See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/279846589

Analysis of suspended sediment transport data in the River Nile

Conference Paper · September 2014

DOI: 10.1201/b17133-129

CITATIONS READS

2 1,221

4 authors, including:

Bahaeldeen Zaid Katinka Koll

Bundesanstalt für Wasserbau Technische Universität Braunschweig
15 PUBLICATIONS   23 CITATIONS    70 PUBLICATIONS   991 CITATIONS   

SEE PROFILE SEE PROFILE

Jens-Uwe Wiesemann
Technische Universität Darmstadt
6 PUBLICATIONS   4 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Entwicklung eines modularen Wärmetauschers mit kombinierter Wasserkraftgewinnung für Fließgewässer (WärmeEnergieTauscher) View project

Stauzielerhöhung eines dynamischen Wehrständers (Federwehr) View project

All content following this page was uploaded by Bahaeldeen Zaid on 02 June 2016.

The user has requested enhancement of the downloaded file.

 River Flow 2014 – Schleiss et al. (Eds)
© 2014 Taylor & Francis Group, London, ISBN 978-1-138-02674-2

Analysis of suspended sediment transport data in the River Nile

B.A. Zaid & Ka. Koll

Leichtweiß-Institut für Wasserbau, TU Braunschweig, Braunschweig, Germany

J.-U. Wiesemann
Lahmeyer International GmbH, Bad Vilbel, Germany

A.S. Elzein
Dams Implementation Unit, Ministry of Water Resources and Electricity, Khartoum, Sudan

ABSTRACT:  The River Nile is the backbone for the water resources management within Sudan. The
heavy sediment load of the river imposes challenges for hydropower and irrigation. A comprehensive sedi-
ment measurement program was initiated in 2004 to develop the design and operation of eight planned
dams and one existing dam. Analysis of the first data set from 2005 to 2008 are presented. Suspended
sediment measured at the station El Koro composes mainly of silt (60%), clay (34%) and a small portion
of fine sand (6%). The discharge varied between 454 and 11,335 m3/s, and the sediment concentration
from 0.004 to 14.6 g/l. The relation between concentration and discharge shows a dominant clockwise
hysteresis for the entire flood season. However, counterclockwise loops have been noted for short periods.
By dividing the flood to rising limb, falling limb and base flow, empirical relations were developed to
account for the hysteresis effect.

Keywords:  suspended sediment; hysteresis; discharge-concentration relation; field data; River Nile

1  introduction

The River Nile carries a large amount of suspended

sediment load. Shalash (1982) analysed long-term
measurements located next to the Aswan High Dam
carried out in the years 1929 to 1979. He reported a
suspended sediment concentration of up to 3.425 g/l
averaged over 10 days in the period 1929 to 1955.
About 95% of the total sediment load is com-
ing from the Ethiopian Plateau through the Blue
Nile and Atbara River (Abdelsalam 2008). The soil
erosion and land degradation in Ethiopia and the
reservoir sedimentation in Ethiopia, Sudan and
Egypt are among the main topics associated with
sediment load in the River Nile.
The first Sudanese dam project out of nine pro-
posed projects started in 2001 with construction of
Merowe Dam (see Fig. 1) which went into opera-
tion in 2008. As part of the requirements for the
implementation and the studies of hydropower Figure 1.  River Nile basin.
projects, and the reservoir sedimentation studies
associated with it, a sediment sampling and analy-
sis program has been initiated. Shellberg et al. 2013, Guzman et al. 2013). Flood
It is well known that the suspended load of a wave propagation affected by river course, chan-
river is dominated by characteristics of the catch- nel geometry, tributaries, etc. further influences
ment due to e.g. land-use and rainfall duration and the suspended sediment concentration, resulting in
distribution (e.g. Seeger et al. 2004, Baĉa 2008, hysteresis between concentration and discharge.

961
 Williams (1989) systematically explored and
grouped the hysteresis loops in five common
classes; single valued, clockwise, counterclockwise,
single valued plus a loop, and eight-shaped. In
general, clockwise hysteresis with higher sediment
concentrations during the rising stage are reported
(e.g. Williams 1989, Ahanger et al. 2008, Baĉa
2008, Wang & Linker 2008).
Ahanger et al. (2008) showed that even for data
obtained in laboratory experiments it is a chal-
lenging task to relate suspended load to water
­discharge. Nevertheless, functional relationships
are required in order to make long-term forecasts Figure 3. Data set at station El Koro (flood periods are
for dams operation, like flushing. For this purpose indicated by the arrows).
a new key station has been established at El Koro
(see Fig.  1) in 2004 where water level, discharge
and suspended sediment concentration are meas- water discharge, suspended sediment transport, and
ured on a regular basis. bed material started in August 2005.
The purpose of this paper is to introduce the The discharge was determined via measurements
station El Koro and the data set measured in the of the flow velocity with a Price AA current meter
years 2005 to 2008. First results of discharge and in 2005. Since 2006 discharge measurements have
suspended sediment concentration analysis are been done with a Workhorse Rio Grande ADCP.
presented. Depth-integrated measurements of the sus-
pended load were carried out with an US P-61 sam-
pler at four verticals within the cross-section. The
2 Data base positions of the verticals were determined by sub-
dividing the cross-section in increments of equal
2.1  Station El Koro discharge. At the same verticals bed material was
The measuring station El Koro was installed in sampled with an US BM-54. The positioning of
2004. El Koro is located about 35 km upstream of the boat, which carried the measuring equipment,
the backwater curve of Merowe Dam and 225 km was determined with a laser range finder referenc-
downstream of the confluence with Atbara River, ing to fixed poles with reflectors on each bank.
the last tributary of the Nile. Thus, the station
allows to collect samples of almost the total sedi- 2.2  Water level and discharge
ment load of the River Nile.
The rectangular cross-section of the station is At El Koro simultaneous measurements of water
ca. 550 m wide and ca. 10 m deep (see Fig. 2). It is level and discharge are available for the period from
located in the last quarter of an 8 km long almost August 2005 to November 2008 (see Fig.  3). The
straight reach of the River Nile. relation between discharge and water level follows
An overview of the data analysed in this paper is a polynomial function with high accuracy, indi-
given in Figure 3. The station is equipped with a staff cated by a coefficient of regression of 0.998. This
gauge to determine the water level. Since May 2004 allows to calculate the discharge for the days when
daily gauge readings are available. Measurements of only the water level was recorded. The hydrographs
of discharge and water level for the years 2005 to
2008 are shown in Figure 4.
During the low flow season from November
to June the average discharge is about 1000 m3/s.
The rain fall events in the catchment areas of the
two main sources White Nile on Equatorial Lakes
­Plateau and Blue Nile on Ethiopian Plateau lead
to a rapid increase of the discharge which peaks on
average at 10,500 m3/s in August.

2.3  Bed material

The bed of the River Nile mainly consists of
fine to medium sand with characteristic diam-
Figure 2.  Cross-section at station El Koro. eters d50 = 0.2 mm, d90 = 0.34 mm, and maximum

962
Figure 4.  Hydrographs of discharge and water level. Figure  6.  Characteristic diameters of suspended sedi-
ment as a function of the discharge.

Figure 5.  Characteristic diameters of bed material as a

Figure 7.  Suspended sediment concentration as a func-
function of the discharge.
tion of discharge measured in 2005 to 2008.

g­ rain-size d100 = 1 mm, on average. A lateral sort- suspended sediment concentration on the cross-
ing of the grain-size distribution of the bed sectional position was determined. Thus, the cross-
material within the cross-section is not evident. sectional averages are calculated and analysed in the
Plotting characteristic diameters as a function of following. The concentration is plotted as a func-
the discharge (Fig. 5) reveals that the bed material tion of the discharge in Figure 7. It increases lin-
becomes finer with increasing discharge. early with the discharge up to a value of 1600 m3/s.
For higher discharges the data scatter remarkably.
2.4  Grain-size distribution of suspended sediment
The grain-size distribution of the suspended sedi- 3 suspendeD sediment dynamics
ment was determined for selected samples in the
years 2005, 2007, and 2008 (see Fig.  3). The sam- The data, presented in Figure 7, scatter distinctly
ples were taken at discharges between 4000 and less when the date of sampling is considered in the
11,000 m3/s. Analysis of the data revealed no depend- rating curve. Figure  8 shows the rating curves of
ency of the grain-size of suspended sediment on the suspended sediment concentration sorted accord-
position within the cross-section. Moreover, the ing to the class of hysteresis as suggested by
characteristic diameters of the suspended sediment ­Williams (1989).
show no relation with neither the discharge, shown In general, the hysteresis have clockwise trends
in Figure  6, nor the magnitude of the suspended with higher sediment concentrations during the
sediment concentration. On average the suspended rising stage, which is typical for large rivers accord-
sediment consists of 34% clay, 60% silt, and 6% fine ing to Baĉa (2008). However, in the years 2005 and
sand with a maximum diameter of 0.09 mm. 2007 counterclockwise loops occurred during the
rise of the flood (Fig. 8a), which was not observed
in the years 2006 and 2008 (Fig. 8b).
2.5  Suspended sediment concentration
The big loops in 2005 and 2007 correspond to sud-
As for the grain-size distribution of bed material den drops in the discharge. The discharge decreased
and suspended sediment no dependency of the in 2005 within three days by about 2200 m3/s, and

963
 Thus, the most likely reason for the counter-
clockwise loops within the clockwise hysteresis
is the first one given by Williams (1989), i.e. that
changes in discharge travel with wave velocity but
the suspended sediment travel with a velocity close
to mean flow velocity, which is slower.
Due to the described hysteresis of the suspended
sediment concentration and the discharge a unique
relationship for a rating curve cannot be found. To
account for the hysteresis and to improve the corre-
lation, first, it was distinguished between dry season
and flood season, and second, the flood was divided
into rising and falling limbs. Thus, the rating curve
could be subdivided into three parts with different
functional relationships according to the discharge.
Figure  9 shows the data, already presented in
Figure  7, sorted according to the three divisions
in a semi-logarithmic plot. A linear relation for a
discharge less than 1600 m3/s had been calculated.
In the rising limb of the hydrograph, a logarithmic
function was fitted, and an exponential trend for
the falling limb of the hydrograph until a discharge
of 1600 m3/s was reached again.
Figure 8.  Rating curve of suspended sediment concen- In the years 2005 to 2008 the concentration of
tration considering time, a) for the years 2005 and 2007, suspended sediment always peaked at a discharge of
b) for the years 2006 and 2008. about 9000 m3/s, i.e. before the maximum discharge
was reached. Thus, the boundary value between the
rising and the falling limb was set to 9000 m3/s.
in 2007 within five days by about 1700 m3/s. In both The base flow below 1600 m3/s is well presented
years the absolute maximum of suspended sediment by the linear trend and also the assumption of
concentration was measured at the first day of the an exponential function for the falling limb is
decreased discharge. With a time lag of one day quite satisfying. However, it is obvious that the
also the sediment concentrations decreased. Also function for the rising limb needs improvement.
smaller counterclockwise loops could be related to Especially, the highest sediment concentrations
steep negative gradients of the discharge. are not well considered by the logarithmic trend.
Williams (1989) gave three possible reasons It is not expected that the values are outliers, but
for counterclockwise hysteresis: the relative travel measuring concentrations larger than 10 g/l only
times of flood wave and sediment flux, high soil four times within four years (three floods) does
erodibility combined with prolonged erosion, and not justify fitting to the extreme values. ­Moreover,
seasonal variability of rainfall distribution and of the large scatter of the sediment concentration
sediment production in the catchment. between 4 and 10  g/l at discharges between 6000
The main source of suspended sediment is
definitely soil erosion in the Ethiopian highlands.
However, information about rain fall intensity
and areal distribution were not available for this
study, and thus cannot be discussed. Moreover,
the ­second and third reason given by Williams
(1989) are more related to the overall behaviour of
the hysteresis and not to short-term effects. Local
sediment sources like bank failures, as observed in
the field after abrupt water level drops, or reservoir
flushing could be assumed to contribute to the sed-
iment concentration. However, dam operations did
not temporally correlate with changes in the con-
centration trends, and the amount of suspended
sediment delivered by bank failures seem to vanish Figure 9.  Rating curve of suspended sediment concen-
compared to the sediment concentration caused by tration and trends sorted according to base flow, rising
soil erosion in the catchment. limb, and falling limb (open symbols are outliers).

964
and 9000 m3/s requires further ­consideration.
­Discussing Figure  8 it has been shown that the
gradient of the discharge plays an important role.
Thus, it is expected that implementing the gra-
dient of the discharge as suggested by Wang &
Linker (2008) will lead to a better estimation of
the suspended sediment concentration.
Figure  10 shows the calculated and measured
hydrographs of suspended sediment concentration
together with the hydrograph of the discharge.

Figure  11.  Comparison of measured and calculated

suspended sediment concentration for the years 2005 to
2008 (diamonds) and 2013 (crosses).

The plots show a reasonable agreement between

the calculated and the measured concentrations but
it becomes obvious that changes of the discharge
within both, the rising and the falling limb are not
reflected well by the empirical fits, yet (e.g. falling
limb in Fig. 10c, and rising limb in Fig. 10d).
Figure  11 presents a comparison of measured
and calculated sediment concentrations for the
data measured in 2005 to 2008 used to fit the
empirical trends. The line of perfect agreement as
well as the ± 25% lines are also plotted.
Sampling did not stop in 2008 but continued
until today. In order to check the quality of the
empirical functions the data measured in 2013
were exemplarily chosen as independent data set.
Measured and calculated sediment concentrations
are included in Figure 11.
As already stated, the empirical equations need
to be improved. However, applying the equations to
an independent data set shows that the data from
2013 do not deviate systematically from the data
used to fit the equations but integrate. One may
conclude that the quality of the fit is not good, but
it has to be taken into account that even formulas
derived from data obtained in laboratory experi-
ments show similar match or scatter, respectively,
between measured and calculated suspended sedi-
ment concentration (Ahanger et al. 2008).

4  Conclusion

The sampling station El Koro located on the main

River Nile just downstream of the last tributary
has been introduced and suspended sediment data
measured during the first four years were analysed.
The sampled bed material mainly consisted of
fine to medium sand. Sorting within the cross-
Figure  10.  Hydrographs of discharge, and measured ­section was not observed, reflecting the rectan-
and calculated suspended sediment concentration in the gular shape of the cross-section and the straight
years a) 2005, b) 2006, c) 2007, and d) 2008. course of the upstream river reach.

965
 Suspended sediment consisted of 34% clay, 60% s­ econd, by including the data measured from 2009
silt, and 6% fine sand, on average, with a maximum until today.
diameter of 0.09  mm. Characteristic diameters
were independent of the sampling position within
the cross-section, the discharge throughout a flood References
and of the magnitude of the suspended sediment
concentration. Abdelsalam A. 2008. Sediment in the Nile River System.
The sediment concentration has been analysed Khartoum, UNESCO-IHP-International Sediment
on large time scale. Within the considered period Initiative.
the discharge varied between 454 and 11,335 m3/s, Ahanger, M.A., Asawa, G.L. & Lone, M.A. 2008.
and the suspended load ranged from 0.004 to 14.6 ­Experimental study of sediment transport hysteresis.
g/l. The sediment concentration was higher during Journal of Hydraulic Research, 46(5): 628–635.
Bača, P. 2008. Hysteresis effect in suspended sediment
the rise of the flood and peaked prior to the water
concentration in the Rybárik basin, Slovakia. Hydrol.
discharge, i.e. a general clockwise hysteresis has Sciences, 53(1): 224–235.
been observed for the entire flood season, which Guzman, C.D., Tilahun, S.A., Zegeye, A.D. &
reflects the dominant impact of the watershed in Steenhuis, T.S. 2013. Suspended sediment concen-
the Ethopian Plateau. However, more or less pro- tration-discharge relationships in the (sub-) humid
nounced counterclockwise loops on short periods Ethiopian highlands. Hydrology and Earth System
were observed, which were related to the relative Sciences, 17: 1067–1077.
travel times of flood wave and sediment flux. Seeger, M., Errea, M.-P., Beguería, S., Arnáez, J., Mart, C.,
Due to the hysteresis in the concentration- García-Ruiz, J.M. 2004. Catchment soil moisture and
rainfall characteristics as determinant factors for dis-
discharge relation the period of a year has been
charge/suspended sediment hysteretic loops in a small
divided into base flow and flood. The flood is fur- headwater catchment in the Spanish Pyrenees. Journal
ther divided into rising and falling limb. Thus, three of Hydrology, 288: 299–311.
trends were fitted to the data: a linear relation for Shalash, S. 1982. Effects of sedimentation on the stor-
the base flow, a logarithmic function for the rising age capacity of the High Aswan Dam reservoir.
limb of the hydrograph, and an exponential trend ­Hydrobiologia, 92: 623–639.
for the falling limb until the base flow is reached Shellberg, J.G., Brooks, P.A. & Rose C.W. 2013. Sediment
again. production and yield from an alluvial gully in north-
Measured and calculated suspended sediment ern Queensland, Australia. Earth Surface Process and
Landform, DOI: 10.1002/esp.3414.
concentrations are in good agreement, consider-
Wang, P. & Linker, L.C. 2008. Improvement of regres-
ing the fact that sediment transport is scattering. sion simulation in fluvial sediment loads. Journal of
­However, the empirical relations can be improved, Hydraulic Engineering, 134(10): 1527–1531.
first, by taking into account the observed influ- Williams, G.P. 1989. Sediment concentration versus water
ence of the gradient of the discharge (presum- discharge during single hydrologic events in ­rivers.
ably responsible for counterclockwise loops), and, Journal of Hydrology, 111: 89–106.

966

View publication stats