat and shale reported by Umitsu
(1993) suggests that the coastal areas of Bangladesh are subsiding at a rate of about 3 mmyr
ÿ1
.His study
indicates that,on average,the shelf area (Sirajganj,Rangpur) is subsiding at a rate of about 4 mmyr
ÿ1
.
Based on the average of these rates,we consider that the basin is subsiding at a rate of 3
.
5 mmyr
ÿ1
.
Every year a signi®cant part of the Bengal Basin is a￿ected by ¯oods due to over¯ow of the Ganges and
Brahmaputra rivers and their tributaries and distributaries.The record of annual ¯ood-a￿ected areas
prepared by BWDB from1954 to 1989 (disregarding the gap from1957 to 1959,and 1979±1981) shows that
the maximum¯ood-a￿ected area was 89 920 km
2
in 1988.The minimum¯ood-a￿ected area was 3140 km
2
in
1982.Our work suggests that,on average,30600 km
2
are a￿ected by ¯ood water.Each year,assuming that
sediment is deposited uniformly at the rate of subsidence (i.e.3
.
5 mmyr
ÿ1
) over an area equivalent to
30600 km
2
,a total of about 289 million tonnes yr
ÿ1
(considering an average rock density of 2700 kg m
ÿ3
) of
sediment will be deposited on the ¯oodplain during the period of ¯ood inundation.This constitutes about
28%of the total suspended sediment load of these two rivers.In using solid rock density as the volumetric
conversion factor,we are assuming a steady-state regolith thickness.it is noteworthy that based on
137
Cs
geochronology of sediment cores and geographic information system (GIS) extrapolation,Allison et al.
(1998) concluded that modern sediment is accumulating on the natural levees of the Brahmaputra at a rate of
about 4 cmyr
ÿ1
.This rate decreases landward to several mmyr
ÿ1
within a fewtens of km.We have assumed
that the remaining 223 million tonnes yr
ÿ1
is deposited on the river bed.This is discussed below.
Maps of Bangladesh show a wide range of variation in the width of both the Ganges and Brahmaputra
rivers.Careful analysis of the Map of Bangladesh (1:633 600,published by the Survey of Bangladesh)
indicates that the width of the Brahmaputra in Bangladesh varies from about 5 to 15 km.Goswami (1985)
assumed the average width of the Brahmaputra in Assamis 8
.
1 km.Map analysis also indicates that the width
of the Brahmaputra in Assamand Bangladesh is similar.An average width of 8
.
1 kmhas therefore been used
in this study.Subramanian (1978) reported the average width of the Ganges to be 5 km.However,map
observation indicates that this value is somewhat higher than the width of the Ganges in Bangladesh.Our
analysis suggests that a width of 3
.
5 km would be more representative.In the same way we assumed the
average width of the combined Ganges±Brahmaputra to be 5
.
5 km.We also assume frommap analysis that
the length of the Brahmaputra extending fromBahadurabad to Goalondo is about 175 km,whilst that of the
Ganges from Hardinge Bridge to Goalondo is about 100 km,and that of the combined Ganges±
Brahmaputra from Goalondo to Mawa is about 60 km.Assuming that the remaining 223 million tonnes of
sediment are deposited uniformly over the river bed,this amount of sediment would aggrade the river bed at a
rate of about 3
.
9 cmyr
ÿ1
.Goswami (1985) reported that in the Assam reach of the Brahmaputra (607 km
long),the river bed aggraded by 16 cm between 1971 and 1979 (i.e.1
.
77 cmyr
ÿ1
).
Remote sensing observations suggest that in the latter part of the monsoon season about 0
.
165 million
tonnes of sediment may be deposited in a single day (30 September 1991) along a 158 km long reach of the
Brahmaputra River,extending from Bahadurabad Ghat to Aricha (unpublished data).Deposition of this
amount of sediment would cause river bed aggradation of 0
.
048 mm in a single day.Extrapolation of this
rate to the whole year indicates aggradation of 1
.
7 cmyr
ÿ1
in this river reach.The same study also indicated
that a net loss of 0
.
17 million tonnes of sediment occurred along a 148 kmlong reach of the Ganges River in
Bangladesh,upstreamof Aricha at the same time.This sediment deposition would cause aggradation of the
Ganges river bed by 0
.
121 mmday
ÿ1
or 4
.
4 cmyr
ÿ1
(after extrapolation).This study leads to the conclusion
that river bed aggradation in the Ganges is much higher (about 2
.
5 times) than that in the Brahmaputra
Copyright
#
1999 John Wiley & Sons,Ltd.HYDROLOGICAL PROCESSES,VOL.13,2907±2923 (1999)
BASIN DENUNDATION AND SEDIMENTATION
2919
during this period.This higher rate can be attributed to the higher magnitude of anthropogenic in¯uences in
the Ganges River.Over the last two-and-a-half decades,a large amount of water has been withdrawn from
the Ganges River during the dry season,mainly for irrigation in India and Bangladesh.At present,dry
season ¯ows are less than 700 m
3
s
ÿ1
,while 1700 m
3
s
ÿ1
is required to maintain the regime which it
dominates.On the other hand,in the headwaters of this river,Nepal's forest has shrunk by 25%over the last
two decades (Khalil et al.,1995).
River bed aggradation reduces the sediment and water-carrying capacity of rivers,and causes river water
to over¯ow the banks more easily and frequently.This assumption is consistent with the suggestion that the
incidence of severe ¯oods has increased in recent times.Over the last three decades,navigable waterways
have shrunk alarmingly due to continued siltation and river bed aggradation.
As with the annual and seasonal variation of sediment discharge in these two rivers,sediment deposition
in the basin also ¯uctuates through the year.The variation of mean monthly sediment deposition during the
monsoon period for the years 1992 and 1993 is shown in Figure 11.The annual variation in sediment
deposition over the period from1989 to 1994 is illustrated in Figure 12.The loads for each year estimated at
the Mawa gauging station were subtracted from the sum of the loads of the year estimated at the Hardinge
Bridge and Bahadurabad gauging stations for the respective years (Figure 5).We assume that the remaining
load for each year is equivalent to the sediment deposition in that year.Figure 12 suggests that minimum
deposition occurred in 1991 and the maximum in the following year,1992.The reason for the signi®cant
variation in these two successive years is not clear.Detailed study is needed to explore this variation.
CONCLUSIONS
This work suggests that on average the Ganges transports 316 million tonnes of sediment into
Bangladesh each year,whilst the Brahmaputra carries more than twice the load (i.e.721 million tonnes)
of the Ganges River.There are several estimates of the sediment loads of these two rivers.Compared with
many other estimates,the proposed loads could be considered as conservative.Most of the average sediment
Figure 11.Monthly ¯uctuations in sediment deposition on the ¯oodplains and on the river beds of the Ganges and the Brahmaputra
rivers in the monsoon season
Copyright
#
1999 John Wiley & Sons,Ltd.HYDROLOGICAL PROCESSES,VOL.13,2907±2923 (1999)
2920
M.R.ISLAM ET AL.
loads documented in the literature range from316 to 520 million tonnes yr
ÿ1
for the Ganges River and from
617 to 800 million tonnes yr
ÿ1
for the Brahmaputra.The annual sediment load of the Ganges River ranges
from155 to 863 million tonnes over a period of 13 years.The sediment load of the Brahmaputra ranges from
455 to 992 million tonnes yr
ÿ1
over a period of 6 years.
Both rivers experience sharp increase in sediment discharge during the monsoon period (June to October).
However,the peak discharge occurs at di￿erent months of the same season.The Brahmaputra experiences its
lowest discharge during the period January to March.The Ganges reaches its lowest discharge during the
period February to May.Interannual and seasonal variations in water discharge are also seen in these rivers.
Maximumwater discharge occurs in the monsoon period.There is also interseasonal variation in the timing
of peak water discharge,which ranges from June to October,but the ¯uctuation of sediment and water
discharge does not always followthe same trend.The lack of a close relationship between water and sediment
discharge and between water discharge and SSC values indicates that their relationship might be in¯uenced
by both ¯ow hydraulics and sediment availability.
In this work,it is estimated that the mechanical denudation rate of the Ganges Basin is 303 mm,whilst
that of the Brahmaputra basin is 458 mm10
3
yr
ÿ1
,assuming a bulk density of 2700 kg m
ÿ3
.These values
stand at 818 mm and 1240 mm10
3
yr
ÿ1
,respectively,when we assume a bulk density of 1000 kg m
ÿ3
.The
average denudation rate for the whole Ganges±Brahmaputra Basin is 365 mm10
3
yr
ÿ1
.Compared with the
average denudation rate of 60 mm10
3
yr
ÿ1
for North America (Judson,1964) and a global average of
30 mm10
3
yr
ÿ1
,the denudation rate for the Ganges and Brahmaputra Basin is very high.This study also
suggests that denudation rates in these basins have not changed appreciably over the last 20 million years.
The mean annual speci®c suspended sediment yield for the Brahmaputra is 1126 tonnes km
ÿ2
yr
ÿ1
and that
of the Ganges is 774 tonnes km
ÿ2
yr
ÿ1
.
A combination of several factors appears to be responsible for these high denudation rates.The Ganges
and the Brahmaputra systems are characterized by very high channel gradients in their upper reaches and by
high relief ratios and high runo￿.Both rivers pass through one of the most tectonically unstable regions of
the world.Both rivers have a relatively large basin area.Basin lithology may also be contributing factor.The
lower and the middle courses of these two rivers pass through erodible rocks.In Bangladesh,particularly,
both the rivers ¯ow through loosely compacted recent alluvium (Coleman,1969).The rapid uplift of the
Figure 12.Interannual ¯uctuations in sediment deposition on the ¯oodplains and on the river beds of the Ganges and the
Brahmaputra rivers
Copyright
#
1999 John Wiley & Sons,Ltd.HYDROLOGICAL PROCESSES,VOL.13,2907±2923 (1999)
BASIN DENUNDATION AND SEDIMENTATION
2921
Himalayan Mountains (55 cm10
3
yr
ÿ1
;Sharma et al.,1978) must also contribute to the high denudation
rates in these drainage basins.
On average every year the Ganges and the Brahmaputra together carry 1037 million tonnes of sediment
into Bangladesh.Of this load,about 51% (525 million tonnes yr
ÿ1
) reaches the sea,28% (289 million
tonnes yr
ÿ1
) is deposited on land to balance the basin subsidence and the remaining 21%(223 million tonnes
yr
ÿ1
) is deposited on the river beds.Assuming that the sediment is deposited uniformly over the river bed,
this study suggests that the river beds of the Ganges and the Brahmaputra are aggrading by about
3
.
9 cmyr
ÿ1
on average.The aggradation rate on the Ganges bed is much higher (about 2
.
5 times) than that
for the Brahmaputra.
ACKNOWLEDGEMENTS
We would like to thank Bangladesh Water Development Board for their kind cooperation in this study.
Thanks are also due to Prof.Khalil Chowdhry,Department of Geological Sciences,Jahangirnagar
University,Dhaka,Bangladesh.We also thank the reviewers whose suggestions helped to improve this
manuscript.
REFERENCES
Abbas,N.and Subramanian,V.1984.`Erosion and sediment transportation in the Ganges River Basin (India)',J.Hydrol.,69,
173±182.
Alam,M.1989.`Geology and depositional history of Cenozoic sediments of the Bengal Basin of Bangladesh',Paleogeogr.Paleoecol.
Paleoclimat.,69,125±139
Alam,M.1996.`Subsidence of the Ganges±Brahmaputra Delta of Bangladesh and associated drainage,sedimentation and salinity
problems',in Milliman,D.J.and Bilal,V.H.(Eds),Sea-level Rise and Coastal Subsidence:Causes,Consequences and Strategies.
Kluwer Academic Publishers,The Netherlands.pp.169±192.
Allison,M.A.,Kuehl,S.A.,Martin,T.C.,and Hassan,A.1998.`Importance of ¯ood-plain sedimentation for river sediment budgets
and terrigenous input to the oceans:insights from the Brahmaputra-Jamuna River',J.Geol.,26,175±178.
BWDB (Bangladesh Water Development Board) 1972.`Sediment investigations in Main Rivers of Bangladesh,1968 & 1969'.BWDB
Water Supply Paper No.359.BWDB,Bangladesh.
Curray,J.R.and Moore,D.G.1971.`Growth of the Bengal Deep Sea Fan and denudation in the Himalayas',Geol.Soc.Am.Bull.,82,
563±572.
Curray,J.R.and Moore,D.G.1974.`Sedimentary and tectonic processes in the Bengal Deep Sea Fan and Geosyncline',in Burk,C.A.
and Drake,C.L.(Eds),Geology of Continental Margins.Springer,New York.617±627.
Coleman,J.M.1969.`Brahmaputra River:channel processes and sedimentation',Sediment.Geol.,3,129±139.
Einsele,G.1992.Sedimentary Basins.Springer-Verlag,Berlin.
Emmel,F.J.and Curray,J.R.1985.`Bengal Fan,Indian Ocean',in Bouma,A.H.,Normark,H.R.,and Bares,N.E.(Eds),Submarine
Fans and Related Turbidite Systems.Springer,New York.pp.107±112.
Emmett,B.E.1981.`Measurement of bed load in rivers,in erosion and sediment transport measurement',IAHS Publ.,133,3±15.
Goswami,D.C.1985.`Brahmaputra River,Assam,India:physiography,basin denudation,and channel aggredation',Wat.Resour.
Res.,21,959±978.
Holeman,J.N.1968.`The sediment yield of major rivers of the world',Wat.Resour.Res.,4,737±747.
Hossain,M.M.1992.`Total sediment load in the Lower Ganges and Jamuna',J.Inst.Engrs.Bangladesh,20,1±8.
Judson,H.D.and Ritter,D.F.1964.`Rates of regional denudation in the United States',J.Geophys.Res.,69,3395±3401.
Khalil,M.G.,Hossain,M.M.,and Hoque,M.M.1995.`On siltation in the rivers of Bangladesh:causes and consequences',J.Nat.
Ocean.Mart.Ins.,12,11±20.
Khan,F.H.1991.Geology of Bangladesh.The University Press Limited,Dhaka.
Ludwig,W.and Probst,J.L.1998.`River sediment discharge to the oceans:present day controls and global budgets',Am.J.Sci.,298,
265±295.
Meybeck,M.1976.`Total mineral dissolved transport by world major rivers',Hydrol.Sci.Bull.,21,265±284.
Milliman,J.D.and Meade,R.H.1983.`World-wide delivery of river sediment to the oceans',J.Geol.,91,1±21.
Milliman,J.D.and Syvitski,P.M.1992.`Geomorphic/tectonic control of sediment discharge to the ocean:the importance of small
mountaneous rivers',J.Geol.,100,524±544.
Morgan,J.P.and McIntire,W.G.1959.`Quaternary geology of the Bengal Basin,East Pakistan and India',Bull.Geol.Soc.Am.,70,
319±342.
NEDECO (Netherlands Engineering Consultants) 1967.East Pakistan Inland Water Transport Authority,Survey of Inland Waterways
and Ports,1963±1967.NEDECO,The Hague.
Rennell,J.J.1781.`An account of the Ganges and Brahmaputra Rivers',Phil.Trans.R.Soc.London,71,91±103.
Copyright
#
1999 John Wiley & Sons,Ltd.HYDROLOGICAL PROCESSES,VOL.13,2907±2923 (1999)
2922
M.R.ISLAM ET AL.
Sarin,M.M.and Krishnaswami,S.1984.`Major ion chemistry of the Ganges±Brahmaputra systems,India',Nature,312,538±541.
Sharma,K.K.,Saini,H.S.,and Nagpanl 1978.`Fission track aealing ages of apatites fromMandi Granite and application to tectonic
problems',Himalayan Geol.,8,298±312.
Singh,S.1988.`Nature of chemical and sediment load in the Ganges River between Sone and Koshi'.PhD Thesis.Jawaharal Nehru
University,New Delhi,India.
Subramanian,V.1978.`Input by Indian rivers into the world oceans',Proc.Indian.Acad.Sci.,87,97±104.
Subramanian,V.1987.`Environmental geochemistry of Indian river basins Ða review',J.Geol.Soc.India,29,205±220.
Subramanian,V.and Ramanathan,A.L.1996.`Nature of sediment load in the Ganges±Brahmaputra River system in India',in
Milliman,D.J.and Bilal,V.H.(Eds),Sea-Level Rise and Coastal Subsidence:Causes,Consequences and Strategies.Kluwer
Academic Publishers,The Netherlands.pp.151±168.
Summer®eld,M.A.and Hulton,N.J.1994.`Natural controls of ¯uvial denudation in major world basins',J.Geophys.Res.,99,
13871±13883.
Umitsu,M.O.1993.`Late quaternary sedimentary environments and landforms in Ganges Delta',Sediment Geol.,83,117±186.
Walling,D.E.and Webb,B.W.1981.`The reliability of suspended sediment load data',IAHS Publ.,133,177±194.
Walling,D.E.and Webb,B.W.1988.`The reliability of rating curve estimates of suspended sediment yield:some farther comments',
IAHS Publ.,174,337±350.
Walling,D.E.and Webb,B.W.1996.`Erosion and sediment yield:a global overview',IAHS Publ.,236,3±19.
Copyright
#
1999 John Wiley & Sons,Ltd.HYDROLOGICAL PROCESSES,VOL.13,2907±2923 (1999)
BASIN DENUNDATION AND SEDIMENTATION
2923