Conference PaperPDF Available

Reservoir sedimentation analysis of the proposed Noordoewer/Vioolsdrift dam on the Orange river and evaluation of sediment control mitigation measures

Authors:

Abstract and Figures

The Orange River Re-Planning Study (ORRS, 1997) recommended that the proposed Noordoewer/Vioolsdrift Dam (NVD) site be further investigated as possible dam development downstream of Vanderkloof Dam, in the lower reach of the 2200 km long Orange-Senqu River. The Lower Orange River Management Study (LORMS, 2005) recommended that a re-regulating dam be constructed at the NVD site on the border between Namibia and South Africa in order to increase the availability of water to meet both the future human and ecological water requirements for the Lower Orange River and the river estuary. This paper addresses the NVD feasibility study findings based on the joint study by Namibia and South Africa related to sedimentation. The sedimentation investigation consists of two phases i.e. (1) the determination of the sediment yield at the proposed NVD site and (2) the reservoir sedimentation hydrodynamic modelling, including the flood level simulations due to sedimentation and the feasibility of flushing sediment from the reservoir. From the sediment yield analysis, the proposed long-term sediment load at the NVD site is 16.4 million t/a. Based on the 2D hydrodynamic modelling, the proposed NVD reservoir would have 47% and 77% of the original storage capacity after 100 years of operation, for the 70 m and 90 m high dam scenarios respectively. The reservoir traps 97 % of the sediment load. Reservoir sedimentation mitigation measures were investigated to extend the life of the dam, based on ICOLD (1999) guidelines. Simulations were carried out to evaluate reservoir drawdown flushing during floods and local pressure flushing at the dam outlet for irrigation. Based on the expected relatively large sustainable equilibrium FSC of 70 % of the original FSC for the NVD project, it is proposed that the 70 m high dam (or a lower dam), with a smaller dead storage for sedimentation, is designed with drawdown flushing during floods.
Content may be subject to copyright.
Reservoir sedimentation analysis of the proposed
Noordoewer/Vioolsdrift dam on the Orange river and
evaluation of sediment control mitigation measures
Prof G.R. Basson, Dr O. Sawadogo and J.K. Vonkeman
Department of Civil Engineering
Stellenbosch University
Stellenbosch 7600
South Africa
The Orange River Re-Planning Study (ORRS, 1997) recommended that the proposed
Noordoewer/Vioolsdrift Dam (NVD) site be further investigated as possible dam development
downstream of Vanderkloof Dam, in the lower reach of the 2200 km long Orange-Senqu River. The
Lower Orange River Management Study (LORMS, 2005) recommended that a re-regulating dam be
constructed at the NVD site on the border between Namibia and South Africa in order to increase the
availability of water to meet both the future human and ecological water requirements for the Lower
Orange River and the river estuary. This paper addresses the NVD feasibility study findings based on the
joint study by Namibia and South Africa related to sedimentation.
The sedimentation investigation consists of two phases i.e. (1) the determination of the sediment yield at
the proposed NVD site and (2) the reservoir sedimentation hydrodynamic modelling, including the flood
level simulations due to sedimentation and the feasibility of flushing sediment from the reservoir. From
the sediment yield analysis, the proposed long-term sediment load at the NVD site is 16.4 million t/a.
Based on the 2D hydrodynamic modelling, the proposed NVD reservoir would have 47% and 77% of the
original storage capacity after 100 years of operation, for the 70 m and 90 m high dam scenarios
respectively. The reservoir traps 97 % of the sediment load.
Reservoir sedimentation mitigation measures were investigated to extend the life of the dam, based on
ICOLD (1999) guidelines. Simulations were carried out to evaluate reservoir drawdown flushing during
floods and local pressure flushing at the dam outlet for irrigation. Based on the expected relatively large
sustainable equilibrium FSC of 70 % of the original FSC for the NVD project, it is proposed that the 70 m
high dam (or a lower dam), with a smaller dead storage for sedimentation, is designed with drawdown
flushing during floods.
1. Background
The feasibility study for the proposed Noordoewer/Vioolsdrift Dam required further investigation
addressing the reservoir sedimentation because the Orange-Senqu River basin is one of the most
significant river basins in southern Africa. It is shared among four countries, namely Lesotho, South
Africa, Botswana and Namibia. Fig. 1 shows the schematic representation of the annual sediment loads (in
red) for the Orange/Senqu River and its tributaries (in purple) relative to the proposed NVD site. The total
average sediment outflows from catchments areas within the Orange-Senqu River basin were computed at
different gauging stations and reservoirs along the river, based on the South African Water Research
Commission (WRC, 2012) methodology.
The impact of reservoir sedimentation is significant in the Orange-Senqu River basin. Some reservoirs
have lost substantial volumes of their original storage capacity due to sedimentation. The loss in storage
directly reduces the available yield of the reservoir, resulting in the possible under supply to relevant
water users. Sedimentation could result in serious socio-economic losses, environmental and aesthetic
problems, considering the dependence on water stored in reservoirs for potable, irrigation, recreation,
hydropower production and flood control purposes. Specifically, food production from irrigated
agriculture along the Orange River can be affected by reduced water storage volumes in reservoirs.
Fig. 1. S
c
2. Sed
Table 1
s
p
ropose
d
sedimen
t
of 11.2
m
p
revious
Departm
data of t
h
in the ca
t
p
ossible
Upin
gt
(millio
n
1)
R
-
-
2)
T
25.
5
N
o
Vio
c
hematic repr
e
iment
y
iel
d
s
ummarizes
t
d
that the hig
h
t
ation of the
p
m
illion t/a to
studies. The
ent of Water
h
e //Khara H
a
t
chment. Th
e
but limited
fu
t
on
n
t/a)
Vio
(mi
l
R
eviewin
g
p
Me
a
Hi
g
T
he re
g
iona
l
5
2
P
roposed
o
ordoewer /
olsdrif Dam
Site
e
sentation of t
r
d
t
he long ter
m
h
long term
s
p
roposed N
V
compensate
f
re is confide
n
and Sanitati
o
a
is Local M
u
e
proposed s
e
fu
ture land d
e
Tab
l
olsdrift
l
lion t/a)
revious stu
d
a
n: 16.6
T
c
1
i
n
g
h: 25.0
l
methodolo
g
2
9.9
T
m
S
V
IOOLSDRIFT
r
ibutaries and
on the W
R
sediment lo
a
s
ediment loa
d
V
D Dam at V
i
f
or limited l
a
n
ce in this v
a
o
n’s (DWS)
T
u
nicipality at
e
diment load
e
gradation an
l
e 1. Long ter
d
ies of the O
R
T
hese studies
oncentration
970 and 197
5
n
the Lower
O
gy
develope
d
T
hese relativ
e
m
ethod was c
S
ubsequent t
o
UPINGTON
sediment loa
d
R
C (2012) me
t
a
ds obtained
d
of 16.4 mil
l
i
oolsdrift. T
h
a
rge flood T
S
a
lue because
T
otal Suspe
n
Upington an
of this study
n
d climate ch
a
r
m sediment l
o
R
RS (1997)
were based
o
data. The in
d
5 had a signi
O
range Rive
r
d
b
y
the W
R
e
ly high load
s
alibrated wit
o
this period,
d
s on the Oran
g
t
hodology)
from differe
n
l
ion t/a is us
e
h
is is 46 %
m
S
S data but is
it was calibr
a
n
ded Solids (
T
d observed p
is conservat
i
a
nge.
o
ads at Viool
s
Comm
e
and LORM
S
o
n pre-dam
d
d
ication is t
h
ficant impac
t
r
since 1976.
R
C (2012) fo
r
s
are attribut
e
h pre-dam hi
additional s
e
g
e-Senqu Rive
n
t data sets a
n
e
d to evaluat
e
ore than the
m
similar to th
a
ted against
S
T
SS) data at
U
ost-1976 res
e
i
vely high an
s
drift
e
nts
S
(2005) b
y
t
d
evelopment
s
h
at the dams
c
t
in decreasi
n
r
un
g
au
g
ed
c
e
d to the fact
storical sedi
m
e
diment yiel
d
e
r catchment a
r
n
d assumpti
o
e
the reservo
i
m
ean long t
e
h
e mean annu
S
outh Africa
U
pington, tu
r
e
rvoir sedim
e
n
d should cat
e
t
he DWS
sediment
c
ompleted b
e
n
g the sedim
e
c
atchments
t
that the W
R
m
ent (1929 t
o
d
data was o
b
r
ea (based
o
ns. It is
i
r
rm load
a
l load of
s
r
bidity
e
ntation
e
r for
e
tween
e
nt load
R
C
o
1969).
tained
from reservoir sediment deposition data. However, not much data was
available in the Lower Orange River in Namibia and Botswana for
calibration due to the nature of the ephemeral rivers. The sediment loads
for these regions were based on a comparative analysis of erosion hazard
classes of the adjacent river catchments located in South Africa.
3) Suspended sediment concentration data sampled by SA’s Department of Water & Sanitation
4.7 9.1
The TSS concentration data collected by the DWS at the flow gauging
stations of Upington (D7H005) and Prieska (D7H002) were used to
calculate sediment loads, found to be similar and combined to give a
larger database for the post-dam period 1976 to 2015.
The results were scaled up from 23.5 million t/a to obtain the same 43.8
million t/a sediment load for the 20 year period 1949 to 1969, based on
published data for Upington and Prieska (Rooseboom, 1992). The post-
dam data was also scaled up at Upington, due to limited data being
available for this period, to rather represent an upper envelope for the
observed sediment loads at river flows above 1000 m3/s.
If the pre-dam sediment rating was applied to the period from 1976 to
2015, a sediment load of 42.4 million t/a would theoretically have been
observed. The difference in the pre- and post-1976 sediment loads is 37.7
million t/a (or 1507 million t over a 40 year period). The reduction in
sediment load can be accounted for by the 1547 million t sediment
deposition observed in the Orange-Senqu River basin reservoirs (based
on reservoir survey data from the DWS dam list of 2015).
6.8 11.2
The flow record from 1976 to 2015 used in the TSS analysis does not
have large floods and therefore a 50 and 100 year flood was added to the
flow record.
4) Turbidity data recorded by the //Khara Hais Local Municipality at Upington
High: 12.0 High: 16.4
The recent TSS dataset from DWS is relatively small and underestimates
sediment loads for larger flows above 2000 m3/s so it was supplemented
with turbidity data sampled every 2 hours at the raw water intake of a
local water treatment works from 2000 to 2015. The corresponding TSS
data was used to convert the turbudity data to sediment concentrations
and loads. There are uncertainties in the long term sediment yield due to
the scatter in the data and the limited data points above 2000 m3/s.
Therefore, it was decided to also consider a high sediment load
relationship. It was based on the reliable TSS data for low discharges and
on the //Khara Hais Local Municipality data for larger discharges.
Note that the suspended sediment concentrations and loads were
increased by 25% to allow for bed load and non-uniformity in suspended
sediment concentrations across the river.
3. Hydrodynamic modelling
The long-term reservoir sedimentation after 50 and 100 years of the proposed NVD reservoir was
simulated by the 2D hydrodynamic model Mike21C. Two different dam height scenarios were evaluated,
namely a 70 m high dam with a Full Supply Level (FSL) of 230 masl and a 90 m high dam with a FSL of
250 masl. The reservoir bathymetry over a 125 km reach was obtained from LiDAR survey data. The
curvilinear grid was set up with cell sizes of 80 m wide and 240 m long in the flow direction. Typical
hydraulic roughness Manning n values were used: 0.025 for the reservoir, 0.035 for the main channel and
0.050 for the river floodplain. The upstream flow boundary was based on the 100 year scaled daily inflow
record of the DWS gauging station DH003 while the downstream reservoir levels were based on the mass
balance of the dam considering daily rainfall and evaporation, monthly demands abstracted and spillage.
The upst
b
ased o
n
0.033 m
m
cohesive
0.182 m
m
storage
o
3.1
R
Fig. 2 s
h
and 90
m
causes t
h
b
ased o
n
Fi
g
Table
2
Volum
e
Perce
n
Design
fo
(minimis
ream bound
a
n
the calibrat
e
m
was deter
m
sediment fr
a
m
. The long
t
o
perated rese
r
R
eservoir se
d
h
ows longitu
d
m
high dams.
h
e sediment t
o
n
the high se
d
g
. 2. Sediment
2
. Simulated
s
O
T
e
of sedimen
t
FSC rem
a
n
ta
g
e of the
o
fo
r sedimenta
ation of spil
l
160
170
180
190
200
210
220
230
240
0
Elevation (masl)
160
170
180
190
200
210
220
230
240
250
260
0
Elevation (masl)
a
ry of the mo
d
e
d sediment l
m
ined from a
a
ctions were
t
erm deposit
e
r
voir in Sout
h
d
imentation
w
d
inal section
s
The 70 m hi
g
o
be deposit
e
d
iment load o
deposition af
t
s
ediment vol
u
Dam FS
L
O
ri
g
inal FS
C
T
ime period
t
in the rese
r
a
inin
g
after
o
ri
g
inal FS
C
tion and dea
d
l
age) reservo
10000 20000
10000 20000
d
el also incl
u
oad-dischar
g
grading anal
y
a
lso implem
e
e
d sediment
d
h
Africa.
w
ithout flus
h
s
of the simul
g
h dam has
m
e
d much furt
h
f 16.4 millio
n
t
er 50 and 100
u
mes in the re
s
L
(masl)
(million m3
)
in years (X)
r
voir after
X
X-
y
ears (mi
C
remainin
g
d
storage wa
s
ir is typicall
y
30000 40000
Initial Bed level (masl)
30000 40000 5
0
Initial Bed level (masl)
u
ded a time s
g
e rating at
U
ysis of sedi
m
e
nted: 30% o
d
ensity was s
p
h
in
g
l
ated bed lev
e
m
uch lower o
p
h
er into the r
e
n
t/a are indi
0
years for (a)
servoir after
5
flushing
)
)
X
-
y
ears (mill
i
i
llion m3)
after X
y
ea
r
s
b
ased on I
C
y
sized to ac
c
50000 60000
Distance (m)
230 FSL Dam
Bed_level (after 50yr)
0
000 60000 7000
0
Distance (m)
250 FSL Dam
Bed_level (after 50yr)
FSL = 230
FSL = 25
0
eries of cohe
pington. Th
e
m
ent samples
f 2.751 mm,
p
ecified as 1
e
l change aft
e
p
erating lev
e
e
servoir. Th
e
c
ated in Tab
l
the 70 m high
5
0 and 100 y
e
i
on m3)
5
1
r
s (%)
C
OLD (2009)
c
ommodate t
h
70000 80000
9
Bed_level (after 100
y
0
80000 90000
Bed_level (after 1
0
masl
0
masl
sive sedime
n
cohesive se
d
collected fro
30% of 0.53
1
.35 t/m3 whi
c
e
r 50 and 10
0
e
ls than the 9
0
deposited s
e
l
e 2.
dam and (b) t
h
e
ars of operat
i
230
2240
50 10
0
5
89 117
651 106
74 47
: “
A
storage
h
e expected
5
9
0000 100000 11
0
y
r)
100000 110000 12
0
0
0yr)
n
t concentrat
i
d
iment size
o
o
m the field.
N
1 mm and 40
c
h is typical
f
0
years for t
h
0
m dam an
d
e
diment volu
m
h
e 90 m high
d
ion without s
e
2
5
0
0
50
8 589
2 4441
88
operated
5
0-year sedi
m
0
000 120000
0
000 130000
i
on,
o
f
N
on
% of
f
or a
h
e 70 m
this
m
es
d
am
e
diment
2
50
0
30
100
1178
3852
77
m
ent
volume
a
yield an
a
water fi
r
is only t
h
storage
c
the 230
m
in Afric
a
for in th
e
3.2
R
Sedimen
flushing
sedimen
t
required
.
reduce t
h
flushing
ICOLD
B
the two
r
Table 3
g
storage
o
sedimen
t
modelli
n
3.2.1.
L
Pressure
sedimen
t
dam is f
u
b
ed elev
a
m water
volume
o
3 hours
w
2-year fl
a
llowing for
t
a
lysis, which
r
m yield.” Bo
h
e colloidal
f
c
apacities re
m
m
asl FSL da
m
a
of 0.6 % pe
r
e
reservoir: 1
R
eservoir se
d
t can be rem
o
during flood
t
over the le
n
.
Flushing w
i
h
e initial req
u
since the va
l
B
ulletin 115
r
atios Kt = F
S
g
ives the Kw
o
perations, a
n
t
from the re
s
n
g for genera
l
MA
R
MA
S
L
ocal Pressu
r
flushing for
t
transport.
A
u
ll. The outl
e
a
tion. The m
o
depth, 0.8 m
m
o
f approxim
a
w
hich shoul
d
ood of 1800
t
rap efficien
cy
means that
o
th dams trap
f
raction (3%)
m
aining after
m
is at 0.53
%
r
annum (IC
O
6.4x50x0.97
/
d
imentation
w
o
ved by pres
s
s. Pressure f
l
n
gth of the re
s
i
th water lev
e
u
ired dead st
o
l
ley is narro
w
(1999) provi
d
S
C/MAR (y
r
> 0.2 and Kt
n
d not drawd
o
s
ervoir was f
u
l
flushing by
Dam FS
L
R
river inflo
w
S
(million m
3
Kw (
y
Kt (
ye
r
e Flushin
g
S
the 70 m hi
g
A
bottom outl
e
e
t dimension
s
o
del was set
u
m
particle si
z
a
tely 50 000
m
d
be done dur
i
m3/s is exce
e
Fig. 3.
B
e
d
y
. This volu
m
o
nly after 50
y
almost all th
that is spille
100 years o
f
%
of the orig
i
O
LD, 2009).
/
1.35=589 m
i
w
ith flushin
g
s
ure flushing
l
ushing woul
d
s
ervoir, wate
r
e
l drawdown
o
rage capaci
t
w
but in this
c
d
es guidelin
e
r
) and Kw =
F
> 50 values
f
o
wn flushin
g
u
rther invest
i
water level
d
Table 3. NV
D
L
(masl)
w
(million
m
3
/a) (
@
1.35t
/
ears)
e
ars)
S
imulation b
y
h dam was s
i
e
t was desig
n
s
are 11.3 m
x
u
p with 200
0
z
e and 0.01 s
m
3 after 82
m
i
ng periods
o
e
ded.
d
levels simula
m
e is conside
r
y
ears of ope
r
h
e incoming
s
e
d downstrea
m
f
operation.
T
i
nal FSC per
A dead stor
a
i
llion m3.
g
g
without wat
d
keep the lo
r level draw
d
during flood
t
y and dam h
e
c
ase, there is
e
s on when fl
F
SC/MAS (
y
for the two
d
g
operations,
i
gated by 3D
d
rawdown.
D
project K
t
a
m
3/a)
/
m3)
y
3D CFD M
o
i
mulated by
A
n
ed to discha
x
11.3 m wit
h
000 cells wi
t
time step. F
i
m
inutes. The
r
o
f flood inflo
w
ted by 3D pre
s
r
ed as dead
s
r
ation will s
e
s
ediment loa
d
m
.
N
onethel
e
T
he worst rat
e
year, which
i
ge for sedim
e
e
r level dra
w
cal intake ar
e
d
own with fr
e
s can be use
d
e
ight. The
NV
little excess
w
ushing oper
a
y
r), where M
A
d
am scenario
s
are feasible.
modelling f
o
a
nd K
w
value
s
o
del
A
NSYS Flue
n
rge the 5-ye
a
h
an invert le
v
h a 0.4 m ce
l
i
g. 3 shows t
h
r
equired dur
a
w
into the re
s
s
sure flushing
s
torage in th
e
dimentation
s
d
(97% trapp
i
e
ss, both da
m
e
of sedimen
t
i
s slightly le
s
e
ntation for
5
w
down or by
w
e
a clear of se
e
e outflow d
u
d
to extend t
h
V
D reservoir
w
ater availa
b
a
tions could
b
A
S = mean a
n
s
which sugg
e
The feasibili
t
o
r local press
s
230
3008
12.1
0.74
184
n
t with user
d
a
r flood of 3
5
v
el of 162 m
a
l
l size, a 34
m
h
e simulated
tion of press
u
s
ervoir, typi
c
simulations
e
water reso
u
start to imp
a
i
ng efficienc
y
m
s have relati
v
t
ation experi
e
s
s than the ty
p
5
0 years was
w
ater level d
r
e
diment but t
o
u
ring floods
i
h
e reservoir l
i
r
shape is ide
a
b
le for flushi
n
b
e feasible, b
n
nual sedim
e
e
sts that nor
m
ty of flushin
g
s
ure flushing
2
3
0
1
1
4
d
efined func
t
5
00 m3/s whe
n
asl at the cu
r
m
sediment d
e
scour hole
w
u
re flushing
w
c
ally when th
u
rces
a
ct on the
y
) and it
v
e large
e
nced in
p
ical rate
allowed
r
awdown
o
remove
i
s
i
fe and to
a
l for
n
g.
a
sed on
e
nt load.
m
al
g
and 2D
2
50
0
08
2.1
.67
4
14
t
ions for
n
the
r
rent river
e
pth, 34
w
ith a
w
ould be
e current
3.2.2.
D
Emptyin
g
done wh
e
reservoi
r
when th
e
b
ecause
t
p
ossible
days are
Fi
g
D
rawdown
F
g
the reserv
o
e
n the dam i
s
r
are plotted
i
e
reservoir in
t
here is no e
x
to draw dow
required for
g
. 4. Daily res
e
Fig. 5. Se
d
F
lushin
g
D
uri
n
o
ir to flush d
u
s
spilling un
d
i
n Fig. 4 for
t
flow exceed
s
x
cess runoff
i
n the 70 m h
i
drawdown f
r
e
rvoir water l
e
d
iment deposi
t
ng
Floods
u
ring floods
c
d
er storage o
p
t
he 70 m and
s
500 m3/s. E
v
i
nflow avail
a
i
gh dam tho
u
r
om a full res
vels and com
b
t
ion after 50 y
e
c
ould result i
n
p
eration con
d
90 m high d
a
v
idently it is
a
ble to fill th
e
u
gh before a
n
s
ervoir to li
m
b
ined reservoi
r
e
ars for the 7
0
n
a reduced
f
d
itions. The l
o
a
ms without
not practica
l
e
reservoir a
f
n
y floods > 5
0
m
it the free o
u
r
flushing, spi
l
0
m high dam
w
f
irm yield an
d
o
ng term dai
l
flushing and
l
to flush the
ft
er drawdow
n
0
0 m3/s arriv
e
u
tflow discha
r
l
lage and abst
r
w
ith and with
o
d
would hav
e
l
y water leve
with flushin
g
90 m high d
a
n
flushing. I
t
e
at the dam.
r
ge to 2000
m
r
action disch
a
o
ut flushing
e
to be
ls in the
g
done
a
m
t
is
Twelve
m
3/s.
rges
2D hydrodynamic model flushing simulations were also done for all the observed floods from 1976 to
2026 with river flows > 500 m3/s. The same Mike21C model setup was used but with lower water levels at
the dam during flushing. Fig. 5 shows the simulated bed levels along the 70 m high dam for storage and
flushing operation after 50 years. With flushing, most of the incoming sediment load is sluiced through
the reservoir without deposition. Table 4 shows the simulation results for 50 and 100 years.
Under storage operation 589 million m3 sediment (97% trapping) would be deposited in the dam after 50
years, but if drawdown flushing is done only 107 million m3 sediment (18% trapping) would be deposited
in the dam. The actual operation of the dam could however deposit more sediment in the reservoir if the
operating rules are not followed judiciously, if a bottom outlet gate has to undergo maintenance, etc.
Other dams in South Africa with flushing operation have long term storage capacity of about 25% to 40%
of the original FSC but are poorly operated for flushing and the equilibrium reservoir storage capacities
could have been much larger. To be conservative the storage loss with drawdown flushing is estimated at
268 million m3 after 100 years (23% trap efficiency) for the 70 m high dam (12% FSC lost).
Table 4. Comparison of FSC in the 70 m high dam without flushing and with drawdown flushing for Q > 500
m3/s
Mode of operation Without flushing With flushing
Time period in years (X) 50 100 50 100
Volume of sediment in the reservoir after X-years (million m3) 589 1178 107 268
FSC remaining after X-years (million m3) 1651 1062 2133 1972
Percentage of the original FSC lost after X years (%) 26 53 5 12
Fig. 6 shows the expected change in FSC over time of the proposed 70 m dam. Without drawdown
flushing the FSC will decrease to almost zero after 200 years of operation. Drawdown flushing releases
only 58 million m3/a more water than would have spilled with storage operation. This corresponds to a
1.9% reduction in the MAR while pressure flushing would cause a 0.3% reduction. With drawdown
flushing, the long term equilibrium storage capacity is estimated at a minimum of 70 % of the original
FSC, equal to 1568 million m3, reached by 190 years after commissioning. These findings indicate that the
drawdown flushing of sediment could be feasible for the 70 m high dam and could even be more effective
if a lower dam is considered.
Fig. 6. Expected FSC changes due to sedimentation over a 300 year period for the 70 m high dam
3.3 Other Sedimentation Mitigation Measures
Other possible sedimentation mitigation measures at the reservoir include:
An upstream check dam is expensive and has a limited life due to large floods
Bypassing sediment or off-channel storage is not feasible at the NVD as the reservoir is too long
Passing incoming sediment loads through the reservoir by density current venting is not
applicable to the NVD site as it has turbulent suspended sediment transport
Dredging sediment from the reservoir has a high cost for large volumes, a large disposal area is
needed and it is not feasible economically or environmentally
0
500
1000
1500
2000
2500
0 50 100 150 200 250 300
Full supply storage capacity (million m
3
)
Years
Normal storage operation FSC (mcm) Flushing operation FSC (mcm)
Compensating for reservoir sedimentation by raising the dam and allowing additional dead
storage may be cost-effective in the short-term but it does not provide a long-term solution.
4. Conclusions
From the sediment yield analysis, the proposed long-term sediment load at the NVD site is
16.4 million t/a. Based on the 2D hydrodynamic modelling, the proposed NVD reservoir would have 47%
and 77% of the original storage capacity after 100 years of operation, for the 70 m and 90 m high dam
scenarios respectively. It would be beneficial to design the NVD project with sediment flushing to
maintain a long-term equilibrium storage capacity of the reservoir while ensuring that small floods from
the dam can still deposit increased sediment loads at the estuary. Reservoir sedimentation mitigation
measures were investigated to extend the life of the dam, based on ICOLD (1999) guidelines.
The ICOLD (1999) general dam flushing requirements indicate that the proposed reservoirs are relatively
too large for the runoff to have enough excess water for flushing during the full flood season and water
mass balance calculations indicated that the 90 m high dam cannot be flushed. Local pressure flushing and
reservoir drawdown flushing during floods > 500 m3/s were however evaluated for the 70 m high dam by
3D and 2D numerical modelling. It was found that flushing could be feasible and maintain a long-term
storage capacity, while without flushing the FSC will be almost zero after 200 years.
Based on the expected relatively large sustainable equilibrium FSC of 70 % of the original FSC for the
NVD project, it is proposed that the 70 m high dam (or lower dam), with a smaller dead storage for
sedimentation is designed with drawdown flushing during floods. The dam should have large bottom
outlets at the current river bed for drawdown flushing during floods exceeding 500 m3/s. Pre-releasing
should commence at least 12 days before the flushing starts and the flushing duration could be anything
from a few days to longer than a month based on the inflow flood hydrograph. Pressure flushing may be
required for the 70 m high dam with storage operation because the sediment deposition will occur near the
dam due to low water levels coinciding with large floods.
References
1. Le Grange , A. P., Badenhorst, D. B., & Basson, G. R. (1997). Sedimentation of Reservoirs in the Orange River
Basin. Orange River Re-planning Study (ORRS). Report No PD 0010015495. Department of Water Affairs. Pretoria.
2. DWAF. (2005). Dam Development Options and Economic Analysis Report: Appendix G: Sedimentation Analysis. Pre-
feasibility study into measures to improve the management of the lower orange river and to provide for future
developments along the border between Namibia and South Africa (LORMS), Report number PB D000/00/4403.
3. WRC. (2012). Sediment yield prediction for South Africa – 2010 Edition. Compiled by Msadala, V., Gibson, L., Le
Roux, J., Rooseboom, A. & Basson, G.R. SA Water Research Commission.
4. Rooseboom, A. (1992). Sediment transport in rivers and reservoirs – a Southern African perspective. WRC Report No.
297/1/92. Water Research Commission. Pretoria, South Africa.
5. DWS. (2015). Dam list - Reservoir survey data base. SA Department of Water and Sanitation.
6. ICOLD (1999). Dealing with Reservoir Sedimentation. Guidelines and case studies. Bulletin 115.
7. ICOLD (2009). Sedimentation and Sustainable Use of Reservoirs and River Systems. BULLETIN 147
Acknowledgements
The authors wish to thank the South Africa-Namibia Permanent Water Commission (PWC) and AECOM for
their permission to publish this paper. The contents of this paper remains the intellectual property of the PWC.
The opinions and views presented in this paper are, however, those of the authors and do not necessarily reflect
those of the PWC and AECOM.
The Authors
Prof G.R. Basson is the Water and Environmental Engineering Division Head of the Civil Engineering Department of
Stellenbosch University. He obtained a PhD from Stellenbosch University in 1996 and has more than 30 years’ experience in
the fields of river hydraulics, fluvial morphology and the design of large hydraulic structures. He has worked on projects in
21 countries. He is honorary Vice President of ICOLD.
Dr O. Sawadogo graduated with a PhD from Stellenbosch University in 2015. He has more than 9 years’ experience in
hydrodynamic mathematical modelling applied in river hydraulics and fluvial morphology, flood hydrology and flood line
determination. He currently holds a position of Hydraulic Modeller and Researcher at the Department of Civil Engineering,
Stellenbosch University.
J.K. Vonkeman is a post-graduate water engineering student and junior research engineer at Stellenbosch University. She
completed her BEng (Cum Laude) in Civil Engineering at the University of Pretoria in 2014 and is currently developing a
numerical model for her PhD on bridge pier scour.
ResearchGate has not been able to resolve any citations for this publication.
Sedimentation of Reservoirs in the Orange River Basin. Orange River Re-planning Study (ORRS)
  • A P Le Grange
  • D B Badenhorst
  • G R Basson
Le Grange, A. P., Badenhorst, D. B., & Basson, G. R. (1997). Sedimentation of Reservoirs in the Orange River Basin. Orange River Re-planning Study (ORRS). Report No PD 0010015495. Department of Water Affairs. Pretoria.
Sediment yield prediction for South Africa -2010 Edition
  • Wrc
  • V Msadala
  • L Gibson
  • J Le Roux
  • A Rooseboom
  • G R Basson
  • Sa Water Research Commission
WRC. (2012). Sediment yield prediction for South Africa -2010 Edition. Compiled by Msadala, V., Gibson, L., Le Roux, J., Rooseboom, A. & Basson, G.R. SA Water Research Commission.
Sediment transport in rivers and reservoirs -a Southern African perspective
  • A Rooseboom
Rooseboom, A. (1992). Sediment transport in rivers and reservoirs -a Southern African perspective. WRC Report No. 297/1/92. Water Research Commission. Pretoria, South Africa.
Dealing with Reservoir Sedimentation. Guidelines and case studies
  • Icold
ICOLD (1999). Dealing with Reservoir Sedimentation. Guidelines and case studies. Bulletin 115.
Sedimentation and Sustainable Use of Reservoirs and River Systems
  • Icold
ICOLD (2009). Sedimentation and Sustainable Use of Reservoirs and River Systems. BULLETIN 147
Dam Development Options and Economic Analysis Report: Appendix G: Sedimentation Analysis. Prefeasibility study into measures to improve the management of the lower orange river and to provide for future developments along the border between Namibia and South Africa (LORMS)
  • Dwaf
DWAF. (2005). Dam Development Options and Economic Analysis Report: Appendix G: Sedimentation Analysis. Prefeasibility study into measures to improve the management of the lower orange river and to provide for future developments along the border between Namibia and South Africa (LORMS), Report number PB D000/00/4403.
Dam list -Reservoir survey data base. SA Department of Water and Sanitation
  • Dws
DWS. (2015). Dam list -Reservoir survey data base. SA Department of Water and Sanitation.