-
I
Rodney White
I ThomasTelford
-
I
\
I
i . ..,, :-
'\
Published by Thomas Telford Publishing, Thomas Telford Ltd, 1 Heron Quay, London E14 4JD.
URL: http://www.thomastelford.com
i
:-1
~I
-j
I
-I
I
II
A catalogue record for this book is available from the BIitish Library :--1
In many areas of the world the life span of reservoirs is determined by the rate
of sedimentation which gradually reduces storage capacity. Eventually, this
process destroys the ability of the scheme to deliver the benefits for which it was
built. Many major reservoirs are approaching this stage in their life.
There are various options available for positively managing sedimentation in
reservoirs.
a sluicing - the process of passing sediment laden flo09 waters through the
reservoir. This method involves the reduction of water levels in the reservoir
during the flood season and is applicable mainly to very fine sediments (clays
and silts).
" density current venting - this method has the attraction that it is not necessary
to lovver water levels but is only applicable in very exceptional circumstances
where sediment-induced density currents carry very fine sediments towards the
dam. The number of cases where density current venting has been successful
is minimal.
iii
EVACUATION OF SEDIMENTS
All the above methods can be used to extend the useful life of reservoirs.
However, their technical, economical and environmental feasibility depend on a
number of specific factors including:
iv
PREFACE
worldwide and draws together this existing knowledge into a concise manual for
practising engineers.
The book begins by assessing the scale of the problem of reservoir
sedimentation. It assesses the volume of storage that is likely to be lost to
sedimentation and compares this volume with the net volulne of storage that is
likely to be required to meet continuing demand. The book provides a review of
the current state of knowledge of reservoir flushing, and then considers the
worldwide experience of flushing to date. Areas of the world are then identified
where flushing is likely to be most useful. The final section of the book describes
the more detailed investigations which must be carried out when considering
sediment flushing at a particular dam site.
v
Acknowledgements
This book describes work which was funded principally by the Department of the
Environment, Transport and the Regions (DETR) as part of the Partners in
Innovation programme. The work was undertaken jointly by HR V/allingford, as
the lead partner, TAMS UK, Binnie Black and Veatch, and LAWGIBB.
It is a pleasant duty to acknowledge the valuable contributions made by
Laurence Attewill and Atila Bilgi of TAlVIS UK, Ed Atkinson and Andrew Nex
of ILl{ Wallingford, John Ackers, Chris Scott and Robert Jones of Binnie Black
and Veatch, and Richard Wingfield and Mary-Ellen Cromack of LAWGIBB.
HR Wallingford is an independent specialist research, consultancy, software
and training organisation that has been serving the water and civil engineering
industries worldwide for over 50 years in more than 60 countries. \Ve aim to
provide appropriate solutions for engineers and managers working in:
• water resources
II» h~gation
o groundwater
• urban drainage
• rivers
• tidal waters
• ports and harbours
o coastal waters
• offshore.
vii
Notation
DDR the ratio of the height of water at the dalTI during flushing to the
maximum height of water at the dam (to reservoir retention
level), both measured above original river-bed level (non-
dimensional)
DSOT 50 percentile size of sediment in transport (IILrn)
D50B 50 percentile size of river-bed material (mm)
F~VR' the ratio of the natural width of the flushing channel and a
representative bed width for the reservoir (non-dimensional)
H fiush the height of water at the datu during flushing, measured above
original river-bed level (m)
Hmax the maximum height of water at the dam (to reservoir retention
level), measured above original river-bed level (m)
LTCR the sustainable storage capacity divided by the original storage
capacity of the reservoir (non-dimensional)
lv!in the average sediment inflow rate (t/yr)
LV the interval between flushing operations (yrs)
P sand proportion of total sediments in motion which exceed 0-06 rom
in size (sand and coarser material)
Qf the flushing discharge (m3/s)
Qm mean annual flow (m 3/s)
Qs the sediment transporting capacity of the flow in the incised
flushing channel (tis)
S the longitudinal energy gradient trliough the reservoir (non-
dimensional)
SBR the ratio of sediment flushed to sediment depositing (non-
dimensional)
SBR d the specific value of SBR related to flushing with maximum
reservoir drawdown (non-dimensional)
SSR sand-size ratio, DsoTIDsoB .~
TE the trapping efficiency of the reservoir, i.e. ratio of sediments
retained within the reservoir to sediments~.;,entering (non-
dimensional) .
Tf the duration of flushing (days)
TIVR the ratio of the natural top width of the flushing channel and a
representative top width for the reservoir
\tV the bed 'width of the incised flushing channel (m)
¥Ibed the representative bottom width of the reservoir, taken as the
bottom width of the reservoir just upstream of the dam (m)
ix
EVACUATION OF SEDIMENTS
Wtep the representative top width of the reservoir, taken as the top
width of the reservoir just upstream of the dam (m)
Wmin the lesser of Wand Wbed (m)
a the angle of the side slope of the incised channel formed during
flushing (zero is horizontal) (degrees)
p the density of the deposits expressed as weight of dry material
per unit volume (tlm3 )
"¥ constant related to the sediment type (non-dimensional)
x
I
Contents
Dlustrations xv
1. Executive summary 1
1.1. Introduction, 3
1.2. Summary of conclusions, 4
1.2.1. Review of sedimentation in reservoirs, 4
1.2.2. Research into factors Wllich influence
sediment flushing, 7
1.2.3. vYorldwide experience of sediment
flushing, 9
1.2.4. Geographical areas suited to flushing, 11
1.2.5. Site-specific investigations and design
considerations, 13
xi
EVACUATION OF SEDIMENTS
reservoirs, 28
2.6.1. Worldwide, 28
2.6.2. Distribution of construction of storage, 30
2.6.3. Comparison of storage construction with
demand, 30
2.7. Rate and distribution of loss of storage, 31
2.7.1. Rate of loss of storage, 31
2.7.2. Distribution of loss of storage, 32
2.8. Trends in the rate of loss of storage, 34
2.9. Reservoir size and rate of loss of storage, 35
2.10. Requirements for new storage, 36
xii
CONTENTS
7. References 131
8. Bibliography 141
Appendices 149
Appendix 1. Reservoir data, 151
Appendix 2. Numerical model case study, 163
.8..Dlue:nWlX 3. Flushing case studies, 171
Erosion, 211
251
xiii
I!' ..
Illustrations
Tables
Table 1.L Demand for storage 5
Table 1.2. Geographical demand for new storage 5
Table 1.3. Demand for new storage, South and Central
A.merica 6
Table 1.4. Demand for new storage, Africa 6
Table 1.5. Gross storage requirements to 2010 8
Table 2.1. Distribution of reservoir storage volume 19
Table 2.2. Growth in 'world population 20
Table 2.3. Prediction of global demand (after Shiklamanov) 21
Table 2.4. Growth in irrigation area 21
Table 2.5. Growth in energy generated by hydropower 22
Table 2.6. Comparison of actual andeconolTlJcal potential
energy 22
Table 2.7. Annual gro\tvth rates and increase in storage 23
Table 2.8. European growth in irrigation and hydropower 24
Table 2.9. North American growth in irrigation and
hydropower 25
Table 2.10. South American growth in irrigation and
hydropower 26
Table 2.11. Annual growth rates, South and Central America 26
Table 2.12. African growth in irrigation and hydropower 27
Table 2.13. Asian and Oc~anian growth in irrigation and
hydropower 28
Table 2.14. Annual growth rates, Asia and Oceania 29
Table 2.15. Regional demand for new storage 29
Table 2.16. of construction of new storage 29
Table 2.17. Distribution of storage increase 30
Table 2.18. Regional sedimentation rates 31
Table 2.19. Extent of sediment data 32
Table 2.20. Distribution of sediment rate and storage loss 33
Table 2.21. Ringlet reservoir, sedimentation 34
Table 2.22. Gross requirement for new storage 35
Table 3.1. Application of sediment balance and long-term
capacity ratios to existing reservoirs 46
xv
EVACUATION OF SEDIMENTS
Figures
Figure 2.1. Growth in world population 20
Figure 2.2. Comparison of growth rates 23
Figure 2.3. Historic growth in reservoir storage 30
xvi
ILLUSTRATIONS
xvii
e u rn a y
;'.: ' .
-- ..
....
I. Executive summary
1.1. INTRODUCTION
There are around 40 000 large reservoirs worldwide used for water supply, power
generation, flood control, etc. Between a half and one per cent of the total storage
volume is lost annually as a result of sedimentation and 300 to 400 dams, at the
cost of around £5 million per dam, would need to be constnlcted annually to
maintain current total storage. The introduction of flushing systems in some old
dams, where appropriate, and in the design of new dams could save 10 per cent
of these costs, i.e. £200 million annually. This book provides guidelines on the
design aspects of flushing systems and indicates where such systems could be
used beneficially. '
The benefits attributable to dams and reservoirs, most of which have been built
since 1950, are considerable and they have improved the quality of life
worldwide. These benefits can be classified under three main headings.
Irrigation
About 20 per cent of cultivated land worldwide is irrigated, some 300 million
hectares. This irrigated land produces about 33 per cent of the worldwide food
supply. Irrigation accounts for about 75 per cent of the world water consumption,
far outweighing the domestic and industrial consumption of water.
Hydropower
About 20 per cent of the worldwide generation of electricity is attributable to
hydroelectric schemes. This equates to about 7 per cent of worldwide energy
usage.
In many areas of the world the life span of these reservoirs is determined by the
rate of sedimentation which gradually reduces storage capacity and eventually
destroys the ability to provide water a.Tld power when sedinlents clog low level
outlets. Many major reservoirs are approaching this stage in their life.
One way of preserving reservoir storage is to flush sediments through purpose-
built outlet works within the dam. This technique can be applied to existing dams
(with adaptation of the works) and to ndw dams. However, the
technique is only effective under certain favourable· conditions and is not
applicable universally. The alternative is to build more dams to replace the
3
EVACUATION OF SEDIMENTS
depleting storage of the existing stock. However, there are fewer and fewer good
dam sites available and new dams can have serious environmental and social
consequences.
Dams designed within the last ten years or so, have sometimes incorporated
design features which will allow flushing to be undertaken when appropriate.
However, these designs have been based on considerations which are site-
specific. The purpose of this project was to produce a generally applicable design
manual which provides guidance on the necessary hydrological, hydraulic,
sedimentological and topographical features for successful flushing. It is based
on a review of recent research and field experience worldwide and draws together
this existing knowledge into a concise manual for practising engineers.
The book starts by assessing the scale of the problem of reservoir
sedimentation. It compares the volume of storage that is likely to be lost to
sedimentation and compares this volume with the net volume of storage that is
likely to be required to meet continuing demand. The book provides a review of
the current state ofknowledge of reservoir flushing, and from this proceeds to
consider the worldwide experience of flushing. Areas of the world are then
identified where flushing is likely to be most useful. The final section of the book
describes the more detailed investigations which must be carried out when
considering sediment flushing at a particular dam site.
-.i
1.2. SUMMARY OF CONCLUSIONS
Distribution of storage
The worldwide distribution of existing storage and storage under construction, as
determined from the International Commission on Large Dams (ICOLD)
Register (1998), is shown in Table 2.1. The Americas, together with northern
Europe and mainland China, account for 70% of the existing world stock of
reservoir storage.
J
EXECUTIVE SUMMARY
Distribution of demand
Europe. Although the dema.'1d for new storage is sensibly zero in much of
Western Europe, it does appear that for the region as a whole there is a small
demand, of the order of 1% per annum, for new storage for hydropower, mainly
concentrated in Eastern Europe.
North America. Although the data show that the energy generated by
hydropower, as well as the area of land under irrigation, continued to ,grow
through to the 1990s, the fact that no new storage was constructed in that period
..;"'.... ~;".::, ....'v .. ..., that the data are influenced by operational factors. Therefore, although
49 51 54
5
EVACUATION OF SEDIMENTS
Table 1.3. Demand for new storage, South and Central America
South and Central America. The data show that the energy generated by
hydropower, as well as the area of land under irrigation, grew strongly through
the 1980s and into the 1990s. Furthermore, it is estimated that only 21 % of the
economically-feasible hydropower potential has so far been developed, so that
short to medium term growth is unlikely to be constrained by shortage of sites.
From the rates of growth for population, water consumption, irrigation area and
hydropower, the growth rates for demand of storage, shown in Table 1.3, are
postulated.
Africa. The data show that the energy generated by hydropower, as well as the
arya of land unde~ irrigation, grew weakly through the 1980s and into the 1990s
at a rate well below the rate of growth of population. This trend is likely to
continue, despite strong demand and great potential, so an annual growth rate of
2% is postulated.
Asia and Oceania. Given the rates of growth for population, irrigation area and
hydropower, the growth rates for demand of storage, shown in Table 1.4, are
postulated.
2000-2010 2-0
')0')0..')010
~
6
EXECUTIVE SUMMAR.Y
apparent fall in the rate of construction during the 19908 is dramatic and may, in
part, be due to inadequate data.
The distribution of the growth of new storage is set out in Table 2.14.
7
EVACUATION OF SED1MENTS
Europe 49 54 103
I.
• Reservoirs where the annual run-off is large compared with the volume of the
reservoir are suitable for sediment flushing.
• Reservoirs where there is a regular annual cycle of flows and a defined flood
season are suitable for sediment flushing. This favours sites in monsoon areas
and sites where flood flows are generated by annual snowmelt in the spring and
summer months.
• Reservoirs where release of significant quantities of water for flushing does
not significantly affect the ability to satisfy water demands at other times of the
year.
• Graded bed sediments produce conditions which are the most conducive to the
efficient flushing of sediments. Such conditions are typical of gravel rivers
with a varying bed material composition. In large rivers this situation is found
where the longitudinal bed gradient is between, say, 0·001 and 0·002. In
smaller rivers the equivalent range nlay be 0·002 to 0·005.
8
EXECUTIVE SUMMARY
c From the point of view of sediment alone, delta deposits of fine sand and
coarse silt are the most easily flushed. Coarser material is difficult to move and
tends to deposit at the upstream end of the reservoir. Finer Inaterial which
deposits in the body of the reservoir outside any incised channel will not be
available for reworking during flushing.
Site-specific factors
The most suitable conditions for flushing are to be found in reservoirs that are
approximate i.T1 shape to the incised channel which develops during flushing.
Long, relatively narrow, reservoirs are better suited to flushing than short, wide,
shallow reservoirs.
Summary
Reservoirs in tb.e 'upper and middle reaches' of rivers are likely to be best suited
to sedinlent flushing for the following reasons.
\I In the lower reaches, reservoirs are likely to have inundated areas that have
previously been flood plains and these areas would not be reached by the
incised flushing channel which is inevitably of limited width.
fII The longitudinal slope available for the flushing channel is relatively small,
thus limiting the amount of sediment transport.
II Reservoir volumes in the lower reaches are likely to be larger compared with
the mean ap. .11ual
. run-off and hence water availability becomes a restraint on
sediment flushing. .-
9
EVACUATION OF SEDIMENTS
Operational limitations
Operational considerations, such as water and power demands, can inhibit the
ability to flush successfully, but they must not be allowed to prejudice the long-
term preservation of an important resource.
Downstream impacts
Downstream impacts can act as a constraint in the planning and operation of
flushing. In some cases flushing may be ruled out, whereas sluicing, which
approximately preserves the seasonal distribution of sediment load, maybe a
practicable alternative.
10
!" • y
E.rosion rate
The erosion rate depends on a complex interaction of the following factors.
Transport of sediment
In order for reservoir flushing to be needed, it is necessary for sediment to be
eroded in the catchment, transported down the river system and deposited in the
reservoir. The efficiency of the transport process is expressed by the sediment
delivery ratio, which is the proportion of sediment eroded from the land that is
discharged into rivers (Morgan and Davidson, 1986).
The sediment delivery ratio is generally higher for sediment derived from
channel-type erosion which delivers sediment to the main channels of the
transport system more quickly and directly than in the case of sheet erosion.
The poor correlation between sediment yield and erosion rates makes it
difficult to estimate the sediment load entering a reservoir on the basis of
erosion rate within the catchment (Morris and Fan, 1997). Most studies that have
attempted to relate the delivery ratio to catchment characteristics have found that
the delivery ratio decreases as the catchment area increases (Walling and Vvebb,
1983).
Climatic zones
An understanding of the precipitation regimes throughout the world may allow
the definition of climatic zones based on temperature and precipitation regimes.
This may permit definition of areas of high and low erosion rates. It is
difficult to classify distinct climatic zones as they tend to merge into one another
rather than have sharp boundaries, but a number of general models have been
produced.
II
EVACUATION OF SEDIMENTS
There have been many climatic classifications produced but one of the most
common is based on the original Koppen classification, with eight climatic
regions based on four temperature zones and one moisture zone and the seasonal
domination of air masses. Details of this classification are given in Chapter 5 and
a discussion of alternative classifications is given in Appendix A4.3.
The eight Koppen climatic regions are as follows.
Hydrological characteristics
Experience has shown that low reservoir water levels provide the most effective
conditions for sediment flushing. To allow water levels to be lowered requires
confidence that rainfall can be relied upon to refill the reservoir. It follows that
well defined wet and dry seasons will be favourable for a sediment flushing
regime. Such a climate is defined by Koppen as tropical wet and dry: Aw, Am and
BS. Also, there are areas in the mid-latitudes where spring snowmelt provides a
regular and predictable annual pattern of high flows.
River discharges must also be sufficient to transport sediment loads through
the reservoir. Regions of low precipitation like the Sahara and other desert
environments therefore will not be suitable for flushing even if they exhibit a
defined seasonal effect. The availability of water will also affect the duration and
discharge rate of the flow required for flushing. Where there is a limited amount
of water it is better to use a high discharge for a short period of time than a low
discharge for a long period of time. This increases the amount of sediment that
is removed.
12
EXECUTIVE SUMMARY
• areas in Central Asia where the rivers are fed by the Himalayas, including
Pakistan, India and Nepal
& parts of Asia, including Calnbodia, Vietnam and Thailand.
Site investigations
Site investigations are required to identify t.~e most compact and efficient
geometry for the flushln.g outlets and the energy dissipation works. The reservoir
itself requires a detailed survey to establish its topography.
Hydrological investigations
Inflows to the reservoir need to be established with confidence. This involves the
acquisition of historical records of river flows going back at least 30 years and
preferably longer and/or the development of a longer sequence from rainfall
records using catchment modelling.
Sediment investigations
The amount and nature of the sediment entering, or likely to enter, the reservoir
needs to be established. This requires measurements of sediment transport rates
in the rivers feeding the reservoir over many years to establish the results with the
confidence that is required.
In the case of existing reservoirs, information about the amount of sediments
entering the reservoir can be augmented by surveys of the amount and nature of
the material settling within the reservoir. is required, however, to allow. for
the amount of material, mainly fine, which passes through the reservoir without
deposition.
Bed material sampling should be undertaken in the reservoir and in the rivers
which feed the reservoir. A sound knowledge of the nature of these sediments,
including their size, specific gravity and degree of compaction, is an essential
requirement to provide inputs for numerical models which simulate sediment
moven1ent, see below.
Hydraulic modelling
Numerical (computer) modelling of the way sediment is likely to behave within
the reservoir and the amount and nature of the sediment which will be passed to
the downstream reach is the cornerstone of any detailed evaluation of flushing
facilities.
Computer simulations of reservoirs ideally use representative, long-term
sequences of water and sediment inflows to the reservoir. The models are capable
13
EVACUATION OF SEDIMENTS
14
•
f •
atio
er\f Irs
t
2. Review of sedimentation in reservoirs
2.1. SUMMARY
In this chapter a summary of the total volume of reservoir storage, and its
distribution is given. An attempt is made to quantify the future demand for new
storage, especially for hydropower and irrigation, and this estimate is compared
with the historic rate of reservoir constnlction throughout the twentieth century.
The rate of loss of storage due to sedimentation is made, so as to arrive at a
prediction of both the net and gross future storage requirements. '
s Japan reported only dams greater than 30 m high. .c_ "" _"
= Russia reported mainly hydropower dams. ' '' . '''., . .
o Some countries failed to respond 8....1J.d for these countrie$ .data:vvasretained
from the earlier edition.
--
The register gives the total number of dams reported by the 80 member countries
and the 60 non-member countries as 25 410. No exact "sUlTilllary of storage
volume is provided but in the introduction it is stated that the total volume of
storage is 6000 km3 • From the analysis of the data in the register, the total gross
storage volume of the reservoirs reported by ICOLD is 6465 km3• This includes
490 km 3 of storage registere9 as under construction.
In order to estimate the total world storage it is necessary therefore to assess
the extent to which the register under-reports the total number of dams and the
number of dams less than 5 m high (and their storage). It is evident from the
res!:ister that the ratio of dams less than 30 m high to the total number of dams
varies from about 90% in the case of India to 5%~in the case of China. From this
it can be infened that many countries, but China in particular, under-report dams
17
EVACUATION OF SEDIMENTS
2.2.3. Conclusion
The best estimate of world storage in reservoirs (excluding natural lakes used as
storage for power and irrigation) is 6815 km3 ,
18
SEDIMENTATION IN RESERVOIRS
2.4.1. Population
The world's population in 1990 is estimated to have been 5286 million, growing
at an annual rate of 1·5%. This rate of growth is forecast to decline in the coming
decades so that the predicted future world population is as shown in Table 2.2
and Figure 2.1.
Water demand is expected to continue to grow at a faster rate than that
predicted by population growth alone. This is because the present -per capita
19
EVACUATION OF SEDIMENTS
2000 6158 I
1·34
I
I
2010 7032 1·15
I
i
2020 7887 0·95
2050 9833
12000 1·8
1·6
10000
... ---------- --
Population
-- _
(J)
--- -- -- ~
c0 6000
......
~
0)
0·8
'3 Cii
::I
0..
0 4000 c
a.. 0·6 c
<:
0·4
2000
0·2
OL-------~------~--------~ __~--~--------~----~O
1990 2000 2010 2020 2030 2040 2050
Year
20
,-
I
SEDIMENTATION IN RESERVOIRS
the two. It is pertinent therefore to examine the historic growth in irrigation and
energy generation.
2.4.2. Irrigation
The gro\vth in demand of water for irrigation is illustrated by the growth in
irrigated area, given in Table 2.4.
These data give only an ~indication of the growth of demand for water since
they are a record of the irrigated area actually planted and are subject to annual
variations in water availability: the effects of the drought in the 1980s is apparent.
Moreover, no data are available on the proportion of the total demand met from
storage reservoirs compared with other sources.
2.4.3. Hydropower
Historic groV'lth
\-Vorldwide power consumption is growing at a faster rate than population growth
as nations industrialise. The share of energy generated by hydropower is difficult
1995 260
21
EVACUATION OF SEDIMENTS
1995 2474000
to predict as the viability of new schemes is influenced heavily by the price of oil
and gas and the safety and environmental concerns associated with nuclear
power. The cost of new hydroelectric schemes increases in real terms as the best
sites are used up.
The historic energy generated by hydropower is shown in Table 2.5.
As with the data for irrigation, the link between these data and reservoir
storage is tenuous: the energy generated is affected by water availability and does
not directly reflect either the growth of installed capacity or storage. The high
growth rates in the 1970s is a reflection of the quadrupling of oil prices in 1972.
Potential
The potential for new hydropower development is indicated in Table 2.6, which
compares average energy generated by hydropower with the estimated
economically-feasible energy in each continent.
These data (Table 2.6) show that future hydropower development is unlikely
to be constrained by the lack of suitable sites within the foreseeable future.
22
SEDIMENTATION IN RESERVOIRS
2.4.4. Conclusion
The rates of growth of population, global water demand, irrigation area and
hydropower generation are compared in Figure 2.2.
From the rates of growth for population, water consumption, irrigation area
and hydropower, the following growt...1-t rates for demand of storage are postulated
and are shown in Table 2.7.
2.5.1. Europe
Population
The population of Europe (including Russia) is predicted to grow from 722
million in 1990 to a maximum of 730 million in the year 2000, and thereafter
decline to 723 million in 2020 and to 678 million in 2050. The overall effects of
23
EVACUATION OF SEDIMENTS
Conclusion
Although the demand for new storage is sensibly zero in much of Western
Europe, it does appear that for the region as a whole there is a small demand, of
the order of 1% per annum, for new storage for hydropower, and this is mainly
concentrated in Eastern Europe. However, against this must be considered the
relative shortage of suitable new sites and the strong opposition to new reservoirs
from environmental groups. It is therefore concluded that the overall demand for
new storage in Europe will grow at about 0·5% per annum.
24
~.
!
.~
·i
~<;
f
i SEDIMENTATION IN RESERVOIRS
of population will result In regional shortages and surpluses which may generate
demand for reservoir storage.
Conclusion
, Although the data show that the energy generated by hydropower, as well as the
area of land under irrigation, continued to grow through to the 1990s, the fact that
no new storage was constructed in that period suggests that the data are
influenced by operational factors. Therefore, although there remains a large
undeveloped potential resource in Canada, environmental pressures will
probably preclude any further development.
25
EVACUATION OF SEDIMENTS
5·3
Conclusion
The data show that the energy generated by hydropower, as well as the area of
land under irrigation, grew strongly through the 1980s and into the 1990s.
Furthermore, it is estimated that Qnly 21 % of the economically-feasible
hydropower potential has so far been developed, so that short to medium term
growth is unlikely to be constrained by shortage of sites. From the rates of
growth for popUlation, water consumption, irrigation area and hydropower, the
growth rates for demand of storage shown in Table 2.11 are postulated.
2.5.4. Africa
Population
The population of Mrica is predicted to grow from 663 million in 1990 to 832
million in the year 2000, rising to 1348 million in 2020 and 2141 million in 2050
- with the growth rate declining from 2·77% at the present time to 1·53% by
26
SEDIMENTATION IN RESERVOIRS
2050. Because of the relatively low per capita consumption at the present time,
water demand could increase at a rate well in excess of these rates.
Conclusion
The data show that the energy generated by hydropower, as well as the area of
land under irrigation, grew weakly through the 1980s and into the 1990s at a rate
well below the rate of growth of population. This trend is likely to continue,
despite strong demand and great potential, so an annual growth rate of 2% is
postulated.
27
EVACUATION OF SEDIMENTS
million ha % I %
I
I
1975 123·2 1·9 235000 4·9
The annual growth rate has been reasonably constant over this 20 year period
at an average rate of 1·97% per annum - well above the population growth rate.
The average growth rate of hydropower over this 20-year pedod is 4·25%, about
three times the population growth rate, reflecting the strong growth of the
economies of many of the Asian countries in this period. It is unlikely that this
differential is likely to persist in the future.
Conclusion
From the· rates of growth for population, irrigation area and hydropower, the
growth rates for demand of storage are postulated and are shown in Table 2.14.
2.5.6. Summary
Based on the above, the forecast future demand for storage is shown in
Table 2.15.
2.6.1. Worldwide
The historic rate of constluction of storage worldwide is shown in Table 2.16 and
Figure 2.3.
The overall annual growth rate for the century as a whole has been 6·5%. It is
interesting to note that neither the two World Wars nor the Depression have made
any serious impact on the rate of growth: in this context the apparent fall in the
rate of construction during the 1990s is dramatic and may in part be due to
inadequate data.
28
SEDIMENTATION IN RESERVOIRS
2000-2010 2·0
2010-2020 1·5
2020-2030 1·0
1910 21
I 11·6
1920 63 6-7
1930 121 I
7·6
I
29
EVACUATION OF SEDIMENTS
1
6000 12
5000 /
I
A
\
\ I
I
I '"
" 10
1.
I \ I \
I \ I \
1
CO)
E
.::<!. I \ I \
4000 I \ I \
CD
E I \ ............ , I \
::::l
"0
I "'............" I \
> , I
CD 3000 , I \
Ol
~ 'J \
.e \ ,,
til
ta
;e
2000 - - - World gross storage ,,
- - - - - Annual growth rate ,,
1000 ......... 2
.......... ..... ,
oL-~~~~==~~~~==~--~----~--~__-L__~O
1900 191.0 1920 1930 1940 1950 1960 1970 1980 1990 2000
Year
ear I
Annual growth rate in following decade: % ,1
Europe North America South America Africa Asia
I
1950 0·05
I
2000
I
30
1
1
SEDIMENTATION IN RESERVOIRS
• the complete loss of storage of the Wetzman reservoir on the River Gail in
Austria withinone year of commissioning in 1883 (Cyberski, 1973)
• the almost negligible loss of storage (0·1 % per annum) found from the suriey
of 95 reservoirs in the UK (White et at., 1996).
In order to assess the variation in the rate of loss storage around the world, data
from approximately 2300 dams in 31 countries have been gathered and analysed.
The summary of the results of the analysis is given in Table 2.18. The countries
have been grouped by geographic regions, taking account of the global map of
sediment yield (Walling, 1984).
China 2·30 22
~"I
Middle East 1·50 ~J
31
EVACUATION OF SEDIMENTS
J
The estimates are based on varying quantities of data gathered, as shown in
Table 2.19.
The estimates of annual loss of storage due to sedimentation have been used
in conjunction with the gross storage volume data available in the ICOLD World
register of dams (1998) to estimate the magnitude of the sedimentation problem.
The results of the analysis are displayed in Figure 2.4.
In summary, the analysis shows that by the year 2000 approximately 567 km3
(10% of the current gross available storage in the world) has been lost to
sedimentation. From the data available from the 1325 registered dams under
construction, it can be seen that the average gross storage volume of new
reservoirs is approximately 370 M.m3 • Therefore, in order to replace the volume 1
lost to sedimentation, over 2000 average-sized dams would have to be
constructed around the world.
1 844 530
, 109980 6·0
North America
i
32
SEDIMENTATION IN RESERVOIRS
6000
5000
C'l
E
.::£
Q.j 4000
01
~
~ 3000
m
;0
~ ::;:f'-_- -'-,~.= =~ ; .; ; J.'- ':-" "L. .-: " -;'. .l:. . -_-.l-: ~ _- L_- -JL. -. _J-. _. . l. -_- -L. ._- L_.- -J
1900
.-.=..-=
..- ""_1.
1910 1920 1930 1940 1950 1960
Sedimentation
33
EVACUATION OF SEDIMENTS
It is interesting to note that the annual loss of world storage estimated by this
study, 0-48% per annum, is almost half that estimated by Mahmood (1987), who
quotes a figure of 1%. Although it is not clear how Mahmood estimated his
figure, it can be argued that the figure determined by this study may
underestimate the problem. This is further supported by estimates made by other
authors, such as Goldsmith and Rildyard (1984), who estimate the annual loss of
storage in central Europe to be 0·5%, compared with the estimates for Europe by
this study of 0·2% and 0·17% per annum.
34
SEDIMENTATION iN RESERVOIRS
4·0
3·5
E
::::l
c:
c: 3·0
...
CiS
<V
Based on data from
c.. 1105 US reservoirs
~ 2·5
1ri
CI
e:!
0 2·0
t;
'0
Co"
tfl 10S
.Q
(l)
CI
e 1·0
ID
~
0·5 t "i"":
Continent
demand
2000-2010
Europe 49
35
EVACUATION OF SEDIMENTS
1% per annum. At the other extreme, 23 of the reservoirs studied had a storage
volume in excess of 1233 M.m3 and an average rate of loss of storage of 0.16%
per annum.
36
earc In fa tor
•
ue u Ing
30 Research into factors which influence
flushing
3.1. INTRODUCTION
This review of the factors which contribute to the efficiency of sediment flushing
operations' for reservoirs is based both on field experience, which has been built
up over the past 30 years or so at dams that have fluspjng systems LTl regular
operation, and on research findings using simulation models, either numerical or
physical, which in tllln rely on fundamental experimental data concelning the
detailed physics of the movement of sediments in water.
Field installations provide information which includes:
The data are very valuable in looking at the efficiency of flushing operations.
They represent 'real' situations and there are no scale effects or other simulation
deficiencies to mask the findings. However, tt:lere are shortcomings:
• there are only a few reservoirs being flushed at present and these do nbfcover
the full range of conditions where flushing might be considered: i.e. the data
forms a sparse matrix
• the data are rarely comprehensive enough to cover the complex situation found
in the field
• the records are short in terms of sediment deposition and they do not,
therefore, necessarily represent the hue long-term situation
o the historic development of flushing systems started, for economic reasons,
with very modest installations of low capacity - which turned out to be
relatively inefficient. There is thus a dearth of information for installations
with high capacity, more efficient, flushing systems:
39
EVACUATION OF SEDIMENTS
had the benefit of the field data in order to check their validity. Simulation
techniques can thus be used to:
This holistic approach, using field data, models and fundamental knowledge of
the sediment transport process, has enabled a clearer understanding of the
requirements for efficient flushing systems to be developed.
40
FACTORSWHICH !NFLUENCE FLUSHING
'. (a)
(c)
Figure 3.1. Longitudinal profiles during flushing: (a) flushing with full drawdol,w1.;
(b) flushing yvith insu:fficient drm·vdown; (c) final conditions after a long period of
flushing with insufficient drawdown
41
EVACUATION OF SEDIMENTS
sufficient to scour and remove sediment. This raises many technical, economic
and environmental issues. -.,
• The shape of the reservoir may preclude the formation of a scouring channel
that is capable of removing significant quantities of sediment.
• River discharges may be insufficient to transport large sediment loads through
the reservoir.
• The drawdown of reservoir level reduces the capacity to generate power and
the release of high volumes of water for flushing may also reduce the annual
water yield from the reservoir. - \
• . There is a need to be able to predict ahead so that the flushing operation is not -Ij
undertaken if it may jeopardise future power or irrigation supplies.
..... The environmental consequences of passing sediments, that may have been in
I
, the .reservoir for .s.ome considerable time, to the downstream reach. ~ !
-I
!
(1)
where Qs is the sediment transporting capacity (tis) of the flow in the incised
flushing channel, n is 86400 (seconds per day), Tf is the duration of flushing
(days), N is the interval between flushing operations (yrs), Min is the sediment
inflow rate (tiyr) and TE is the trapping efficiency of the reservoir.
42
J.
FACTORSWHICH INFLUENCE F.LUSHING
Efficiency of flushing
More generally, the non-equilibrium state of sedimentation in a reservoir can be
expressed as a sediment balance ratio SBR, the ratio of sediment flushed to
sediment depositing, which can be defined as:
(2)
(3)
43
EVACUATION OF SEDIMENTS
The coefficient, '1', varies with particle size, the larger the sediment size the
smaller the value of W. It follows, therefore, that fine sediments are more easily J
flushed from reservoirs than coarser sediments.
2. The longitudinal energy gradient, S.
This is a parameter which depends on the degree of draw down at the dam
during flushing. The maximum energy gradient is obtained when the drawdown
is maximum and under these conditions it approximates to the slope of the
original river bed prior to impoundment. It is a parameter which can be
controlled by the choice of the amount of drawdown and which influences the
efficiency of flushing operations because of its effect on sediment transport rates,
see Equation (3)~
3. The bed width of the incised channel, W. _.0
Equation (3) requires the bed width of the incised channel, W, to be input.
Channels formed by flushing in reservoir sediment deposits correlate well with
flushing discharge. Figure 3.2 shows the relationship and the data from which it
was derived. The fitted line is described by the equation (in SI units):
(4)
In some cases, channel bed widths may be constrained by the reservoir width:
In general, though, the width of the incised channel is determined by the flow and
is independent of sediment size.
• Sanmenxia
+ Guanting
E @ Guernsey •
~ EB Baira
.s=
:0
+
.~
1i5
c:
c:
ctS
..c:
()
100
•
"'0
~
<l.'l
(/)
.0
0
10~--~~~~~~--~--~~~~~--~~~~~~----------
1 10 100 1000
Flushing discharge, 0.: m3/s
44
FACTORSWHICH INFLUENCE FLUSHING
/II the need for the operation in the ~ght of the success of previous years'
operations
Ii hydrological forecasts for the follovving months
e likely demands for power and irrigation water supplies for the following
months, taking into account alternative sources of supply.
depends on the volume of water in the reservoir which in tum depends upon
reservoir water level. 'Vater levels in most reservoirs, particularly those used f6r
annual water storage and supply, vary throughout the year. The flushing of
sediment demands large fluctuations in reservoir water level and hence the
application of Brune's (1953) curves does present some problems.
45
EVACUATION OF SEDIMENTS
of the inputs to repeat predictions. For example, a low SBR may imply flushing
should be performed at a time of higher discharge and a high SBR may imply that
flatter slopes can be expected upstream from the dam (as shown on Figure 3.lc).
If a value of SBR well above 1 cannot be achieved, then flushing is not feasible.
In the comparison with data presented in Table 3.1, SBR was computed for a
slope defined as the drawdown water surface elevation below maximum water
surface elevation divided by reservoir length.
Table 3.1. Application of sediment balance and long-term capacity ratios to existing
reservoirs
Reservoir Country . Initial Long-term capacity ratio Sediment
. capacity: (LTCR) balance ratio
M_m3 (SBR)
46
FACTORSVVHICH INFLUENCE FLUSHING
In the list above only the side slope steepness is not known. In well-
consolidated sediments, near vertical channel sides can occur, while slopes as
low as 2·5% have been observed for poorly consolidated material. Therefore, a
technique to predict this slope is vital to a reliable prediction of sustainable
capacity.
Atkinson (1998) recommends, with some reservations, the use of the
following expression:
,..,.,Ian a. = 06'"1
. .Jp4,7 (5)
'where CL is the angle of the side slope (zero is horizontal) and p is the density of
3
the deposits expressed as weight of dry material per unit volume (tlm ). p ca.TJ. be
47
EVACUATION OF SEDIMENTS
. ---I
-I
I
1
-II
i
1750 i
E
C
I
~I
0
~
>
Q)
Q5
~l
"C
Q)
a:l 1745
I
1
-I
1740~----------------~-------------------L--------------__~
I
!
o 100 200 300
(a) -\
(Before flushing) I
I
325 I
J I
E -i
320 Sept. 1973 - -
C I
0 (after flushing) i
~ I
~
Q5 -I
"C
m
Q)
315
-I !
r)
310
~
~I
Aug. 1960
(at end of
construction period)
305
0 1000 2000 3000 4000 - )
Lateral distance: m
(b) i
1
~.
Figure 3.3. Cross-sections of flushing channels: (a) Heisonglin reservoir, China;
(b) Sanmenxia reservoir, China
predicted from the composition and age of the deposits using Lane and Koelzer's
(1953) method. Atkinson (1998) found that slopes computed using Equation (5)
could be in enor by as much as a factor of ten and clearly this method needs to J
be treated with caution.
A simple criterion for assessing sustainable reservoir capacity can be
developed by fitting a simplified reservoir shape as shown in Figure 3.4. A cross
section just upstream from the dam can be taken as representative of the entire
reservoir, and then the area of the trapezoidal flushed section can be compared to
the original cross-section area. The ratio of these areas then gives a long-term
capacity ratio (LTCR), which is an estimate of the reservoir capacity that can be
sustained in the long term by flushing.
J
----~.
: ----------------------------------~----------------------------
48
J
FACTORSWHICH INFLUENCE FLUSHING
(a) (b)
.-/-
------
Section
"'C7
(e)
Section
""=7' J
(d)
~\ B_~ I I
__
,-- - - - - - - - -
_ _ _ Water level during flushing
lA'
),t
--It- .
"I"
Bed width, WbO!
Area B
Area A plus Area 8
49
EVACUATION OF SEDIMENTS
If the long-term capacity ratio, LTCR, is greater than 0·5 then flushing is likely
to be successful in terms of maintaining live storage in the reservoir and is likely
to be economic if the shortfall in generating capacity during the flushing period
is not too severe.
50
r'"
(6)
\Aihere Hfjush is the height of water at the dam during flushing and Hma,'( is the
maximum height of water at the dam (to reservoir retention level), both measured
above original bed-level. DDR less than about 0·7 indicates some degree of
constraint owing to insufficient drawdown.
., the flushing outlets may be too small and may restrict discharge
., the river discharge may not be adequate at the appropriate time for flushlIJ.g
e the water may have to be retained in the reservoir for needs in the immediate
future, thereby restricting the duration of the flushing operation.
Under these circumstances the amount of sediment enteIing the reservoir will
exceed that which is removed by flushing until the storage is reduced to the point
where a new, but unsatisfactory, sediment balance is achieved. The retained
reservoir storage will be low compared with its original value and it will have a
much lower natural trapping efficiency to match the constrained flushing flows,
see Equation (1). .
For a given flushing discharge, Qf' the maximum rate at which sediment can
be flushed from the reservoir occurs under the conditions of maximum
drawdown. The sediment balance ratio llsed to assess the adequacy of flushing
flows should t.l}erefore be based on this maximum drawdown. This specific value
51
EVACUATION OF SEDIMENTS
of the sediment balance ratio is designated SBR d and is calculated using the
original river slope, that is for conditions of full drawdown. SBR d < 1·0 indicates
a constraint due to the inadequate capacity to flush sediment.
The amount of sediment removed during the flushing period depends, for a
particular reservoir, on the flushing flow and the duration of flushing. The relative
importance of discharge and duration are illustrated in the following example.
By using Equation (4) to determine flushing channel widths and then
substituting in Equation (3), this yields values of sediment discharge rates, Qs'
The total quantities of sediment removed during the flushing period are then
obtained by considering the duration of flushing. Table 3.2 shows an example of
the results from this type of analysis. It is based on a notional reservoir with a bed
slope during flushing of 0·0006 and a flushing water volume of 8·64 x 109 m3• It
is assumed that the reservoir contains 0·1 mm sand and hence the coefficient in
Equation (3) is 100 (Atkinson, 1998).
If there is a restriction on the quantity of water available for flushing, it is
clearly better to use a high discharge for a short period than a low discharge for
a long period. This increases the amount of sediment that will be removed. The
penalties of extended flushing periods are considerable, quite apart from the
Table 3.2. The relative importance of the discharge used for flushing and the duration
of flushing for a particular volume offlushing water
Flushing Flushing Mean sediment I Sediment Sediment removed
discharge: duration: concentration: removed per during flushing
m 3/s days I ppm day: period:
Mt Mt
310
I
52
FACTORSWHICH INFLUENCE FLUSHING
Reservoirs that. are too wide for the flushing channel to reach tie perimeter
The natural top width of the flushing channel may be less tha..TJ. the representative
top width of the reservoir and under these circumstances accumulated sediments
vviU remain along the perimeter of the incised channel forming a high level
terrace.
If the top width of the section scoured by L.~e flushing channel is not restricted
by the reservoir sides, then the top width of the flushing channel is a constraint.
top width ratio is defined as:
53
EVACUATION OF SEDIMENTS
Table 3.3. The influence of sediment size on the amount of sediment removed from,
reservoirs
Flushing conditions Sediment removed during the flushing period: Mt
By taking the example given in Table 3.2 and by considering the specific
combinations of flushing discharges of 500 m3/s and 5000 m3/s for 200 days and
20 days respectively, the effects of sediment size are as given in Table 3.3.
The sizes of the sediments deposited in reservoirs are an important factor in
deciding whether flushing will be effective.
(9)
Table 3.4 presents the relationship between this ratio and the number of days
of flushing required annUally. It is assumed that flushing is performed at a
discharge of twice the mean annual flow, as recommended above. Full drawdown
of water levels is also assumed and the calculations were performed for a series
of assumed sediment size ratios, SSR. In preparing the table, the updated Ackers
and White sediment transport predictor, Ackers (1993), was used and the river-
bed material was divided into ten fractions. The equation derived by Tsinghua
University and reported by IRTCES (1985) could not be used as it cannot be
applied to sediment transport rates in rivers.
There is no universal relationship between the sediment size ratio, SSR, and
the number of days of flushing required annually because the reservoirs
compared are of different sizes. If a general rule is to be made, then it may be
suggested that rivers where the SSR is less than about 0-03 are generally suited
for flushing.
54
FACTORSWHICH INFLUENCE FLUSHING
It should be noted that flushing has been proposed at the Tarbela reservoir, but
not, to our knowledge, at the other two reservoirs. The Tarbela and Tungabhadra
reservoirs are large (>3000 M.m3), while the planned Rooiport reservoir is about
800 M.m3 •
The parameters required to determine the SSR can be derived as follows.
• DSOT: the 50 percentile. size of the sediment in transport in the river can be
obtained, if possible, from sediment sampling during periods of high
river discharge or (if the reservoir has been constructed) from samples
taken from deposits. Otherwise values can be obtained either from
estimates derived from ·other rivers in the region or by prediction using
the bed material grading .
• Dsos: the 50 percentile size of the river-bed material can be obtained from
representative bed material samples.
In each case, silt and finer material can be excluded as it is usually not a
constraint to a sediment balance.
0-2 38 58 84
I
0-1 23 30 58
I
0-06 15 Ii 21 46
0-05 13 19 43
0-04 11 17 40
0-03 9 . 15 36
I
55
EVACUATION OF SEDIMENTS
The sediment sizes will also affect where in the reservoir the· material will
settle. Some of the silt, the sand and the coarser material tend to deposit in a delta
at the upstream end of a reservoir, while the finer silts and the clay can deposit
throughout a reservoir. After a period of flushing an incised channel will be .
formed in the deposits in the delta, which will quickly refill with incoming sand
and coarser sediments. Thus, most of this coarser material will be flushed from
the reservoir when the incised channel reforn1s during the subsequent flushing
operation. In contrast, the finer deposits formed nearer the dam will be
distributed across the reservoir, and so will be only partially removed by the
flushing of an incised channel. The impact of this process will be to extend
reservoir life at sites with less fine sediment, even when the long-term capacity
achieved by flushing is quite smalL
Table 3.5 quantifies this. effect. It gives predictions of increases in reservoir life
due to flushing for various values of the proportion of sand and coarser materials,
P sand ' and for flushing discharge at the three reservoirs listed in the Table 3.4. Psand
is defined as the proportion of the liver sediment load that consists of sand and
coarser material. In each case, the following assumptions were made:
Table 3.5. Effect of the proportion of sand and coarser material on extending the life
of reservoirs
PropOition Qllushin/ Qmean Factor by which reservoir life is extended
of coarse
sediment I Tarbela Tungabhadra Rooiport
56
FACTORS WHICH INFLUENCE FLLJSHING
Sensitivity to the exponent in this equation was slight (about 5% when the
exponent \-vas doubled to 2.4)
$ the silt deposits downstream from the main sedimentation delta (this is a
conservative assumption, if a proportion of the silt is known to deposit in the
delta then that proportion can be included in P sand)
Q a sediment balance is achievable.
The analysis technique outlined in Atkinson (1998) was used and the assumed
proportion of sand and coarser material, P sand ' was varied from 0·2 to 0·9 in each
case.
These results indicate that where a large proportion of the material deposits in
the delta, say P sand >O·8, then flushing for 30 days annually can greatly extend
reservoir life. This would apply even at sites where flushing does not produce an
acceptable reservoir volume in the very long term. Sensitivity to other flushing
periods "vas found to be slight, for example reducing the period to 10 days only
reduced the factor by which reservoir life is extended by between 2% and 120/0.
(It has been assumed that a sediment balance can still be achieved with L.ie
reduced period of flushing.)
Summarising both the analyses presented in this section provides the
following conclusions .
57
EVACUATION OF SEDIMENTS
The results in the second part of Table 3.6 indicate that at two reservoirs, Ichari
and Shuicaozi, changes to the outlet structures at the dam could potentially
remove all constraints to successful flushing, while at the other reservoirs, site
conditions constrain the success of flushing.
58
FACTORSWHICH INFLUENCE FLUSHING
level must be held constant at as Iowa level as possible throughout the flushing
period. To achieve this:
&Reservoirs ~vvhere the annual run-off is large compared with the volume of the
reservoir are suitable for sediment flushing.
o Reservoirs where L.~ere is a regular annual cycle of flows and· a defined flood
season are suitable for sediment flushing. This favours sites in monsoon areas
and sites where flood flows are generated by annual snowmelt in the spring and
summer months.
• Reservoirs where the release of significant quantities of water for flushing does
not significantly affect t.he ability to satisfy water demands at other times of the
year.
Flushing discharges of twice the mean annual flow are recommended and the
quantity of water required for flushing is unlikely to be less than 10% of the mean
annual run-off. This is based on worldwide experience from reservoirs which are
being flushed on a regular basis, see Atkinson (1996, 1998), Basson and
Rooseboom (1997a and 1997b) and Mahmood (1987), together with detailed
numerical modelling of proposed flushing systems, see Attewill et al. (1998) for
example. Note:
• flushing discharges of at least twice the mean annual flow are required
e flushing volumes of at least 10% the mean annual run-off should be
anticipated.
(j) Graded bed sediments produce conditions which are the most conducive to the
efficient fiusr.ting of sediments. Such conditions are typical of gravel rivers
with a varying bed material composition. In large rivers this situation is found
where the longitudinal bed gradient is between, say, 0·001 and 0·002. In
smaller rivers the eauivalent
1.
range
...
may between 0·002 to 0·005.
59
EVACUATION OF SEDIMENTS
.. From the point of view of sediment size alone, delta deposits of fine sand and
coarse silt are the most easily flushed. Coarser material is difficult to move and J
tends to deposit at the upstream end of the reservoir. Finer material which
deposits in the body of the reservoir outside any incised channel will not be
available for reworking during flushing.
o In the lower reaches, reservoirs are likely to have inundated areas that have
previously been flood plains and these areas would not be reached by the
incised flushing channel which is inevitably of limited width.
Q The longitudinal slope available for the flushing channel is relatively small,
thus limiting the amount of sediment transport.
o Reservoir volumes in the lower reaches are likely to be larger compared with
the mean annual run-off and hence water availability becomes a restraint on
sediment flushing.
61
EVACUATION OF SEDIMENTS
J
1
1
I
62 1
I
1
I WI
e
len e
G
..
0
4. Worldwide experience of sediment flushing
4. J. INTRODUCTION
sy'stems?
/I what downstream effects occur?
The answers to these questions will allow general findings to be made with
regard to the success (or failure) of current flushing operations, the factors that
influence the outcomes, and will provide an indication of the trends in the
designs for flushing systems.
The principal method of appraising worldwide experienGe_has been a
comprehensive literature review, drawing on the reference lists of previous
reviewers and including fresh searches of library references available orieD and
the Internet. Appendix 3 contains descriptions of a number of case studies, where
sufficient information has emerged from the literature searches. For each case
study, the history and physical features of the reservoir are described, the
sedimentation evidence reviewed and the flushing measures which have been
implemented are described.
In the overall stock of dams worldwide (over 40 000 with dams higher than
15 m according to Morris and Fan, 1997), flushing, in one form or another, must
have been attempted in many hundreds, probably th"Ousands of dams.
Unfortunately, the amount of accessible documentary evidence amounts to only
about 50 cases, with substantial quantitative and qualitative data readily available
for only about half of these.
Inevitably, the degree of science applied to the design and execution of the
flushing process must have varied considerably, while the degree of success
would depend on factors such as:
65
EVACUATION OF SEDrMENTS
- i
.j
~
• whether the reservoir and dam were designed taking account .of local
sedimentation data, with facilities to enable flushing to take place
• the fundamental suitability of the reservoir and dam for undertaking successful
flushing
• the degree of operational flexibility to allow an effective regime of sediment
fl ushing to be undertaken as needed
• the application of sufficient know ledge and experience to allow the optimum
flushing regime to be developed.
J
The physical factors, such as the hydrological setting, reservoir basin
geometry and outlet pipework elevation and discharge capacity, that influence the
suitability of the reservoir for successful flushing are discussed in detail in r
I Chapter 3.· A few of the main points are given below, as these help to explain J
some of the experiences of flushing performance.
4.2. FLUSHING
66
WORLDWIDE SEDlMENT FLUSHING
4.3. I. Overview
Table 4.1 lists the reservoirs for which the literature search has revealed evidence
of flushing, although in a few cases it appears that SOlne form of sediment routing
(sluicing or density current venting) may be the major method of sediment
discharge. Excluding those cases where there is no hard evidence of flushing,
leaves 50 cases, for which the locations and purposes are summarised in Table
4.2. In many cases, the purpose is not provided and in some cases multiple
purposes apply, so that the total numbers are not equal to the sums of the
purposes. (In a few instances where the reference cites two or three reservoirs
operated in or parallel for flushing, only a single case is included in this
table.)
By far the greatest number of examples is in China, but this is not surprising,
because of the size of the country, the numbers of reservoirs (18 800 dams higher
than 15 ffi, according to the 1998 ICOLD vVorld register of large dams), and the
high sediment yield, particularly in the basin of the Yellow River. It is notable,
however, that 42% of the reservoirs listed in Table 4.1 in China, which contains
52% of the dams higher than 15 m and 30% of those higher than 30 m (ICOLD,
1988).
"". . . ,............. , it may also be noted that the majority of the examples are from
........ ,..".............. . ., with high sediment yields.
Attempts to relate L;e number of examples of flushing in different countries to
their stock of large dams and their typical sediment yields would not be fruitful,
because of the relatively small sample sizes in most cases, together with a
number of other factors which come into play, such as:
G> t.1.e resources - fin~ncial and technical available to those countries for
researching and dealing with sedimentation problems
• policies for open dissemination of the lessons learned from sedimentation and
flushing experience
11 the financial resources to allow attendance at international conferences
• languages in which technical papers might be written or receive
publication.
Of those flushed reservoirs for which the purpose is k..l1own, Table 4.3 lists the
numbers falling into each purpose or combination of purposes.
67
EVACUATION OF SEDIMENTS
,I
Sediment removal
<.)0-
Reservoir/dam Country ~
~
I:l..
.;:.'"
~
~~
.~
..,
""0
0
::lE
~0
""0
;::
!:
c
0
jL
3 .§~
c c
E-
'00:
~~
:l~
~.g
U I
Principal
reference
Comment
i 0 Q '" ,t:tl E
Baira India H 1981 0·1 F i F E Yes J&K(1984)
Barenburg Switzerland F F 17 M&F(1997)
Bajiazui China F I M&F(l997)
Cach! Costa Rica H 1966 J.6 F F F Yes M&F(I997)
Chirurt ' Former USSR F 25 , M&F(l997)
Chiyu I China F M&F(l997)
Dalingkou China L F M&F(1997) Parallel resrs
Dashidaira ' China 1985 F I M&F(1997)
Dashikau China F F EI 1
M&F(1997)
Donfanghong China F F 12-18 I M&F(l997)
I i
Fergoug Algelia F F Qian (1982)
Ferrera Switzerland -. F F I 38 M&F(l997)
Gebidem --; ... ~wi cierI and H 1968 2·) F F L F 17-21 i Yes M&P(l997)
Genshanpei China F F :
11 M&F(l997)
Groiind , Austria H 1945 0·7 F F : Yes i R&S (1982)
Guanshan China ! F I
La M&F(1997)
~
1
Grimsel H F 1 i SNCOLD (1982)
Guanting FWH 1953 182 FD P Yes B&P(1986)
Guernsey HI 1927 4·3 F P 5880 Yes J&M (1963)
Heisonglin China : IF 1959 61 FRD F iF La Yes M&F(l997) I
=±1
Iril Emda i Algeria 1953 76 D UNESCO (1985) I
68
~.
Algeria 1 1
Austria 1 1
I I
China 2 4 5 15 21
I I
Costa Rica 1 1
Former USSR 2 1 2 4
II I
I
Guatemala 1 I 1
India
I
I 3 3
Iran 1 1 1
l
i
I I
Japan 1 1
I
New Zealand 1 1
Pakistan 1 1 1
Puerto Rico 1 1 2
I I
Sudan 1 1 1
Switzerland 3 2 5
I I
Taiwan
Tunisia
US A 1 1 2 3
I I I I
Venezuela 1 1
I
Totals 19 11 5 25 50
, I
cases, five were considered to involve a degree of sluicing and in five density-
current venting was considered to be an important contributor to sediment
removal.
Table 4.4 Slll1lll1arises worldwide evidence regarding the flushing techniques,
covering whether the drawdown is partial or complete, the season when it is
undertaken and whether it is enhanced by techniques such as lateral channels or
a fluctuating pool level. The total sample number for the amount of drawdown
and the flushing season is taken as the number flushed in Table 4.1. In all cases
where a flushing season is given, it is also stated whether the drawdown is full
69
EVACUATION OF SEDIMENTS
Flood control
Multipurpose
Total 25
70
WORLDWIDE SEDIMENT FLUSHING
by flushing are silts and fine sands. A proportion of finer materials nlay either be
discharged with the water passing through the impoundment under normal
operations~ or may be discharged as a density current.
Coarser sands, gravels and cobbles are likely to be deposited in the upstream
part of the reservoir basin. Given suitable flushing (or sluicing) conditions, these
can be drawn down into the lower part of the impoundment and ultimately
discharged downstream, but this tends to be a longer-term process, generally
associated with a permanent rise in the form of a delta and braided channel at the
upstream end of the basin and with associated permanent loss of reservoir storage
capacity.
4.4.1. Summary
Over 20 case studies are included in Appendix 3. These appear in chronological
order of constnlction completion or first impounding, in order, to some extent, to
illustrate the development of knowledge in sediment problems and remedial
measures, including flusriing. The salient features and key findings from the case
sIDdies are given below, in the same sequence. Table 4.5 summarises the mrun
descriptive and quantitative information on these and a few other reservoirs.
Table 4.6 presents the main quantitative infonnation for the case studies only,
including the key ratios concerning reservoir volume, annual inflow and sediment
load and flushing discharges and volumes, where available. Also included in
Table 4.6 is a subjective assessment of whether flusPing has been successful,
'which is discussed furt.~er below.
71
Table 4.5. Detailed list of reservoirs subject to flushing
Reservation/dam Country Purpose Year built or
sll\rlcd
Year
modified
I Calchmenl
i area: km'
Max depth: m Surface
area: km'
Basin
1enlllb:km
Original
capacity:
Annual inft ow:
M.m'
I AM",'
sediment
Designed for flushing?
Baita India Hydropower 1981 5] (dam height) 4-1 24 1990 I .0.3 Yes, via diversion tunnel
I
35 (diversion I (Atkinson, (Atkinson,
lunnel) 1996)3500 1996)
(J984ref)
Cachi Costa Rica Hydropower 1966 785 1500 (appro~.) ()'81 Single bottom oUllel suitably
I 69 3·24 6 54
localed
Gebidem Switzcrland Hydropower 1968 1996 200 lB 14 I 9'() 420 0·5 Yes, two flushing lunnels below
I bydropower 'nlake
Gmund Austria Hydropower 1945 1967 156·3 311 0·J24 0·94 0·93 135 0·210·7 Yes
Gaunting China Flood control 1953 43400 43 229 30 2270 i 1250 73 (19505) Generously-sized bOllom outle~
7 (19805) but flushing not practicable
13 (B&P because of downstream impacts
I value)
!
Guernsey USA lITigation & 1927 42000 29 9·6 (1963) 24 91 2100 J.7 No
hydropower (1927-57)
I
Probably, as lhe disch.tEe capacity
of]O mS/s is much grelltctlban the
mean inflow
Hengshan China Irrigalion & flood 1966 163 65 I'() 13·3 15·8 1·18 IYes, although main flushing outlet
control is 15 m above original river-bed
Honglingjin China Irrigalion (probnbly) 1960 1364 <42 16-6 43·2 0·77 Probably, as lbe discharge capacity
I I
of 28 m'/s is much greater Ihan Ihe
mean inllow
Icbari India Hydropower 1975 37 II 11-6 5300 5·7 (1976-84 Apparently not for basin, but
!
mean) 2·2 lIushing provided for sediment
(median) exclusion at inlBke
Jensanpei Taiwan Waler supply (sugar 1938 1955 10·6 7·0 Flushing gallery added 1955
cane)
Khashm EI Gima Sudan lITigation, hydropower 1964 950 84 Seven 7 m )( 7·3 m bottom outlets
& water supply controlled by radial gales
Loiza (Carraizo) Pucno Rico Water supply 1954 534 23 2·7 9 27 449 0·38
i
Mangahao New Zealand Hydropower 1924
! I No, upstream reservoir buill as
sediment trap
I
!
Plllagnedru Swil.7.crland Hydropower 1952 1974 138 55 2·6 5·5 3041 (199 in 0·08 NO! known, but bollom outlel
SINCOLD available
I I
1982)
Sanmenxia China Muhipurpose . 1960 1966-71 & 688400 55 120 9640 43000 1600 Yes, bu! facilities bad to be
1990 (m340) upgraded in Stages
Sanlo Domingo Venezuela Hydropower 1974 None 427 65 Approx.0·2 -[ l'() 3'() 450 ()'2 Yes, apparently
Sefid-Rud Irrigalion (primary) & 1962 56200 82 25 1760 5000 50 Probably, us thare are five low-
lImn level outlets
hydropower
!
I
Shuicaozi I China Hydropower 1958 28
" 9·6 514 ()'63 No (no bottom outlet)
Zilcnziliang China lrri~li"n (prohably) 1958 1740 <14 36-6 53-8 342 Protr.bly. as the discharge capacity
of 57 m'/s is much greater th..... the
I I mean inflow
Mt::m mit: ~ri~inall\' t!$limalt:d a.__ 0'0'12 M.m.l/vr (.lbmn 4% Used Llivcniiun lUllltcl. ck::uin!! O·J8 M .m·' in 40 huurs Yes. :t((c:rrm.'hlern~ A6.1 ~5
o(ori~inal 'l';r"~e): bUI ()·45 M.nr' ao:urnulalcd in firsl I~ interrupt ion lOgo:nt:rJliun~ ann"'utli nashing P(\'Poscd futlnd DETR
mCf':lhs thcrt!aftc:"
E;ilim;JlI.:d lhat 18% nUW! thtl~uyh wiltKJut deposition. 54~ Commenced 1973 :lnd undcruIkcn I~ [imc~ in 13 \lars. JM''''''' (19921: R:unfre7. & RUdri;u.:7.\I99Z): M".,.is & Cil 19
pas~c:; by dCtUiit!l current "cnlin~ and 18% dcposill!L.! reduced lr.lppinl:: i'rt1m 8~~ [('I ~7C;r. (fi~,t1tt:.'i on Id:t nc~d Fan 11~9il DETR
da.rifyin~ - ufter nushln~ SI:utr.:L.!'.')
Virtually nu sccJimcnl .3CCUI11UI:ltiun. h..:C:lU~C of gOl".;e:·lyre R~:;.crJl1ir cmpltcJ flu!-J. day~ per yc:v ;,md about J ~·t.nl" Phy!\il:al DUWUl1s ",,,I. (1982): SNCOLD (1982), UNESCO 1I9H5): A6.! 99
1;l!omt.·try J.nd annual tlu~h.in:; of WOlter used; unt!crta~cn rrum ('Iutsct; hypa.'is tunnel Alkin.on 1.1996); Mom, & F:m ( 1997) Ch~1
considered. but ~jcch",-d un C(tst grounds DETR
0·2 Mllyr initially. ".'ducin~ to 0·07 Mllyr ufter upsl=m Flusbing undeOUlken intermillCnlly 194:1-60: annuully Yc., Rieno..1 &: S<hnollc (1932): Alkin50n (1996) Ao.3
","ervoir built 1967 Ih."",rLOr DETR
350 M.m' dCpt"ilCd 1953-60: sUbsequenrly m~ny Only one nu..t1ing operalion (1954) reported, removing IRTCES (1985); UNESCO (1935); Atkinson (1996); A7.1 20
up~ueam l"C'ScrYoirs constructet.!. 5ubs(:lntially n:ducing 10% of annu:::al inUow, parlly hy venting clen:oity currt!nt!' Moni, & F:ln 11997); Binnic & Partners (1986) (some dot. DETR
sedimcm inl1ow:,. B&P value for annual scdim1!nl inllow rut incon5istent betw~n source.c;)
recommended for plill1ning pu"",ses in 1986
~9 ·3% ofstorn2C lust 1927-'7 when ~ime"I<onlriburin1:!: AlIen1fued in Four y~ 19'9-62: nc;u ccmsidc::n.=d J.r~"Cki&: Murphy (1963): UNESCO (1985): Mahmood A7.2 26
C'atchm.en( red~t!d from 14000 to 1800 km! - ccnntlmic:li ur en'ectiv~. 0lS rt!covcreti <Q·2% of origin.II (1987): AllUn",n 11996); Moms 8< Fan (1997) DETR
1·62 M.m·' deposilion in fir~t :; yt!Ol'S of op.!r.ltion (6% , From 1962. llc::nsity cum:nt venting and tklod Se3.C;lln Zhan;., 01. (1976): Xia i1~SO): UNESCO (198;1: A7.J
slur"lI" los., p<r year): capadty reduced to j·87 M.m' by sluicing r<:duced lrap efficiency 10 about 15%: lalenll Atkinson (1996); I-Io.,.is ," Fan (1997) CIt25
1973 erosion IAO<:hniquc successfully implemented from 1980. DETR
ra:ovt!ring some losl storage: long·u:nn c:::1pxity c~pccled
!Q be 30-35% of ori¥inal
3·19 M.m·· dopa,ilcd 196C~73. rcuching deplh of 27 m :II Emplied & nu..hod ror 37 days in 1974. removing O·M IRTCES (19S5): UNESCO 119S'): Atkin.'iOn (1996): AM 77
dom M.mJ of depo.sits: 52 days in 1979 removcd 1·03 M.m" Moms & F:tn (1997} (su,,"..: d~b inc,,"si~lent bt:"tween DETR
I SOLH't.~)
I
0·57 M.m"' deposited per year 1960-63 in impound!ng \Varer 1evel lowered in nood lensan, resuttin!! in suh"t~\ntial ZI,.ng er al. (1976): IRTCES (1985); UNESCO (198;):
mode. representing 3·5% s[omgc loss per ye:u- reduction in r:ue of "crag. In.'. 10 0",,' M.m; p<r ye:" Alldnson (1996)
1~6.J.-7J: technique is e.~5entially routing/!iiiluicing
S.:c!imcnlO1tion reached spillway crest after one year; 85% Flushed annually by fully-op.ning spillway gaI.s: Bharg.Y:1., al. (lq87): Mohan ., al. (1982); Alkinson A7A 36
tr:::apping much gre-Jler (h:L"l indicated by Brune's curvcs problems with :abr.1sion dumage to spilhv:::ay ~d roller (19961 DETR
( '.~,!): .lllLicipatcd long.lcrm c:1pacit:{ :wout 3S%
SLOr:J~t: loss ~·26 M.m' 1938-55. repre!>enting 3-4% per Flushing commenct:!d 1955 for 1·5 month." aa1l1uillly, Hwong (1985); Paul.l:: Dhillon (19881: Monis & Fon DETR
year virtuaity mesting ~ub~qucnl sedimcntarion. hut not (I'1~7)
re.... toring c3f':Jcity. minor mising of im['lOunding level :1bout
1942 and 1958
I
Cap~ity seriously de~ie1l!d Flushing opo<"Jtions in 1971 .nd 1973 c.ch ",moved 85 Mt UNESCO 1198S): EI H.gToyeb (1980): El FailhSiI:ld DETR
(1980); CFGB (J 973 '" 1982)
L..:.lS{ 53t':~ ot' ..:a9:.u.:i(y lCJ5J-~~~ :.hrcc {lOU mm low·lcv~J Mechanic:Ji mt:thods auemptcd unsuccessfully in 199J.; HEC6 (Morris & Hu. Webb &: Soi~r·L.orc7. (1997): Morri • .l; Fun (1997) Ch2Q
aw:le.s blocKctl ~...'!!im"nt
pass-through planned and .,petted to <educe 1992)
.~cdimc:H.:ltk'll':;
dredging con.;idered in (995; \cchniq!J~ :s
I cssctuially r,outing/sluicing
59% of Slara!e 1o", by 1958: prublem inc,••• in~I)· ",riOu..
by mid-I 96C.'
I Flu.hod in 1969 through low·l".,.cl dive"ion tunnel anLi
75% of accumulated .!<diment removed in one monlh;
JOW"lr (1984); Alkinson (\ 996) OETR
CilPllCity reduced
about B<X~ by ""dy 19iO..
LO iiOOUL 50% hy ! 950. bUl :ccuvcr:t.! to I
1
ilo!totn oulict> ung.l<:ci prior to 1970.• 0 fl\1Sllinll appears
to have been j);&IUr.:tI
IRTCES (19851; L'NE5CO (19851; A.kinson (1996)
Bod level, nJ>" up to 23 m by 1969, ,edimen. volume ISluicrd for ~ month., annually sine. 1963 IRTCES (1985): UNESCO (19851: Arkinson (1996) A7.5 10
oppe-.Jrs to have stabilised at 30 M.m.' sir:<:c t953 DETR
I .
1973 Hood c;tuscd l·S M.:rr' dC(l<~.liit:on (33% of origin:J.1 Flushing (aioe'!! by m""banic;ll pl.nt) II/iS-03179 Lie::hi & Hacoerli (I 97O'J: SNCOLD 119~2): AtKinscn A6.3 100
.slor~f!} and .liuhnK:~c.:d
bmtom \lur!el rcmovc:u 2,.", ~I.mJ: 1760 m 10n1: scdiment bv!)a.5.S tunnel (1996) OETR
Rehanili!.lLion from 1966 includod con;[ruttion of Inrger Zhang &: long (1980): UNESCO (19851: Atkinson (19961: 11.7.6 39
low .. It:'IIct Out[CL;; flushed for 4 months annually; six Mom, &: F,n (1997) Ch14
Sc',ere. causing less or 2· t % of !.he :.tor.lge c:lp:!clty ~r Flus hing (:::about 4 month"/yrJ commcncdi in 198C: after 7 Toio"i •• r al. (1991): Mahmood (IY'7); Aikin,'on (1996); A7.7 13
year up 10 19S0 (Imp efficiency 73%); mas I nf sedim"", years 25% of lost storage h:r.tJ been reco\'cred: from t 99~ ChD
relc:JSC' occurred in d..:n!ittj' currents fluodplain erosion ~nn:lllced using diversion ciumnel!:: DETR
~."(pe.;:~~d mat long·umn staragt: CJposcity could be up to
90% uf original
8,: 3 ~'l.m·l {~3% d' 5tarJg~) tos~ :~1:::1-81 ; ~eu !~,:c!s ::.t rmp!:::,'!·.:n:et.! ::''tpcrimt:~t::~!y from 19€5; b:.1t !imi!:!c by I i (WeHR i !9~~): IRTCES (19~51: liNESCO (IQ·i 5r I A7.Pt 39
dam oni:- "] m ~dow impoundin~ tc.:\'ct nigh elcv:1tion uf !ipillway and short dl.!rltiun :;u1nual1y lo A'kin,on ( 1996) I DETR
abOUt one third of inflow
I
DETR
f':r yt:lr I
1
4·3 M.m.' dcp0:iitl"C.I
r.1vUl!. rcp(!!..:cntin~
p~ryt::tr 195lJ-6 1 in impoundln~
12% ${Qi.!g..! )o!t; per YC:lr
\\:'alcr !e\lcllowcn:d in Ihmtl.,ea.'ion. rt:~uILin~ in :;ub!:.lambl
;eduction in iJ.tc orstoiJ~~lo~ t~)O · 77 M.m' pc:- ."~".U'
I Zh:;.ng t:( tTl. !!~76j
I·I
I Itjrjl-1J: tC'chnique:s ·.!!i5t!'nt1i!lIy rm.!li;"1gJ~bicir.g
m
~
()
C
~
o
Table 4.6. Summary of key flushing parameters z
o
~
Reservoir Year Capacity: Mean Mean or Flushing Flushing Ratios: % LTCR% Comment ""T1
M.m3 annual flow: median discharge: volume: Vl
m
M.m 3 sediment m 3/s M.m3 CIl SC: Q/I VII Atkinson Current C"-.
Guernsey 1927 91 2100 1·7 125 54 4·3 1·9 188 2·6 26 N Note I
Zemo-Afchar 1927 6600 5 450 30 215 0·5 20 Y? Original capacity not given
Gmund 1945 0·93 135 0·07 25 0·45 0·7 7·5 583 0·3 98 85 Y
Palagnedra 1952 5·5 199 0·08 2·8 1·5 100 100 Y Assisted by bulldozers
Shuicaozi 1958 9·6 514 0·63 50 4·32 1·9 6·6 307 0·8 39 Y?
;;:.,'
-~ -1~",,";,,,j J~"-=",,, -~:
~ -1. - --' -
-l - j - -- - L.~ ---L-J' L-J 1 -' -~
'-
NOles: ~
I. Flushing diseharge given as 120-1L10 ml/s at 12-13 III drawdown; Q and V from Atkinson (1996) 0
;0
2. Value or sediment inflow ('or planning, as recommended in 13 innie & Partners (1986) report; flushing not acceptable due to downstream constraints r
3. Sediment management includes sluicing during 1100d scason and lise of lateral channels I 0
4. Main sediment management technique is sluicing during flood season
5. By using longitudinal flushing channels, anticipated Ihal long-term capacity could be up to 90%
~
6. More inrormalioll may be available in 1980 references by EI Hag (1980) and El Faith Saad (1980) 0
n1
7 . Much greater Ilushing disc\larges possible via a higher outlet (abollt 25% of water deplh above base of dalll) (/I
8. Mllch t~realer Illlshing discharges possible if needed rn
9. No hollom ourlet; llushing via gated spill.way only 0
10. Mean anllual !low taken as mean of two values reported in li rerallll'e 3:
fl1
Z
-I .:0,
11 " . :..
r
C
VI
I
Z
(.)
EVACUATION OF SEDIMENTS
76
VVORLDVv'IDE SEDIMENT FLUSHING
77
J
EVACUATION OF SEDIMENTS
J
about 30% of the original capacity. The sediment released from the reservoir is
used beneficially for agricultural warping.
]
Sanmenxia reservoir (China, 1960)
This is a large dam, built on the world's most silt-laden river, so it was originally J
designed with sediment management measures. However, the assumptions upon
which these were based proved to be highly optimistic and major sedimentation
problems arose very quickly, threatening increased flood risks up to 260 km 1
upstream of the dam. Hydrologically, Sanmenxia is just small enough to be
classified as 'small', with a capacity/inflow ratio of 22%. After about five years
40% of the original capacity had been occupied by sediment deposition. 1
The reduction and control of sediment deposition and the development of a
r
sustainable sediment. management . regime became a high priority for the
Sanmenxia reservoir. The development occurred over six stages between 1962 1
and 1978, and the final technique is largely one of sediment routing by sluicing
through the high-flow season and impounding for irrigation and hydropower in
the low-flow season. Since 1975, the net storage capacity below elevation of 1
330 m (10 m below top water level) has been stable in the range of 50-55% of
the original capacity at that elevation. The lessons learned at Sanmenxia have
guided subsequent projects, including the Three Gorges.
]
Warsak reservoir (Pakistan, 1960)
This is a hydrologically small reservoir at less than 1% of the mean annual 1
inflow, on a river with a bed load which includes gravels and cobbles. The total
annual sediment load is equivalent to an annual accretion of about 8% and, after
20 years of operation, the reservoir was filled with sediment up to conservation 1
level, except for a channel leading to the power intake. There is apparently no
substantial bottom outlet to the dam, and flushing, which was attempted by way
of the gated spillway on five occasions between 1976 and 1979, was I
unsuccessful. The Warsak reservoir has apparently reached a broad equilibrium,
with virtually no residual live storage.
I
Ouchi-Kurgan reservoir (Former USSR, 1961)
The Ouchi-Kurgan reservoir is hydrologically small, with a capacity of less than
1% of the mean annual inflow. Soon after construction it has been flushed 1
annually, but apparently with only a limited drawdown. The volume of deposited
sediment has stabilised at 50-55% of the original capacity since 1968.
1
Sefid-Rud reservoir (Iran, 1962) I
This reservoir is hydrologically large, with a capacity/inflow ratio of 35%.
Sedimentation was a serious problem over the first 17 years of its operation, I
reducing the storage capacity at a mean annual rate of 2·1 % and reaching a
minimum of 63% in 1982-83, before recovering as a result of flushing measures.
The flushing measures comprise emptying the reservoir from October to I
Febluary, outside the irrigation season, then refilling it dUling the early part of the
flood season in time for the strut of irrigation in May.
I
78
L
WORLDWIDE SEDIMENT FLUSHING
r and the potential to absorb over 4% of the original storage per annum. The
reservoir has been flushed annually in the flood season and this has resulted in
virtually the entire storage capacity being preserved. This is attributed to the
r gorge-like geometry of the basin and the steep valley slope. There have been
problems with downstream sediment accretion, where the valley slope reduces,
which were expected to be overcome by deploying greater flushing discharges.
79
EVACUATION OF SEDIMENTS
reservoir was operated without flushing and the total sediment accretion was
about 25%. The first flushing operation, in 1978, with full draw down , was
estimated to have removed 50-60% of this accretion in three or four days. Over
a further three weeks, with assistance from bulldozers, the original storage
capacity was virtually restored, with an estimated 3-5% loss remaining.
It was concluded that, in the future, flushing should be undertaken annually,
preferably towards the end of the high-flow season. It was also considered that
empty flushing should occasionally be inten"upted and be followed by a short
period of pressure flushing to concentrate sediment removal on the immediate
areas of the bottom outlets.
80
WORLDWIDE SEDIMENT FLUSHING
4.4.2. Findings
The findings from this review of t.~e case histories, together with the limited
information on a number of other reservoirs where flushing has been undertaken,
can be considered under the following subhea?ings:
Hydrology. "';'
The hydrology of the catchment needs to be properly researched and understood,
as it is central to the consideration of the other issues which affect the
practicability and likely success of flushing. It is necessary to k.TJ.OW the typical
patterns of run-off within the year, together with the ranges of variations
encountered within the year and from year to year. This information is important
both for 'broad-brush' assessments and for mathematical model simulations of
reservoir sedimentation.
If local fio\v gauging records are inadequate, additional expert hydrological
appraisals will be needed. These might make use of national or regional
hydrological paran1eters, supplemented by techniques such as flow gauging data
transposition (from within or outside the drainage basin) and flow gauging data
record extension, using correlations with longer periods of rainfall records.
The data intervals used need to take account of the size of the catchment,
varying from an hour or less when considering local floods in very small
81
1
EVACUATION OF SEDIMENTS
1
.catchments, up to. perhaps a week fer very large catchments, particularly these !
,.--L
where a large part o.f the flo.o.d flew is go.verned by glaciers and sno.wmelt.
Upstream reservo.irs, that are already in existence o.r planned in the future, can
have the fo.llo.wing effects en water inflews to. the reserveir:
• en the to.tal annual inflo.w, if flews are diverted fer irrigatio.n er transferred into.
ether catchments, o.r subject to. increased evapo.rative lesses in sto.rage
• en the seasenal distributien ef inflo.ws (unless the upstream reservo.irs are few 1
in number and all hydro.lo.gically small)
• en the availability and co.ntro.l o.f flushing flews when they are needed . _1
. In the case o.f Guanting reservo.ir, fer example, the censtructio.n o.f ever 300
reserve irs upstream and increased water use fer irrigatien between the 1950s and
1970s, reduced the tetal river flews entering the reservo.ir. At the Gmund 1
reservo.ir, the censtructio.n o.f an upstream reserveir allo.wed a high degree o.f
co.ntrel to. be exercised ever inflews during flushing.
Acco.unt also. needs to. be taken o.f ether land-use trends affecting the
catchment hydro.lo.gy, such as urbanisatio.n and defo.restatien.
Sedimentology 1
The cellectio.n o.f useful sediment data is a vitally impo.rtant issue, because large
errers in predictien can be made if proper acceunt is net taken o.f the large
variatio.ns in sediment co.ncentratio.n which naturally o.ccur. An insufficiently
1
frequent and rigo.reus sediment data co.llectio.n pro.gramme is" liable to.
underestimate severely the large co.ntributien to. the annual sediment lead - beth
suspended and bed lead - which derives fro.m the highest discharges. Wherever
1
po.ssible, the available sediment data sheuld be tested against experience o.f
reserveir sedimentatio.n in the area, to.gether with natio.nal and regienal data en
sediment yields fer the seils and geelo.gical co.nditio.ns feund in the catchment.
The ultimate ebjective ef the sediment studies is to. o.btain reliable values fer
the mean annual sediment lead and the degree o.f variability fro.m year to. year,
tegether with particle size distributio.ns and to. derive hydro.graphs that give the
1
seasenal prefiles o.f sediment lead which co.mplement the seaso.nal flew
hydro.graphs.
Catchment changes also. need to. be taken into. acco.unt, as in the hydro.lo.gical
studies. Reductio.ns in sediment leads entering reservo.irs due to. the develepment
o.f upstream reservo.irs have been repo.rted fer the Guernsey, Gmund and
Guanting reservo.irs. At the Guanting reservo.ir, fer example, the reductio.n in 1
sediment leads, due to. reserveir and irrigatio.n develepment between the 1950s
and the 1970s, was much greater than the reductio.n in annual inflo.ws.
Co.nversely, fer the Mangahae reserveir, the expected develo.pment o.f additienal
upstream reservo.irs, to. reduce sediment leads, did · net o.ccur. Where the
censtructio.n o.f upstream reserveirs prevides a respite in sediment lead, it sheuld
be remembered that this may be ef limited duratio.n, as the upstream reserveirs
fill with sediment, er as sediment flushing and sluicing measures are
implemented.
82
'vVORLDWIDE SEDIMENT FLUSHING
Storage capacity
The key storage capacity parameter is the ;hydrological size' which is the ratio
of the storage capacity to the mean annual inflow. Table 4.6 lists reservoirs with
ratios of between 0·1 % and almost 200%.
A ratio of 30% may be considered as an approximate boundary between
hydrologically large and small reservoirs. If the ratio is less than about 30%,
there is a reasonable prospect of having sufficient flow available to allow the
reservoir to be emptied for flushing annually, generally in the early part of the
flood season, so that it can be reliably filled in the later part of the flood season.
The smaller the ratio, then the more practicable flushing becomes, from a water
resource standpoint. Smaller ratios, of perhaps 5% or less, allow more rapid
emptying and refilling of the reservoir, and so suit relatively short periods of
flushing.
Su.bject to tlle other constraints (in particular u~e adequacy of the bottom outlet
and the suitability of the basin shape), most hydrologically small reservoirs
appear to have been flushed successfully, examples being Gmiind, Palagnedra,
Cacm, Gebidem, Santo Domingo and Baira. However, this is not a guarantee of
successful flushing if other requirements are not met, such as at Guernsey and
Warsa..1(.
If the ratio is much larger than 30%, then it becomes increasingly difficult to
schedule a flood-season flushing regime that will still meet the water storage
objectives, which generally require the reservoir to be full by the end of the flood
season. Successful flushing has nevertheless been undertaken at reservoirs witJ.1.
a higher ratio, an example being the Hengshan reservoir, aided in that case by its
small size and narrow steep Valley.
o Once the ratio approaches or exceeds 100%, it is clear that an impounding
reservoir (other than one used solely for flood control) is designed for the carry-
over of water from one year to the next, to cover shortages in drought years.
Annual empty flushing of such reservoirs is not possible, but there may be the
possibility of empty flushing once every decade or so, if this is beneficial. There
may also be some benefit in flushing with partial draw down , at the lowest annual
water levels. 0
83
EVACUATION OF SEDIMENTS
Basin shape
Narrow steep-sided 'gorge-like' reservoirs are clearly more amenable to effective
flushing, pruticularly where the longitudinal gradient is steep. To some degree,
this is often offset by poorer mobility of the coarser bed load present at many of
these reservoirs. Several of the hydrologically small reservoirs listed in Table 4.6,
such as Gmund, fall into this category and have been successfully flushed. The
Hengshan reservoir, Which is hydrologically large, but steep-sided and with a
steep valley gradient, has also been successfully flushed.
Broader reservoir basins are vulnerable to sediment deposition over the
flooded floodplains, leading to two problems with flushing:
• when the deposits are exposed during draw down of the reservoir, they tend to
dry out and consolidate
• they are isolated from the flushing flows, which tend to be in line with the
original watercourse, so are not subject to significant erosion.
The operational regime (for example, whether the reservoir is empty or full at
the time when the flood anives) can have an impact on the vulnerability of the
former floodplains to progressive deposition.
As a result, the residual storage capacity resulting from flushing broader
reservoirs is generally limited by the geometry of the channel that can be eroded
by the flushing flows. The eroded channel gradient, width and side-slope angle
are generally functions of the sediment characteristics, valley gradient and the
flushing discharge, but may also be limited by other features of the geometry of
the basin. Guernsey, Naodehai, Heisonglin and Sefid-Rud reservoirs are
examples of broad reservoir basins where the basin geometry acts as a constraint
on the efficacy of flushing.
The flushing of reservoir basins, which include the valleys of tributaries, is
also likely to be influenced by the relative magnitudes of the tributary flows. An
84
WORLDWIDE SEDIMENT FLUSHING
Low-level outlets
The primary requirement, even in reservoirs where empty flushing may not be
desirable for operational reasons or considered likely to be necessary for many
years, is that there should be effective low-level outlets near the bottom of the
basin. \Vithout these, there is no possibility of undertaking empty flushing if and
when required in the future.
If the lowest outlet is at mid-height, for example, this provides a pelmanent
constraint to flushing, limiting it to the less effective partial drawdown. This is
the case at the Guernsey, Shuicaozi, Warsak and Ichari reservoirs.
J.4. common rule of thumb for successful flushing is that the discharge capacity
of L.~e low-level outlets should
. be sufficient to Dass at least twice the meaTl annual
~ .
inflow at a drawdown of the pond level by at least 50%. This drawdown elevation
may allow sediments to be effectively scoured from Ll.e upstream half of the
reservoir length, although some of the coarser material would be expected to be
redeposited in the downstream half. However, this criterion should not be ta.l(en
to imply that only partial drawdown is needed. Full drawdown clearly has the
potential to be more effective.
If the reservoir is hydrologically smail, so that a relatively short flushing
.. period is possible, this discharge capacity criterion is probably a reasonable one
for flushing outside the flood season. In a larger reservoir, where the time taken
to draw the reservoir down would be longer, this discharge capacity may be
inadequate from an operational standpoint, even outside the high-flow season.
There would also be the risk of the drawdown being interrupted and extended by
periods of higher flow, which would also tend to result in further deposition.
For full draw down flushing in the flood season, the discharge capacity would
probably have to be significantly greater than the above rule-of-thumb value. For
t~e design or checking of a flushing system, consideration therefore should be
given to actual hydro graphs at the proposed time of flushing, simulating the time
ta..ken for the reservoir to empty for flushing, then to refill under a range of flow
conditions, together with th~f'robabilities of being able to keep the reservoir pool
at the required level during the flushing period ...
The Sanmenxia reservoir is an example of a reservoir where obtaining an
adequate discharge capacity through the low-level outlets for flushing (or
sluicing) through the flood season was vital for the success of the sediment
managemen t measures.
Operational considerations
The relevant operational considerations regarding flushing are mainly those
associated with the lack of water supplies and/or the reduction in operating head,
as they affect hydropower generation or irrigation supplies. These considerations
vary considerably from site to site. For example, if the irrigation supplies are
abstracted downstream and the irrigation system is designed to accept turbid
water for warping the agricultural land, such as at the Heisonglin reservoir, there
85
EVACUATION OF SEDIMENTS
may be no disruption and even a benefit from flushing. Seasonal demands will
also influence the time when flushing is most convenient and the amount of time
and water which can conveniently be used. At the Jensanpei and Sefid-Rud
reservoirs, industrial water supply and irrigation demands respectively favoured
flushing during the low-flow season.
In the final analysis, of course, operational considerations must not be allowed
to prevent effective flushing to be undertaken, if that is needed for the
preservation of the resource for future generations.
Drawdown
The case histories . show that full draw down to achieve empty flushing is
preferred, but there are cases where this is not practicable, owing to limitations
.!
imposed by the available flushing facilities or from operational considerations.
The available outlets may prevent full draw down due to either their elevation (as
at the Guernsey, Shuicaozi, Warsak and Ichari reservoirs) or because of
insufficient discharge capacity, or a combination of reasons. In cases where a
limited discharge capacity prevents full draw down during the flood season, it
may be possible and more effective to undertake flushing at lesser discharges
during the non-flood season.
Enhancements
Of the enhancements listed in Table 4.1, fluctuating water levels during flushing
have been reported to be beneficial at the Gebidem and Santo Domingo
reservoirs. In practice, the technique has probably been more widely employed
than reported and would be expected to be always beneficial in terms of
enhancing sediment outflows. Fluctuations in flushing discharge have also been
found to be beneficial to the encouragement of slumping failures of the channel
banks.
The other enhancements all involve human intervention, preferably widl
earthmoving plant, on the reservoir ·deposits. In some of the cases where
bulldozers have been used to shift sediment towards the main 'channel and
increase its rate of disposal, the intervention was experimental in the first flushing
and may not be found worthwhile after a regular flushing regime has been
instigated.
The two cases (Heisonglin and Sefid-Rud) where lateral channels, lateral
piping and longitudinal channels have been successfully deployed (and where
sufficient information is available to judge) are hydrologically large reservoirs
with significant areas of deposits over the fonner floodplain. Without these
measures, there would be a significantly worse prognosis for the preservation of
storage in the long term.
Downstream impacts
There have been severe impacts on the downstream aquatic environment in a
number of cases, principally where heavy deposition or high suspended sediment
concentrations affect the habitat and the survival of fish and other wildlife. In this
respect, Sh011 peliods of flushing are particularly problematic. Sediment routing,
86
WORLDWIDE SEDIMENT FLUSHING
by sluicing through most of the flood season, on the other hand, is "much more
benign environmentally.
The potential adverse environmental impacts downstream should always
considered in the light of the alternatives, such as an acceptance of long-term
sediment accretion, requiring the further use of natural resources for such things
as the construction of additional reservoirs or the development and use of
alternative sources of power.
In some cases, reservoir operators are subject to regulatory limits on
downstream sediment loads or concentrations, wrJeh have to be taken into
account in the detailed planning of every flushing operation. Because of the
variable nature of the phenomenon, unless a very large factor of safety is
employed vvith implications for the duration of flushing needed, occasional non-
complial"lce is almost inevitable.
The downstream impacts can be rnitigated substantially if there is dilution
available a short distance downstream, for example, from hydropower releases or
at the confluence wit.~ a larger river, such as in the case of the Rhone downstream
of the Gebidem reservoir. The control of flushing - for example, in response to
downstream concentration monitoring can be aided if inflows to the reservoir
can be controlled, for example, by releases from an upstream reservoir, as at the
Gmund reservoir.
Other downstream interests that may be affected by sediment releases
include:
Ill)reservoirs.
G L~at there should eventually be a balance, over time, between the sediment
inflow to and outflow from the reservoir
Q that the sustainable long-term capacity of the reservoir s.hould be at least a
certain proportion, typically 40-500/0 of the original reservoir capacity.
87
. ,
,
I
EVACUATION OF SEDIMENTS
J
The key test is that a practical compromise should be achieved between the
processes of sedimentation and the requirements for beneficial use of the
reservoir. The long-term beneficial use may be much diminished from what
might have been expected at the time when the reservoir was designed and built,
but in comparison with the poorer or even catastrophic outcome that may result
without intervention, any tangible improvement from flushing must be judged a
success to some degree.
In these terms, therefore, the flushing undertaken at most of the reservoirs
presented in the case studies can be considered successful. Several of these are
cases that were considered unsuccessful by Atkinson (1996), based on the more
restrictive criteria listed earlier. The reservoirs at which flushing is not possible J
or has been unsuccessful are probably limited to the following:
88
WORLDWIDE SEDiMENT FLUSHING
89
ro· hi I
it d to
5. Geographical areas suited to flushing
. (a) Climate
(i) precipitation and run-off
(ii) temperature .
(iii) wind speed and direction.
(b) Geotechnics
(i) geology
(ii) volcanic and tectonic activity
(iii) soils.
(c) Topography
(i) slope
Oi) catchment orientation
(iii) drainage basin area
(iv) drainage density.
(d) Vegetation
93
EVACUATION OF SEDIMENTS
the total sediment load. Regions with less than 50 tJkm2/yr account for about half
of the land area and 2·1 % of the sediment yield. Case studies of erosion rates are
presented in Appendix A4.2.
Table 5.2. Continental variations in sediment yield (Jolly, 1982, taken from Gregory i
Mrica 27 550 J
Asia 600 16160
Australia 45 230 J
Europe 35 330
North America
I
96 1990 J
South America 63 1220
J
94
f
AREAS SUITEDTO FLUSHING ·
approximately 80% of the world sediment total (Jolly, 1982). The largest
sediment yields occur in Oceania at 1000 t/km2/yr including ntllnerous
catchments in. New Zealand, New Guinea and Taiwan with sediment yields two
to three times the world average.
There are considerable differences between the continental figures produced
by the two studies. The table produced by Mahmood (Table 5.1) distinguishes
between yield rates in Australia and Oceania and this leads to the highest
sediment yield rates in Oceanian rivers at 1000 t/km2/yr. This is hidden in the
study by Gregory and Walling (Table 5.2), where Australia is taken to include
Oceania. This produces a higher rate for Australia than found in the study by
Mahmood but a much lower rate than for Oceania. The next highest sediment
yield is produced by Asia with a yield of 380-600 tlkm?/yr. The lowest rates of
sediment yield occur in Australia (owing to aridity) in the Mahmood study, at 28
tfJs:m?/yr, and in Africa, at 27 tlkm?/yr, in the study by Gregory and Walling.
These rivers do not produce the world's highest sediment yields as the list
excludes drainage basins smaller than 10000 km2 which are likely to produce the
highest yields per unit area.
95
J
EVACUATION OF SEDIMENTS
1
Table 5.3. World maximum recorded suspended-sediment yields greater than I
2000 tlkm 2/yr (Jolly, 1982, from Gregory and Walling, 1973) J
River Location Average annual yield:
tlkm2/yr
l
r--'-
Marecchia
...,;. . .
Pietracuta, Italy 4570 J
Semani Urage, Kucit, Albania 4150
I
.---L
valley, and areas of volcanic soils, such as Java, South Island of New Zealand,
Papua New Guinea and parts of Central America (Morris and Fan, 1997).
• The Pacific Asiatic-Australian sector demonstrates the most intensive rates of
erosion. Figures in the range of 10 000 to 50 000 tJkm2jyr have been reported · 1
at stations in China, Taiwan, the Philippines, Indonesia, Java, Kenya, New Guinea
and New Zealand (Walling, 1994) due to active tectonics and volcanism, steep
slopes, high precipitation amounts and intensities, high and irregular lun-off, .
·1
dissected mountain relief composed mainly of sedimentary rocks, and human
influence by agriculture and logging (Dedkov and Moszherin, 1992).
• Taiwan discharges more sediment to the ocean per unit area than any other 1
country in the world. Streams draining the central range produce suspended
sediment yields of 13 760 tlkm2jyr. One small basin exports 31 700 tJkm2jyr f
(Li, 1976). The sediment discharge of Taiwan is nearly five times larger than
that from the continent of Australia, even though it is 210 times smaller. Lower
values of 11-12000 tJkm2 have been reported in Java (Walling and Webb,
1983).
1
• In New Zealand values reach between 20 and 28 000 tlkm2jyr with a mean
value of around 2000 tJkm2jyr. The highest mean annual sp~cific suspended-
j
96
1
AREAS SUITEDTO FLUSHING
Table 5.4. Rates of sedilnent yield for the world's major rivers at ocean level,
excluding basins with an area less than 10 000 km 2 (IYlalirJ'lood, 1987)
Continent Country River Drainage area: Run-off: Sediment: Yield:
million km 2 cm/yr tlkm 2/yr ppm
I I
Oceania New Guinea Purari 0·031 248 2581 1039
I
South AmeLica Peru Chira 0·02 25 2000 I 8000
I
Europe I France
I i
Rhone
I
(·09 I
I
S4 111 204
.' .
97
T
1
EVACUATION OF SEDIMENTS
North America Canada,",. ' ''''M ackenzie " " 1·81 17 55 327
1
North America USA Hudson 0·02 60 50 83 I
~
Asia Iraq Tigris-Eupha 1·05 4 50 1152
Europoean AItic
Africa
USSR
Egypt
Indigirka
Nile
0·36
2·96
15
1
39
38
255
3700
1
South Amelica
Africa
Argentina
Nigeria
La Plata
Niger
2·83
1·21
17
16
33
33
196
208
1
j
Asia USSR Amur 1·85 18 28 160 I
-1
98
AREAS SUITEDTO
sediment yield is 53 SOD t/la.n2/yr for the Huangfuachan river (3199 k1112), a
tributary of the Yellow River in China (Walling and Webb, 1983).
Table 5.5 lists a number of basins with very high yields in various countries,
which were reported in 1983 and which exceed the values of record yields
published in 1973 and shown in Table 5.3. For tributades of the Yellow River,
highly erodible loess, lack of vegetation and the semi -arid climate are the major
controlling factors. The semi-arid climate is a factor in the Kenyan example, but
severe disturbance due to agriculture is also a factor. For Java and New Guinea
steep relief, high rainfall and agriculture are important and in New Zealand the
steep relief, high rainfall up to 9000 mm1yr, and tectonic activity play a role
(vValling and Webb, 1983).
Table 5.5. Values of sediment yield in excess of 10 000 tlkm2/yr (l;Valling and vVebblec
1983)
Country River Drainage Mean annual Source
area: km 2 t sediment yield:
tfkrn'1jyr
II
China Dali 96·1 25600 MOli and Meng (1980)
North Island
New Zealand
[ Waiapu
Waingaromia
1378
175
J 19970
17340
Griffit.~s (1982)
Griffiths (1982)
I
New Zealand Hikuwai 307 13 890 Griffiths (1982)
I
South Island Hokitika 352 17070 Griffiths (1982)
99
EVACUATION OF SEDIMENTS
",
New South Wales, Australia. Values of less than 1 t/km 2/yr have been reported in 1
Poland (Branski, 1975) (from Walling and Webb, 1983). Areas of low sediment
yield are usually flat, arid with inadequate streamflow to transport large sediment
volumes, or arctic regions with low relief, little precipitation and human impact
(Morris and Fan, 1997) or low mountains of the temperate zones that are
underlain by crystalline rocks and covered by dense deserts e.g. Scandinavia, the
Urals, the mountains of South Siberia and the Trans-Baikal region (Dedkov and
Moszherin, 1992). The results quoted for Poland appear to be anomalous.
100
AREAS SUITEDTO FLUSHING
o In northern USA and in Canada there are low sedin1ent yields, ' except in
Alaska and the mountains, in southern USA and along the coast' of California
(Jansson, 1988).
5.2./. Introduction
An understanding of the precipitation regimes throughout the world may provide
a key to the definition of areas of high and low erosion rates. We briefly describe
these in the section which follows. It is difficult to classify distinct climatic zones
as they tend to merge into one another rather than have sharp boundaries but a
number of general models have been produced. These are discussed in Appendix
A4.3.
Seasonal variation
Precipitation distribution during winter (December to February). Figure 5.4 shows
the global distribution of precipitation during the northern winter of 1998.
During this period there were precipitation values of over 100 mm per month
along the western fringe of North America into Canada, a region to the north of
the Gulf of !vIexico, a region of Central America, mainly in Brazil, southern
Africa in the region of fvlozambique, Mada£ascar, Tanzania, Zambia and Angola.
In Asia, high p;ecipitation values- occur in the Pacific Islands and in the northern .
part of Australia.
101
m
~
()
C
. .. .......
.. .. . . . .. . ~
.. ... .. . .. .. ...... ...... .. ...... .
...-............... o
...
. .. . ....................
............................. . . . . .. . . . .. . .
. .. .. ., .. .............. .. z
•
. . . . . . . .. . .. . . . . . . .. . . . . .
.. • • • • • • • .. .. • • .. .. . . It ••• ., ..
o-n
·1········································· --'
"!j ~~:::~; ~:~ ~~ ~mwm lj;m\ij:i{:":""" i (/)
m
o
3:
m
Z
~ '
(/)
t km-2 yr- 1
240
100
50
10
,0 Arid regions
(a)
Figure 5.1. Global patterns of sediment yield: (a) after Strakhov (1967); (b) after Fournier (1960)
103
m
0 <
»
~
()
c
~
0
z
0
"m
Vl
-0
3:
m
Z
-I
Vl
(a)
Figure 5.2. Global patterns of suspended sediment yield: (a) from Lvovich (1991) in Walling and Webb (1996); (b) frOln Walling and
Webb (1983)
" ~,
~ .~
xv b
o
1000
750 »;0
500 m
»
Vl
250
Vl
100
50
),
VI
~ I@
-;
O Deserts and
permanent ice
0
(b) I~
C
Vl
I
o
Figure 5.2_ continued -
Z
Vl
G)
m
-Nean ,monthl}!. precipitation JGPCpmonitoring) _ ~
forthe.- yeor lJan -Dec) -1~98 -.nmm/month ()
C
90N..---------r------,-.-----,------....,..------r-o------. ~
o
z
o
"m
(/)
o
3:
m
Z
-I
30N - t - - - - -- - t - " I (/)
EQ~--------~----~
3~~---------+----~--~
6~~--------~----------~----------r----------+--------~~-------~~
9~~--------~----------~----------~---------+----------~--------~
1BO 12ml 60W (I GOE 1eo
1 - :
~
. . \
.I . \
_..__ ..__ _.. _#_ . . __ _
~_. _ ...._... . ..
~_"' :..a ._~ __ ~ ; ~ !a.' . . ..;:J~ 5~ ~~
Ea~-----------~-----~~~
»
:;;0
m
»
VI
VI
C
-l
m
9~~----------~----------~----------~~------------~----------~~--------~
1fiD 12ml !)OW 0 tiOE 120£ 1SO
o
-l
o
"C
r-
1 1Q 25 50 15 100 1M 200 300 400 VI
I:
Z
Figure 5.4. Precipitation distribution during winter 1998 (December to February) C)
1
EVACUATION OF SEDIMENTS
1
Precipitation distribution during spring (March to May). Figure 5.5 shows the
global precipitation distribution during the nOl1hern spring of 1998. In spring, 1
high precipitation values extend to roughly the same area with a gradual
northwards movement of peak values. In Africa the high values are now more
concentrated on the central-western coast around Cameroon and Gabon and 1
values in Asia are higher in eastern China and Japan with the islands of the
Pacific still experiencing high values.
1
Precipitation distribution during summer Oune to August). Figure 5.6 shows the
global precipitation distribution during the northern summer of 1998. This shows
high precipitation values in America extend from-the northern regions of South 1
America into Central America. In Africa, high values are founq in central regions
and in the westenl zone from Guinea, Ivory Coast, Cameroon, Nigeria and into
Zaire. In Asia, the monsoon period brings high amounts of rainfall to India, Tibet, 1
China, the other eastern Asian countries and to the Pacific Islands.
The climate has consistent high daily temperatures ranging from 20-30°C.
1
Monthly temp~rature averages range from 24-30°C . . The annual range of
monthly temperatures is about 3°C. Precipitation is uniform with a total over
1
108
J
1
, 1 -I
[O-~-----------+---~'-
3(}S - i l - - - - - - - - - I - - - -- -+.i
»
;0
rn
»
Ul
Ul
C
-l
m
D~~----------~-----------F~---------+~--------~----------~~----------,
1 12[}11( 150'1/ 0 1 ZOE 1BO o
-I
o
."
r
1 10 25 50 15 , 00 t flO 200 300 .4QO 600 800 100'D C
Ul
:c
Figllre 5_5_ Precipitation distribution during spring 1998 (March to JV/ay)
zC)
m
o
,' on,itoring,)
Nean. ,monthl.r' Pfacipitation(GPC.C m ~
()
for summer (Jun',Jul,Aug) 1998 In mm/montft C
~
9{JN------...----~--.....--...------.......------..,._----..-;...,.....----_., o
z
o
-n
Vl
m
OON o
3:
m
Z
-I
Vl
EC~----------~--~~
3~~----------r-------~~
6(}S-l-------+------+-------+------+--------t----~-____1
90~J.BO------12.j..rn-'I-----6+0.,,-t-----..f..D------a+oE------1+20-E------IllBO
1 1
L- ' " I..--. ' ' \-. " L, ' t_ " L- " L-." 1-- ' ' L- ' L- ' L- ' 1_ L- ' .1_
M~.' on month,~ precipitation (GPee monitoring)
for autumn lS"P.Oct"No·~r} 199B in mm/month
30N-v------~~'-t-----~-~--F'--,------r_-----....,..------ __-----""""lI
m· D - - - - - - - - I-----.;..,:-;+-I\I
3~) -~------------1----------~~),~~---------1--~~~~~~~-~-----------~
»
;::0
6liS ·u---·-----l·-----------::--+-~----_il_~,..,.,...;..~~,-_I---------II_--------_I m
»
Vl
Vl
C
-l
90S -\-------+------""i-~----_Ii__------_I---
1BO 12mI' - M't.' 0 tiOE
..
---__I------~
120E 1BO
m
Q
-l
o
."
r
10 50 75 100 ffl<J 200 300 400 600 800 1000 C
Vl
I
Figure 5.7. Precipitation distribution during autumn 1998 (September to November) Z
C)
EVACUATION OF SEDIMENTS
Jan. Feb. Mar. , Apr. May June July Aug. Sept. Oct. Nov. Dec. Yr
Temp: °C 27 27 28 28 27 27 27 27 27 27 27 27 27 [
- ,
,
Pptn: mm 554 / 519 557 620 655 655 572 574 561 563 563 512 6905
J
2000 mm. The region lies within the effects of the intertropical convergence zone
.... ,
all year. Convergence and high maritime humidities create cumulus clouds and
thunderstorms regularly. A typical monthly distribution of temperature and .-J
rainfall is given in Table 5.6.
The climate has distinct wet and dry periods. The seasonal pattern is due to the
movement of the intertropical convergence zone. The wet season coincides with
J
the high sun and the presence of the convergence zone. The dry season is due to
more stable air associated with the presence of the subtropical high zone during the
low sun season.
J
During the rainy season the climate is similar to the tropical wet climate. During i
the dry season, semi-desert conditions prevail. Some regions experience intensifi-
cation of rainfall due to monsoons and orographic uplift. A · typical monthly J
distribution of temperature and rainfall is given in Table 5.7.
Tropical desert
• Koppen classification BW.
• Continental tropical air mass all year.
This region is found near the tropics usually, but not always, on the western side
of continents and covers 25% of all land area. It is characterised by: J
• low relative humidity (10-30%) and cloud cover
• low frequency and amount of precipitation
• high mean annual temperature
J
It
Table 5.7. India, 13 oN, elevation 22 m
Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. Yr
Temp: °C 27 27 28 29 29 27 26 26 26 27. 27 27 27 J
Pptn: mm 5 2 9 40 233 982 1059 577 267 206 71 18 3467
J
I
112
J:
J
AREAS SUITEOTO FLUSHING
Temp: °C 15
0
17 21 26 31
, 1
32 32 33
l 30 28 22
I
17 35
, 3
Pptn: mm 0 0 0 0 1 0 0 1 0 0
[ [
I
The region is influenced {~y "'upp~r ail<:~:$ta.bility and subsidence owing to the
presence of the subtropicalhfgh pressure:~'zone . Temperatures are highly variable
daily and annually. vVith theayerage monthly temperatures ranging fronl 29-35°C
and the average diurnal rangejs :}Jetween 14-25°C. A typical monthly distribution
of temperature and rainfall is tsiven in Table 5.8.
Mid-latitude wet
"Koppen classification Cf a#d~f.~: . .;, :."}~.7:'· :'
o NIaritime tropical in sUII1IIl~r:~'d:maritime p~Iru<ln£"mnte~~:~.
In the Northern HeInisphere the ~regi6n is from -60oN to between 25 and 30 0 N
mainly on the western side of continents. In the Soutt1.em Hemisphere the climate.;,·
spans from the south-eastern tip of South America, New Zealand and the south-~
east coast of Australia. Summer is dominated by thunderstorms produced by daily:
heating. Monthly average temperatures range from 21-26°C. Frontal weather
associated with the mid-latitude cyclone dominates the climate of more polar areas
and is more frequent in all regions in winter. .c; .. . . . .
I Jan . Feb. I Mar. I APr.' May I June I July f Aug. I Sept. I oct./ Nov. I Dec. I Yr
Temp : cC , i 15 ,
7
I 10
4 4 7 9 12 16 18 17 11 I 5
I I I I I I !
Pptn: rnm [
54 40
,
' 37 38
I
46 46 56 I
I
I
59 ,
50
I
57
I
64
I
48 , 5"95
113
EVACUATION OF SEDIMENTS
i
-10 !
i
-9 -4 4
i
10 13 17 I 15 11 5 -2 -7 4
I
, 26 i
15
Pptn: mm
I
17 I 20
I
35
I
52
I
88 58 ! 59 35 23 16
I
444
i
J
114
I AREAS SUlTEDTO FLUSHING
Jan. I Feb. I Mar. Apr. May June July Aug. Sept. I Oct. Nov., Dec. Yr
Pptn: mm I 23 I 20 8 5
I 3 5 12
I 19 21
I 16 35 I 37 204
the summer insolation is high with long days, however the albedo of the snow
surfaces reflects up to 90% back. Average monthly temperatures are generally
below O°C. A typical monthly distribution of temperature and rainfall is given in
Table 5.13.
• In the cold steppe climate (BSk) the range in value of annual sediment yield
is betvveen 1 and 16 300 t/km2 • Only 13 out of the 75 rivers have yields of
more than 600 tJkm2; these are mainly in China and Argentina with one in
South Africa and one in ~he USA.
• The two warm temperate humid climates with no dry period (Cfa and Cfb)
have high standard deviations. For the Cfa climate, Taiwan and Italy are
responsible for the high standard deviations. The ten rivers with the highest
sediment yield are in Taiwan with a range of 2605-18 339 tJkm2 while 106 of
the rivers have yields of less than 300 tlkm2 • In the Cfb climate there are 144
rivers with less than 100 tf1lffi12 and 38 rivers (23 in New Zealand) which yield
more than 500 t/km2,
• The cool Mediterranean climate (Csb) has limited data with a wide variation
2
in yields. Four rivers in Australia have yields of less than 2 t/km and four
rivers in the USA yield more than 1000 t/km2•
., The boreal climate without a dry period (Dfb and Dfc) has a low sediment
2
yield. The median value for Dfb is 33 tJkm2 and the mean is 104 tlkrn . There
Temp: °c I -2 I -5 1-10 1-14 I -16 1-16 1-17 -17 -17 I -14 1 -7 1 -3 1-12
Pptn: mm I. 13 I 19 I 51 ! 44 ! 92 i 67 I 77 I 95 52 I 43 I 46 I 26 I 625
115
EVACUATION OF SEDIMENTS
~----------~~~--------------------------------~~~~
A Tropical rain climates
f no dry period
w dry winter
C Warm temperate rain climates
a warmest month >22
b at least 4 months >10
" ~
are eleven mountainous rivers in Romania and two Soviet rivers in the
Caucasus area with more than 500 t/km2• The median value of Dfc climate is
9 t/km2 with a mean of 76 t1km2 • The data are highly variable, with a standard
deviation of 3·5 times the mean.
• The boreal climate with dry winter (Dwb) has too few data points to give
reliable values. The snow climate ET consists of two populations - rivers in
the fonner Soviet Union with up to 50 t/km2 and rivers in Alaska and Iceland
with yields of between 375 and 4000 t/km2 (Jansson, 1988).
The data from the climatic regions were amalgamated by Jansson (1988) into
simplified homogeneous climatic groups, depending on their characteristics. The
groups are set out in Table 5.14.
The classification of sediment yield for each climatic group shows a number
of distinct features which can be seen in Figure 5.9. Over half the rivers in the
Af group have sediment yields greater than 1000 t/km2, with only one river
exhibiting a value less than 100 t/km2 • The groups Cwa and Cs have the next
largest percentages in the highest sediment yield class, with more than 50% of
the basins having sediment yields above 100 tIkln2• Dfa-d and Dwc have 67%
and 100% of the rivers respectively with sediment yields less than 100 t/km2 • Cfb
116
AREAS SUITEDTO FLUSHING
Table 5.14. Reasons for com,bining climates into homogeneous climatic groups .
(Jansson, 1988)
Climates Climatic Reasons for grouping
group
I
Af,Cf Af Cf is found in tropical mountains. Similar precipitation
conditions as Af
I
AW,Am, Cw, Aw Cw is found in tropical mountains. Cw (Argentina excluded)
outside Argentina has a median yield of 246 rll<m2 and a mean of 446 tlkm"
I
BSh,BSk BS Similar rain conditions. Not many low sediment yield values in
BSh. Many low and many high values in BSk
I -.
BWh, BWk BW Similar rain and sediment yield conditions
I
I
Cfa Cfa Few low values cefb in contrast has many low values)
I
Ctb, Cfc Ctb Cfc similar climate as Cfu
l
Df Df Mountains in Cfa or CfD but Df climates have more snowmelt
erosion than Cfa or etb
I
Csa, Csb, Cs Cs As rain and erosion is in winter, the temperature of the warmest
month, whether Csa or Csb, is of no significance. Extremely
low values in Australia both of Csa and Csb
.-
Dsb, Ds Dsb and Ds are found in mountainous areas in Csa or Csb but
Ds
I have more snowmelt run-off than Csa or Csb
Dfa, Dfo, Dfe, Dfd I All have snowmelt run-off in spring when vegetation cover is
Dfa-d
sparse. Temperature similar to snowmelt erosion
I
D we. D\vd I Dwe Similar snO\vmelt and rainfall conditions. Dwa and Dwb hU"'e
different rainfall and snowmelt conditions from Dwc and Dwd
I
ET, ET (Mt) ET Similar climate
I
has about 450/0 within the two lowest classes. Low values are found in Central
__::"._0 and Western Europe, with high values tn the Southern Hemisphere and in
southern-most Europe.
Generally, a tropical Af climate has high sediment yields while Aw has more
variable values. In arid climates BS has high values, while the desert group BW
has low and intermediate values. There are few rivers with yields less than
10 tf1.i.<ll12 in the warm temperate clirrlates except in the Cfb zone. In the boreal
region sediment yields are low, except in the mountains of the warm temperate
zone, Ds and Df, where elevated yields occur (Jansson, 1988).
Jansson (1988) provided a list of countries in each climatic zone in order to
show the geographical distribution of each. This allows an estimate of erosion
117
EVACUATION OF SEDIMENTS
Boreal
1
r--'-
Warm
temperate
:
,--
At
Tropical
and arid
---'-
Figure 5.9. Number of basins within sediment yield classes in climatic groups
rates to be made for each country based on climatic classification. Table 5.15 lists
the climate zones in each country and indicates the number of rivers that occur
in each zone.
A similar exercise has been carried out in this study for a more comprehensive
list of countries in the world, the results of which are shown in Appendix A4.4.
The updated map of climatic zones produced by Alexandersson (1982) based on
the Koppen classification, was used to determine the climate classes that occur in
each country and the rates of sediment yield were based on the map produced by
Walling and Webb (1983) (see Appendix A4.4).
There is a reasonable similarity between the data listed in Table 5.15 and the
results we have obtained using the world climate map. Differences arise mainly
because the data in Table 5.15 are based· on a number of river basins in each
118
AREAS SUITEDTO FLUSHING
Table 5.15. . Countries classified into climatic zones showing nwnber of river basins in
each zone (modijiedfrom Jansson, 1988)
Country (nLlmber of rivers/zone) Climate zone
119
· EVACUATION OF SEDIMENTS
Sweden (14/3/2)
I
Dfc/CfblDfb
Switzerland (9/5) Df/Cfb
Taiwan (16111) CwaJCfa
Tanzania (1)
Thailand (25/13111)
Bsh
Aw/CwaJAm 1
Trinidad (1) BS
Tunisia (6) Csa
Turkey (1) Ds
USA (83116/14/12/916/4/212) CfalBSk/CsblDfblDfaJDflDsb/CfblB sh
USA-Alaska (7/5) DfclET
Venezuela (4/312) Cf/AwIBsh I
f
Yugoslavia (28/311)
Zimbabwe (111)
Cfb/CfaJCs
BShlCwa
1
I
.....I-
country rather than on a global assessment of which climatic zone the country fits
In.
This study emphasises the diversity of climate classes and rates of sediment
1
yield within many individual countries and shows that this base unit will often
not be small enough to produce homogeneous conditions. On a global scale, a
compromise must be reached between generalisation and accuracy and the
1
country provides a manageable unit for some purposes. ,
1
i
~
5.3. GEOGRAPHICAL AREAS SUITABLE FOR FLUSHING
5.3.1. Introduction
There are a number of factors common to areas suitable for the application of
reservoir flushing techniques. A first prerequisite is that there must be ,a high to
medium erosion rate within the catchment. Secondly, the sediment must then be
transported down the river system to the reservoir resulting in the requirement for
its removal for flushing. These two prerequisites are discussed in the first two
sections of this chapter. The hydrological characteristics required for successful
flushing are then considered.
120
AREAS SUITEDTO FLUSHINC;
Precipitation
High rates of erosion occur in regions where there is high intensity of rainfall. It
is not just high precipitation totals that result in high erosion rates but it is the
relationship between precipitation and vegetation. Global relationships between
erosion rates and precipitation show variable results.
Geology
The geology is an important factor deteITPjning the susceptibility of the rock to
the effects of erosive forces. Erosion rates are generally highest in areas of soft
sedimentary rocks.
Soils
The key characteristics of a soil that influence erosion rates are texture, structure,
organic matter content, shear strength and infiltration capacity. High erosion rates
occur where the texture of the soil is high in silt and fine sand and low in clay,
and where the structure is compacted and the organic matter, shear strength and
infiltration rates are low. All these factors cause high run-off rates, leading to
erosion of the soil.
Slope
The gradient and length of the soil sUlface influence the velocity and direction of
run-off and therefore its erosivity. High erosion rates occur where there are long,
steep slopes resulting in movement of water downslope at a high velocity.
Vegetationlland use
Vegetation depends on the interaction of a number of factors including rainfall,
temperature, soils and topography. The presence of a vegetation cover reduces
the erosive power of rainfall by dissipating its energy, increasing infiltration,
reducing ·the velocity of nln-off and by holding soil particles together. High
erosion rates therefore occur where there is sparse vegetation cover either due to
natural climatic conditions or due to land-use practices.
Human impact
Activities such as deforestation, urbanisation and agriculture all affect the
erodibility of the soil. Current erosion rates are more than two and a half times
the historic, mainly as a result of human influences.
121
EVACUATION OF SEDIMENTS
Erosion processes
The sediment delivery ratio is generally higher for sediment derived from
channel-type erosion which delivers sediment to the main channels of the
transport system more quickly and more directly than from sheet erosion.
DepOSitional features
The presence of a depositional area decreases the sediment delivery ratio. Most
of the sediment eroded from the steep uplands of basins may be redeposited at
the base of slopes.
The poor correlation between sediment yield and erosion rates makes it difficult
to estimate the sediment load entering a reservoir on the basis of the erosion rate
within the· catchment (Mon"is and Fan, 1997). Most studies that have attempted
to relate the delivery ratio to catchment characteristics have used an inverse
relationship with catchment area (Walling and Webb, 1983).
122
AREAS SU ITEDTO FLUSHING
5.3.5. Areas of the world which are best suited to reservoir flushing
It is not possible to define precisely which specific areas of the world will provide
conditions for 'successful' flushing. In reality there is a spectrum of conditions
ranging from tllOse sites where conditions are ideal to those sites which are quite
unsuited to sediment flushing.
From the hydrological a11d hydraulic conditions neCeSSfuy' for successful
reservoir flushing the luost likely locations in which to use this technique are
those which are within the Koppen climate classification tropical wet and dry:
classifications Aw, Am and BS. Also, there are areas in the mid-latitudes where
spring snowmelt provides a regular and predictable annual pattern of high
flows.
From the Koppen classification of climatic zones and the mid-latitude spring
snowmelt effect, the requirements for successful flushing are most likely to be
met in the following locations:
123
Si
e
.m:i
pecific
e
In sti I an
• •
d sIgn n era ons
: .' ,
Site-specific investigations and design
considerations
Site investigations
Flushing outlets have to be able to withstand high velocityfiows with high
concentrations of sediment. Such flows are pJghly abrasive and expensive steel
lining will normally be required to avoid undue damage to the structures. Hence
it is important that the site allows for the construction DT relatively compact
flushing facilities, either orifices within the dam itself or relatively short tunnels
or channels. Energy dissipation works will normally be required at the
downstream side and it is an advantage if these facilities can be shared with other
outlets such as high head spillways or irrigation outlets. It is advantageous if the
flushing facilities discharge to the downstream channel well away from any
power station outlets as any local deposition of sediments will increase tailwater
levels and reduce power output.
The reservoir itself requires a detailed survey to establish its topography. This
is required to check whether the reservoir basin is a suitable shape for sediment
flushing and also to provide input data for detailed modelling of the
sedimentation process within the reservoir.
Hydrological investigations
It has been stated that there are certain requirements for successful sediment
fiushin2: which are related to the amount of water available and its reliability year
on yea"r and season by season. Hence inflows to the reservoir need to be
127
-f
I
EVACUATION OF SEDIMENTS !
-!
~- -/
established with confidence. This involves the acquisition of historical records of
river flows going back at least 30 years and preferably longer. Records of river
flows can often be extended further back in time by considering local rainfall
records, which often go back 100 years or more, and by undeltaking catchment
modelling to convert rainfall into run-off.
The ideal situation for sediment flushing is an annual inflow of water of at least
three times the volume of the reservoir (original volume in the case of existing
reservoirs) and an annual hydro graph which shows distinct wet and dry
seasons.
Sediment investigations
j
The amount and nature of the sediment entering or likely to enter the reservoir
needs to be established. This requires measurements over many years to establish
the "results with the confidence that is required. There are variolls approaches to
this task. Most commonly, sediment transport is measured at a gauging stavon
not too far upstream of the reservoir and a relationship between flow rate and
sediment transpolt rate is established. The long hydrological record is then used
to compute the total amount of sediment passing the gauging station by -
integrating over the period of the record. There are some dangers in doing this
because there is no unique relationship between flow rate and sediment transport
rate for fine sediments, the quantities of sediment being determined by the
amount being washed off the catchment not the capability of the river to transport
them. Bed load is difficult to measure and is often estimated as 10% of the total
sediment load. An alternative approach is to calculate the bed load using
established predictive techniques.
. In the case of existing reservoirs, information about the amount of sediments
entering the reservoir can be augmented by surveys of the amount and nature of
the material settling within the reservoir. Care is required, however, to allow for :4
J
the amount of material, mainly fine, which passes through the reservoir without
deposition.
Bed material sampling should be 'undertaken in the reservoir and in the rivers
which feed the reservoir. A sound knowledge of the nature of these sediments,
including their size, specific gravity and degree of compaction, is an essential
requirement to provide inputs for numerical models which simulate sediment
movement.
Hydraulic modelling J
Sophisticated numerical (computer) modelling of the way sediment is likely to
behave within the reservoir and the amount and nature of the sediment that will
be passed to the downstream reach is the cornerstone of any detailed evaluation
of flushing facilities. One-dimensional models with quasi two-dimensional
simulation of the incised channel that develops during sediment flushing are the
most appropriate tools. These models are computationally efficient and are
capable of making long-term simulations, decades rather than hours or days.
They have reached reliability levels which permit thenl to be used 'cold' when
1
128
SITE-SPECIFIC INVESTIGATIONS
new reservoirs are being investigated. When used on existing reservoirs they have
the added benefit of measured sedimentation data for verification purposes.
Computer simulations of reservoirs ideally use representative, long-term ·
sequences of water and sediment inflows to the reservoir. The models are capable
of looking at the effectiveness of various aspects which affect reservoir
sustainability over periods of up to 50 or 60 years, including:
129
7. References
Ackers, P. (1993). Sediment transport in open channels: Ackers and \Vhite update. Proceedings
of the Institution of Civil Engineers Water, Maritime and Energy, 101, 247-249.
Albertson, M. L., Malinas, A. and Hotchkiss, R. (eds) (1996). Proceedings of the International
Conference on Reservoir Sedimenation, Fort Collins, Colorado, USA.
Anderson, H. W. (1975). Sedimentation and turbidity hazards in wildlands. In: Watershed
Nfanagement. Proceedings of a Symposium of the Irrigation and Drainage Division. ASCE,
New York, pp. 347-376.
Atkinson. (1984). Consolidation of reservoir.peposits. In: Sedimentation in reservoirs in the Tana
River basin, Kenya, R. Wooldridge (ed.). Report OD 61, HR Wallingford, UK.
Atkinson, E. (1996). Feasibility of flushing sediment from reservoirs. Report OD 137, HR
Wallingford, UK.
Atkinson, E. (1998). Reservoir operation to control sedimentation: techniques for assessment.
Proceedings of the Conference of the British Dams Society, Bangor.
Attewill, L. J. S., \vrute, W. R., Tariq, S. M. and Bilgi, A. (1998). Sediment man(!gement shldies
of Tarbela Dam, Pa..1dstan. Proceedings of the Conference of the British Dam Society,
Bangor.
Basson, G. R. and Olesen, W. (1997). Modelling flood flushing. International Water Power and
Dam Construction, 49, No.6.
Basson, G. R. and Rooseboom, A. (1997a). Dealing with reservoirsedinuntation: guidelines and
case studies. Bulletin 115, ICOLD, Paris.
Basson, G. R. and Rooseboom, A. (1997b). Dealing with rese:rvoir sedimentation. "Vater
Research Commission Report No. TT91197, Pretoria.
Bhargava, D. N., Narain, L., Tiagi, S. S. and Gupta, P. P. (l987).Sediirientation problems at low
dams in the flimalayas. Water Power and Dam. Construction, Jan., 30-33.
Binnie & Partners. (1986). Report on Guanting reservoir - sedimentation. Report for
Government of PRC (Beijing Municipal Engineering Administration Division), under
assignment by ODA.
Bordas, M. P. and Canali, G. E. (1980). The influence of land use and topography on the
hydrological and sedimentological behaviour of basins in the basaltic r~gion of South Brazil.
In: The influence of man on the hydrological regime with special reference to representative
and experimental basins. International Association of Hydrological Sciences, Publication No.
130, pp. 55-60.
Branski, J. (1975). Oeena denudacji dorzecza Wisley na podstawie wynikow pomiarow
rumowiska unoszonego. Prace Instytuto Nfeteorologii i Gospodarki ¥Iodnej, 6, 1-58.
Bnme, G. 11. (1953). efficiency of reservoirs. Transactions of the A,m.erican Geophysical
Union, No.3, Washington, DC, 407-418.
Budyko, !vi. 1. (1956). Heat balance of the Earth's surface (in Russian). Giclrvn1eteoizdat,
Leningrad, 255 pp.
133
EVACUATION OF SEDIMENTS
134
, REFERENCES
Gregory, K. J. and vValling, D. E. (1973). Drainage basin form and plVcess. Edward Arnold,
London
Griffiths, G. A. (1981). Some suspended sediment yields from South Island catchments, New
Zealand. Water Resources Bulletin, 17, 662-67l.
Griffiths, G. A. (1982). Spacial and temporal variability in suspended sediment yields of North
Island basins, New Zealand. Water Resources Bulletin, 18, 575-584.
Hardjowitjitro, H. (1981). Soil erosion as a result of upland traditional cultivation in Java Island.
In: Proceedings of South-East Asia Regional Symposium on Problems of Soil Erosion and
Sedimentation, T. Tingsanchali and H. Eggers (eds). Asian Institute of Technology, Bangkok,
pp.173-179.
Hwang, J. S. (1985). The study and planning of reservoir desilting in Taiwan. ~Vater
International, 10, 7-13.
International Commission on Large Dams (ICOLD). (1988). World register of dams: updating.
ICOLD, Pa..ris.
International Commission on Large Dams (ICOLD). (1997). General report on question 74.
ICOLD 19th Congress, Florence, Vol. 3.
International Commission on Large Dams (ICOLD). (1998). World register of large dams. CD
ROM.
In[ernational Research and Training Centre on Erosion and ~edimentation (IRTCES) (1985).
Lecture notes of the training course on reservoir sedimentation. IRTCES, Tsinghua"
University, Beijing, Chh"la.
Institute of vVater Conservacy and Hydroelectric Power Research (IWHR). (1983). The rational
operation of drawdownfiushing in Shuicaozi reservoir on the flU River. Summary in English
of Paper SR8301, IWHR.
Jaeggi, A. L. and Kashyap, B. R. (1984). Desi1ting of Baira reservoir of Baira Siul project.
Irrigation and Power, Oct., 375-380.
Jansson, M. B. (1988). A global survey of sediment yield. Geographical Annals, No. 70, Series
A (1-2), 81-98.
Jansson, M. B. (1992). Suspended sediment outflow from Cach! reservoir during the flushing in
1990. In: Sedimentological Studies in the Cach{ reservoir, Costa Rica, IVI. B. Jansson and A.
Rodriguez (eds). UNGI Report No.8l. Department of Physical Geography, Uppsala
University, Sweden.
Jantawat, S. (1985). An overview of soil erosion and sedim,entation in Thailand. In: Soil erosion
and conservation, S. A. El-Swaify, W. C. Moldenhauer and A. Lo(eds). Soil Conservation
Society of America, Ankeny, Iowa. pp. 10-14
Jarecki, E. A. and Murphy; T. D. (1963). Sediment withdrawal investigation - Guen'lsey
reservoil: Paper No. 91, Miscellaneous Publication 970, US Department of Agriculture.
Jowett, I. (1984). Sedimentation in New Zealand hydroelectric schemes. Water International, 9,
172-176.
Jolly, J. P. (1982). A proposed method for accurately calculating sediment yields from reservoir
deposition volume. In: Recent developments in the explanation and prediction of erosion and
sediment yield (Proceedings from the Exeter symposium, July 1982), IAHS Publication No.
137.
Jowett, I. (1984). Sedimentation in New Zealand hydroelectric schemes. ~Vater International, 9,
172-176,
135
EVACUATION OF SEDIMENTS
Judson, S. and Ritter, D. F. (1964). Rates of regional denudation in the United States. Journal of
Geophysical Research, 69, 3395-3401.
Krumdieck, A. and Chamot, P. (1979). Sediment flushing at Santa Domingo reservoir. Water
Power and Dam Construction, Dec., 172-176.
J
Lagwankar, V. G., Gorde, A. K., Borikar, D. A. and Patil, K. D. (1995). Trends in reservoir
sedimentation in India. In: 6th International Symposium, on River Sedimentation, New Dehli,
pp.91-111. J,
Lai, 1. S. and Shen, H. W. (1996). Flushing sediment through reservoirs. Jountal of Hydraulic
Research, IAHR, 34, No.2.
Lane, E. W. and Koelzer, V. A. (1953). Density of sediments deposited in reservoirs. Report No.
9: A study of methods used in measurement and analysis of sediment loads in streams, St Paul
I'
l!¢te~. States Engineering District, St Paul, Minnesota, US. ~.
~
Li, Y. H. (1976). Denudation of Taiwan Island since the Pliocene Epoch. Geology, 4, 105-107.
Mahmood, K. (1987). Reservoir sedimentation: impact, extent and mitigation. World Bank
J
I
Technical Paper, No.7!. ISSN 0253-7494, ISBN 0-8213-0952-8.
Mohan, J., Tiagi, S. S. and Rajvanshi, U. S. (1982). Effects of sedimentation at Ichari reservoir.
ICOLD 14th Congress, Rio de Janeiro, Question 54, Reply 26, pp. 409-418. ,-1..
!
Morgan, R. P. C. and Davidson, D. A. (1986). Soil erosion and conservation, Longman Group,
UK.
Morris, G. L. and Hu, G. (1992). HEC-6 Modeling of Sediment Management in Loiza Reservoir,
Puelto Rico. In: Hydraulic Engineering Water Forum '92, M. Jennings and N. Bhowmik
(eds), ASCE, New York, pp. 630-635.
Monis, G. L. and Fan, J. (1997). Reservoir sedimentation handbook: design and management of
dams, reservoirs and watershedfor sustainable use. McGraw Hill, New York ..
i
Mou, J. and Meng, Q. (1980). Sedilnent delivery ratio as used in the ' computation of the r
watershed sediment yield. Beijing, China.
Neil, C. R. and Mollard, J. D. (1982). Erosional processes and sediment yield in the upper
Oldm,an river basin Alberta, Canada. Recent developments in the explanation and prediction
of erosion and sediment yield. Proceedings of a symposium held during the first lARS
scientific assembly at Exeter, 1982, Publication No. 137.
Noble, E. L. (1965). Sediment reduction through watershed rehabilitation. In: Proceedings of the
Federal Inter-Agency Sedimentation Conference, 1963. US Department of Agriculture,
Miscellaneous Publication 970, pp. 114-123.
136
,
j
REFERENCES
O'Loughlin, C. L., Rowe, L. K. and Pearce, A. 1. (1980). Sediment yield and water quality
responses to clearfe1ling of evergreen mixed forests in western New Zealand In: The influence
of man on the hydrological regime }vith special reference to representative and experimental
basins. International Association of Hydrological Sciences, Publication No. 130, pp.
285-292.
Ongweny, G. S. (1979). Patterns of sediment production within the upper Tana basin in eastern
Kenya. In: The hydrology of areas of low precipitation (Proceedings of the Canberra .
Symposium, December 1979), IAHS Publication No. 128.
Palmieri, A. (1998). The role of technical solutions in the context of international developnzent
and dam investments. International Atomic Energy Agency, Consultant Meeting, Thematic
Planning for Technical Cooperation Isotope Hydrology in Danl Management, Vienna.
Paul, T. C. and Dhillon, G. S. (1988). Sluice dimensioning for desilting reservoirs. Water POrver
& Dam Con.struction, May, 40-44.
Pickup, G., Higgins, R. J. and vVarner, R. F. (1981). Erosier. and sediment yield in Fly River
drainage basins. In: Erosion and Sediment Transport in Pacific Rim Steeplands, T. R. H.
Davies and A. 1. Pearce (eds). International Association of Hydrological Sciences, Publication
No. 132, pp. 438-456.
Pidwimy, M. (1999). Climatic regions of the world: clim.a.tology - the geographic study of the
atmosphere, introduction to physical geography II. Department of Geography, Okanagan
University College, http://www.artslouc.bc.ca/geog/G121Iln.html. pp. 1-9.
Qian. N. (1982). Reservoir sedimentation and slope stability technical and environmental effects.
14tll Congress on Large Dams, leaL!!, Rio de Janeiro. __
RaInfrez, C. and Rodriguez, A. (1992). History of the Cachf Reservoh; Costa Rica. UNOI Report
No. 81, Department of Physical Geography, Uppsala University, Sweden.
Rhoades, E. D., Welch, N. H. and Coleman, G. A. (1975). Sediment-yield characteristics from
unit source watersheds. In: Present and prospective technology for predicting sediment yields
and sources. US Department of Agriculture. Publication No. ARS-S-40, pp. 125-129.
Rien6ssl, K. and Schnelle, P. (1982). Sedimentation of small reservoirs in the high Alps. ICOLD
14th Congress, Rio de Janeiro, Question 54, Reply 54, pp. 65-82.
Rooseboom, A. and Van Harmse. H. R. (1979). Changes in sediment load of the Orange River
1929-1969. Hydrology of areas of low precipitation. IAHS publication 128, proceedings of
conference, Canberra, Dec.
Salas et al. (1997a). Capacity situation in Spanish reservoirs. ICOLD 19th Congress, Florence,
Vol. 3, Question 74, Reply 53.
Salas et at. (1997b). Sediment yield at Spanish reservoirs and its relatior.ship with the drainage
basin area. ICOLD 19th Congress, Florence, Vol. 3, Question 74, Reply 54.
Shahin, M. IV!. A. (1993). An overview of reservoir sedimentation in some African river basins.
In: Sediment problems: strategies for monitoring, prediction and control, lARS Publication
No. 217, Wallingford UK, pp. 93-100.
Sharma, K. D. and Chatterji, P. C. (1982). Sedimentation in Nadis in the Indian arid zone.
Journal of Hydrological Sciences, 27, 345-352.
Shen, H. "'vY and Lai, 1. S. (1996). Sustain reservoir useful life by flushing sediment. International
Joumal of Sediment Research, IRTCES, 11, No.3.
Shen, H . \v. (1999). Flushing sediment through resen'oirs. Journal of Hydraulic Research, 37,
No.6.
137
EVACUATION OF SEDIMENTS
Strakhov, N. M. (1967). Principles of Lithogenesis, Vol. 1, 245 pp. Oliver and Boyd,
Edinburgh.
Subramanian, V. (1982). Erosion and sediment yield in Krishna River Basin, India Recent
developments in the explanation and prediction of erosion and sediment yield. Proceedings of
a Symposium held during the first lARS scient~fic assembly at Exeter, 1982; Publication No.
137.
Swiss National Committee on Large Dams (SNCOLD) (1982). General paper ICOLD 141h
Congress, Rio de Janeiro, pp. 889-958.
Ta Wei Soong, M. and Yean Zhao. (1994). Flood and sediment characteristics of the Lower
Yellow River in China. Water International, 19, 129-137.
Tabuteau, M .. M. (1960). Etude graphique pour les consequences hydro-erosives du climat
mediterrane. Association Geographique Fran{:aise Bulletin, 295, 130-142.
Tolouie, E. (1993). Reservoir sedilnentation and de-siltation. PhD thesis, University of
Birmingham, UK.
United Nations Educational, Scientific and Cultural Organisation (UNESCO). (1985). Methods
of computing sedimentation in lakes and reservoirs. A contribution to the International
Hydrological Programme, IHP-II projectA.2.6.1 Panel (Stevan Bruk, Rapporteur) (essentially
the same as IRTCES, 1985).
United Nations Education and Scientific Organisation (UNESCO) (1990). UNESCO sourcebook
in climatology for hydrologists and water resource engineers. UNESCO Publication,
Catalogue No. 89598, Paris.
Ursic, S. 1. and Dendy, E E. (1965). Sediment yields from small watersheds under various land
uses and forest covers. In: Proceedings of the Federal Inter-Agency Sedimentation Conference
1963. US Department of Agriculture. Miscellaneous Publication. No. 970, pp. 47-52.
US Global Change Research Information Office. (1999). Soil erosion and sediment erosion,
pp. 1-3. Web site: http://www.gcrio.org/geo/
Walling, D. E. (1994). Measuring sediment yields from river basins. In: Soil erosion research
methods, R. Lab (ed.). Soil and Water Conservation Society, Ankeny, Iowa.
Walling, D. E. and Kleo, A. R. (1979). Sediment yields of rivers in areas of low precipitation,
a global view, pp. 479-493. The Hydrology of Areas of Low Precipitation, lARS Publication
No. 128, Wallingford, UK.
Walling, D. E. and Webb, B. W. (1983). Patterns of sediment yield. In: Background to palaeo-
hydrology, K. J. Gregory (ed.). John Wiley, New York, pp. 69-100.
Walling, D. E. (1984). Muddy waters move mountains. The Geographical Magazine, 262-267.
Walling, D. E. and Webb, B. W. (1996). Erosion and sediment yield: global and regional
perspectives. IARS Publication No. 236, Wallingford, UK.
Webb, R. M. T. and Soler-Lopez, L. R. (1997). Sedimentation history of Lago Loiza 1953-1994.
USGS, San Juan.
White, P., Labadz, 1. C. and Butcher, D. P. (1996). The management of sediment in reservoired
catchments. Chartered Institution of Water and Environmental Management Journal, 10,
June, 183-189.
White, W. R. (1982): Sedimentation problems in river basins. Project 5.3 of the International
Hydrological Programme. UNESCO.
138
REFERENCES
vVhite, vY. R. and Bettess, R. (1984). The feasibility of flushing sediments through reservoirs.
Proceedings of the Symposium 'Challenges in African Hydrology and Vlt~1ter ResoLlrces\
Harare, July, 1984. IAHS Publication No. 144.
Wilson, L. (1969) . Les relations entre les processus geomorphologiques et Ie climat modern
comme methode de paleoclimatologie. Revue de Geographie PhysiqLie et de Geologie
Dynamique, Ser. 2, 11, 303-314.
Xia, M. (1986). New desilting methods used in Heisonglin reservoir: lateral erosion on
floodplain . In: Collected Research Papers, Northwest Institution of Hydraulic Research, 2,
Part 2, Yangling.
Yoon, Y. N. (1992). The state and the perspective of the direct sediment removal method from
reservoirs. International Journal of Sediment Research, 7, No.2, 99-116.
Zhang Hao, Xia Mai-ding, Chen Shi-ji, Li Zhen-Wu Xia Heng-Bin ,Jiang Nai-Sen and Lin Ben-
wen (1976). Regulation of sediments in some medium and small sized reservoirs on heavily
silt laden streams in China. 12th Congress on Large Dams, ICOLD, ~lexico.
Zhang Qishun and Long Yuqian. (1980). Sediment problems of Sanmenxia reservoir. 1st
Symposium on Reservoir Sedin'Lentation, Beijing, pp. 707-716.
139
Be lio rop
r
I
I
I
II
I 8. Bibliography
I
j
1
!
~ Abdelhadi, :M. L. (1995). Environment and socio-economic impacts of erosion and sedimenta-
!
i
tion in Northern African countries. 6th Symposium on River Sedimentation, New Delhi, Vols.
1-4, pp. 1141-1152.
Ackers, P. and Thompson, G. (1987). Reservoir sedimentation and influence of flushing. In:
Sediment transport in gravel bed rivers, C. R. Thome, 1. C. Bathurst and R. D. Hey (eds).
John Wiley and Sons, pp. 845-868.
Ackermann, Wbite and Worthington (eds). (1973). Geophysical 1'IIonograph No. 17, American
Geophysical Union, Washington.
Adams, 1. (1980). High sediment yields from major rivers of western southern Alps, New
Zealand. Nature, 287, 88-89.
Ahmed, T. E. S. (1992). Sedimentation in Sudan mUlti-purpose reservoirs. 5th International
Symposium on River Sedimentation, Karlsruhe.
Aitken, A. P. (1981). Aspects of erosion and sediment transport in Java. Problems of Soil Erosion
and Sedimentation, South-East Asia Regional Symposium, Thailand.
Ando, N., et al. (1994). SedLllent removal project at :NIiwa Dam. ICOLD 18th Congress, Vol. Ill.
Question 69, Reply 27.
Annandale, G. W. (1987). Reservoir sedimentation. Elsevier Science Publishers, Amsterdam.
Anon. (1982). Sedimentation of small reservoirs in the high Alps. ICOLD 14th Congress, Rio de
Janeiro, Question 54, Reply 5.
Anon. (1982). Estimation of silting load wid capacity loss rate in irrigation reservoirs in Korea.
ICOW 14th Congress, Rio de Janeiro, Question 54, Reply 12.
Anon. (1982). Some problems concerning reservoir sedimentation on Romanian rivers. ICOLD
14th Congress, Rio de Janeiro, Question 54, Reply 14.
Anon. (1982). Sedimentation aspects of the :tYnssouri river dams. ICOLD 14th Congress,· Rio de
Janeiro, Question 54, Reply 15.
Anon. (1982). Environmental impact a.l1d control of reservoir sedimentation. ICOLD 14th
Congress, Rio de Janeiro, Question 54, Reply 23.
Anon. (1982). Studies on sediment transport in mountain streams of Eastern Alps. ICOLD 14th
Congress, Rio de Janeiro, Question 54, Reply 28:
Anon. (1982). Reservoir sedimentation - technical and environmental effects in Indonesia.
ICOLD 14th Congress, Rio de Janeiro, Question Reply 31.
Anon. (1982). Investigation of average long-term suspended sediment discharge in Brazil.
ICOLD 14th Congress, Ria de Janeiro, Question 54, Reply 32.
Anon. (1982). Silting of reservoirs in southern Brazil - first results and methodological
approach. ICOLD 14th Congress, Rio de Janeiro, Question 54, Reply 33.
Anon. (1982). Sediment deposition in reservoirs on the River Tana, Kenya. ICOLD 14th
Congress, Rio de Janeiro, Question 54, Reply 37.
143
EVACUATION OF SEDIMENTS
Avendano Salas, C., Sanz 1v10ntero, E. and Gomez Montana, J. L. (1997). Sediment yield at
Spanish reservoirs and its relationship with drainage basin area. ICOLD 19th Conference,
Florence 1997, Vol. 3.
Bister. (1991). Ageing of dams. ICOLD, Vienna, Question 65.
Biswas, A. K. (1991). Water for sustainable development in the 21st century - a global
perspective. Water Intemational, 6, No.4, 219-224
Bollinne, A. (1985). Adjusting the universal soil loss equation for use in Western Europe. In: Soil
erosion and conservation, S. A. El-Swaify, W. C. Moldenhauer and A. Lo (eds). Soil
Conservation Society of America, Ankeny, Iowa, pp. 206-213.
Borland, W. M. and Miller, C. R. (1958). Distribution of sediment in large reservoirs. Journal of
the Hydraulics Division, ASCE, 84, No. HY2.
Borland, W. M. and Miller, C. R. (1960). Distribution of sediment in large reservoirs.
Transactions ASCE, 125.
Brune, G. M. (1953). Trap Efficiency of Reservoirs. Trans A117,erican Geophysical Union, 34, No.
3,407-418.
Bruk, S. and Zebidi, H. (1996). Reservoir sedimentation. Proceedings of the St Petersburg
Workshop (May 1994), International Hydrological Programme, UNESCO, Paris.
Buff, B., et al. (1997). Survey of sedimentation in dams in the free sate of Saxony, Federal
Republic of Germany. ICOLD Conference.
Butcher, et al. (1992). Assessment of catchment erosion in the southern Penines using reservoir
sedimentation monitOling. International Association of Hydrological Sciences, No. 210.
Buttling and Shaw. (1973). Predicting the rate and pattern of storage loss in reservoirs. 11th
ICOLD Congress, Madrid, Q40-R37.
Central Board of Inigation and Power. (1981). Sedimentation studies in reservoirs. Technical
Report No. 20, Vol. I and II. Central Board Irrigation and Power, New Delhi, India.
Chain. (1986). Reservoir sedimentation and desilting in Taiwan. Proceedings Conference Water
Forum, Vol. 1, World Water Issues in Evolution, USA.
Chen, 1. L. (1994). Sedimentation studies at Three Gorges. International Water Power and Dam
Construction, Aug., 54-58.
Coleman, D. J. and Scatena, F. N. (1986). Identification and evaluation of sediment sources.
Drainage Basin Delivery IAHS. Pub. 159, Wallingford, UK, pp. 3-18.
Crowder, B. M. (1987). Economic costs of reservoir sedimentation: a regional approach to
estimating cropland erosion damages. Journal of Soil and Water Conservation, 42, No.3,
194-197.
Dai, D. (1991). River sediment problems in China. In: Water conservancy in China. Water
Resources and Electric Power Press, Beijing.
Delft Hydraulics, et al. (1992). The control of reservoir sedimentation - a literature review.
Ministry of Development Cooperation, The Netherlands.
Dendy, F. E. (1967). Sediment deposition in reservoirs in the United States. ARS - Paper
#41-137, US Department of Agriculture.
Dendy, F. E., Champion, W. A. and Wilson, R. B. (1973). Reservoir sedimentation surveys in the
USA. In: Man-nwde lakes: their problems and environmental effects, W. C. Ackerman, E. F.
White and E. B. W0l1hington (eds). Geophysical Monograph, No. 17, AmeIican Geophysical !
Union, Washington, DC.
I
~I
144
BIBLIOGRAPHY
Dendy, F. E. and Champion, W. A. (1978). Summary of reservoir deposition surveys made in the
USA through 1975. US Department of Agriculture, rvIiscellaneous Publication No. 1362.
De Vera and Rewtarkulpaiboon. (1981). Analysis of sediment yield and estimated soil loss of
selected Philippine watersheds. Proceedings of South·East Asian Regional Symposium orz
Problems of Soil Erosion and Sedimentation, Bangkok, Thailand.
Dunne, T. (1979). Sediment yield and landuse in tropical catchments. Joun-zal of Hydrology, No.
42, 281-300.
Engelman, R. and Leroy, P. (1994). Sustaining water: population and the future of renewable
water supplies. Population Action International, Washington DC.
Fan, J. and Morris, G. L. (1992). Reservoir sedimentation II: reservoir desi1tation and long-term
storage capacity. Journal of Hydraulic Engineering, 118, No.3, ASCE, 370-384.
Fan, 1. and Morris, G. L (eds). (1993). Notes on sedinzent .'1-zanagement in reservoirs: natiortal
and international perspectives. Federal Energy Regulatory Commission, Washington DC.
Fan, J. (1995). An overview of preserving reservoir storage capacity. International ~Vorkshop on
Reservoir Sedil7'zentation, San Francisco. Federal Regulatory Commission, Washington, DC.
Fan, J. and Pan, S. (1996). Control of sediment deposition by flood flushing. Proceedings of the
International Conference on Reservoir Sedimentation, Fort Collins, Colorado.
Fan, S. (1988). Twelve selected computer stream sedimentation models developed in. the United
States. Interagency Sedimentation Work Group and Federal Energy Regulatory Commission.
Fan, S. (1989). Reservoirs. In: Civil Engin.eering guidelines for planning and designing
hydroelectric developments, American Society of Civil Engineers.
Fan, S. (1990). lvIajor sedimentation issues and activities at the selected Federal Agencies.
Federal Energy Regulatory Commission.
Fan, S. (1999). Sediment management at hydropower reservoirs. International Journal of
Sediment Research, 14, No.2, 391-397.
Fleming, G. (1969), Design curves for sllspended load estimation. Proceedings of the Institution
of Civil Engineering, 43, No. 1.
Frenette, M. and Julien, P. Y. (1986). Advances in predicting reservoir sedimentation.
Proceedings of the 3rd International Symposium River Sedimentation, U. lVIississippi,
Oxford.
Galay, V. (1987). Erosion and sedimentation in the Nepal Himalaya. "Vater and Energy
Commission Secretariat. Kathmandu.
Garde, R. J. and Kothyari, V. C. (1987). Sediment yield estimation. Journal of Irrigation and
Power, 44, No.3.
Gogus and Yalcinkaya. (1992). Reservoir sedimentation in Turkey. 5th International Symposium
on River Sedimentation, Karlsruhe.
Goldsmith, E. (1985). Sedimentation - the way of all dams - the social and environmental
effects of large dams. European Ecological Action Group, Vol. 1, pp. 226-239.
Gong, S. and Jiong, D. (1979). Soil erosion and its comrol in small watersheds of the loess
plateau. Saentia Sinica, 22, No.1, 1302-1313.
opec Precipitation Visualiser see http://www.dwd.de/research/gpcc/visu_gpcc.htm1
Graf, "V. H. (1984). Storage losses in reservoirs. vverter Power an.d Dam Construction, 36, No.4,
37-40.
145
EVACUATION OF SEDlfVlENTS
Griffiths, G. A. (1979). High sediment yield from major rivers of western southern Alps, New
Zealand. Nature, No. 282, 61-63.
Hadley, R. F. and Shawn, L. M. (1976). Relation of erosion to sediment yield. Proceedings of the
3rd Federal Interagency Sediment Conference Water Resources Council, Washington DC,
pp.1.l32-1.139.
Harrison, L. L. and .Fan, S. (1996). Sedil1'lent management solutions for FERC licensed
reservoirs. Draft Staff Paper, Federal Energy Regulatory Commission.
Harrison, L. L. and Fan, S. (1997). Needs for sediment management at non-federal reservoirs.
International Water Power Conference.
Holeman, J. N. (1968). Sediment yield of major rivers of the world. Water Resources Research,
4, No.4.
Hotchkiss, R. H. (1999). Reservoir sedimentation: research needs and refocusing perspectives.
International Journal of Sediment Research, 14, No.2.
International Association for Hydrological Sciences (IAHS) (1979). The hydrology of areas with
low precipitation. IAHS Publication 128, Wallingford, UK.
International Association for Hydrological Sciences (IAHS) (1986). Challenges in African
hydrology and water resources planning. IAHS Publication 144, Wallingford, UK.
International Association for Hydrological Sciences (IAHS) (1986). Drainage basin sediment
delivery. IAHS Publication 159, Wallingford, UK.
International Commission on Large Dams (ICOLD) (1984). Deterioration of dams and
reservoirs.
International Commission on Large Dams (ICOLD) (1989). Sedimentation control of reservoirs.
Guidelines.
International Commission on Large Dams (ICOLD) (1997) . ICOW 19th Congress, Florence,
Vol. 3, Question 74, Reply 34.
Jiang, J. and Fu. (1997). Problems on reservoir sedimentation and measures for reducing
sediment deposition in China. ICOW 19th Congress, Florence, Vol. 3, Question 74, Reply
5.
Julien, P. Y. (1995). Erosion and sedilnentation. Cambridge University Press, Cambridge, UK.
Lauffer, H. and Sommer, N. (1982). Studies on sediment transport in mountain streams of the
eastern Alps. ICOLD conference, Vol. ill, Question 54, pp. 431-451.
Liu, X., Yueping, H. E., Zhang, X., Qin, Y. and Shuyou, C. A. O. (1999). A study of sediment
problem in the Gongzui reservoir. International Journal of Sediment Research, 14, No.2.
Meade, R. H. and Parker, R. S. (1985). Sediment in rivers in the United States. National Water
Summary 1984, USGS Water Supply Paper 2775, Washington.
Milliman, J. D. and Meade, R. H. (1983). Worldwide delivery of sediment to the oceans. Journal
of Geology, 91, No.1, 1-21.
Mizuyama, T. (1989). Sediment yield and river bed change in mountain rivers. Proceedings of
the IAHR, International Workshop on Fluvial Hydraulics of Mountain Regions, Trento, Italy,
pp.1-5.
Monis, E. L. (1995). Reservoir sedimentation and sustainable development in India: problem
scope and remedial measures. 6th Symposium on River Sedimentation, New Delhi.
Newson, M. D. (1998). The role of geomorphology in monitoring and managing river sediment -f
systems. Journal of the Chartered Institute for Water and Environmental Management, Feb.
146
".t.,","
;,
.1
,,~ BIBLIOGRAPHY
1
I1 Palmieri, A., Shah, F. and Dinar, A. (1998). Reservoir sedimentation and the sustainable
management of dams. World Congress of Enviromnerztal and Resource Economists, Venice',
f
June, pp. 23-27.
1
~
Partheniades, E. (1965). Erosion and deposition of cohesive soils. Journal of Hydraulics
Division, ASCE, 91, No. HYl, 105-l38.
i Patzkowski, L. G. (1992). Rock creek - cresta sediment management project: an overview of
j geotechnical and sediment characterization studies. 9th. Annual Dredging and Placer Mining
Conference, Reno, Nevada.
i
1 Pemberton. E. L. (1978). Reservoir sedimentation. Proceedings of US-Japan Binational Seminar
j on Sedimentation and Erosion, Honolulu, Hawaii.
IA
I
Pitt, J. and Thompson, G. (1984). The impact of sedimentation on reservoir life - challenges in
African hydrology and water resources. Proceedings of the Harare Symposium, IARS
Publication No. 144.
I~ Ritchie, J. C., Cooper, C. M. and McHenry, I. R. (1986). Sediment accumulation rates in lakes
£ and reservoirs in the :Mississippi river valley. 3rd Symposium on River Sedimentation,
!
~ Jackson, Mississippi.
~
f Roehl, J. "V. (1962). Sediment source areas, delivery ratios and influencing morphological
~ factors. IAHS, Publication No. 59.
I!
Rogers, R. S. and Schumm, S. A. (1991). The effect of sparse vegetation cover on erosion and
sediment yield. Journal of Hydrology, 123, Elsevier Science Publication, Amsterdam.
basins.~
~ Roosebooffi! A. (1975). Sediment transport in rivers and river Thesis. University-"of
Pretoria, South Africa ..
Roosebooffi, A. and Mulke, F. 1. (1982). Erosion initiation. Proceedings of Symposium on
Recent Developments in the Explanation and Prediction of Erosion and Sediment Yield, IAHS
Publication No. 137.
Rooseboom, A. (1992). Sediment transport in rivers and reservoirs - a South African
perspective. Water Research Comrrjssion Report No. 297/1/92 South Africa.
Sanderson, M. (1990). UNESCO Sourcebook in Climatology for hydrologists and water
resources engineers, ·UNESCO.
Scheuerlein, H. (1993). Estimation of flushing efficiency in silted reservoirs. Proceedings of the
1st JeHE, Vol. 1, Part A, USA.
Shangle, A. K. (1991). Hydrographic surveys of Indian reservoirs. Report by Central Water
Commission for Workshop on Nlanagement of Reservoir Sedimentation, New DeLhi.
Strand, R. 1. and Pemberton, E. L. (1982). Reservoir sedimentation. US Bureau of Reclamation,
Denver, Colorado.
Swiss National Comrnittee on Large Dams (1982). General paper. ICaLD 14th Congress, Rio de
Janeiro , pp. 889-958.
Tolouie, E., West, 1. R . and Billam, 1. (1991). Sedimen.tation and desiltation in. the Sefid-Rlld
reservoir. British Geomorphological Re.search Group meeting, St Andrews.
United Nations Educational, Scientific and Cultural Organisation (UNESCO) (1977). Atlas of
World Water Balance 1977. Karzoun. VI Editor-in-Chief, Paris.
United Nations Educational, Scientific and Cultural Organisation (UNESCO) (1978). World
Water Balance and Water Resources of the Earth. Leningrad.
147
EVACUATION OF SEDIMENTS
United States Department of Agriculture (USDA) (1943). The control of reservoir silting.
Miscellaneous Publication No. 521.
Varma, C. V. J., Rao, A. R. G. and Natarajan, S. (1992). Sedimentation of Indian Reservoirs. An
assessment. 5th Symposium on River Sedimentation, Karlsruhe, pp. 919-923.
Veltrop, J. A. (1992). The role of dams in the 21st century. US COLD, Denver.
Walling, D. E. (1989). The erosion problem. International Journal of Sediment Research, 14, No.
1, 1-11.
Walling, D. E. and Probst, J. L. (1997). Human impact on erosion and sedimentation. IAHS
Publication No. 245, Wallingford, UK.
Water Power. (1979). World's largest capacity reservoirs. Intenwtional Water Power and Dam
Construction, 31, Nov., 98.
White, W. R. and Rofe, B. H. (1996). Responsibilities and lists associated with large reservoirs.
Intenwtional Conference on Aspects of Conflicts in Reservoir Development and Management,
London.
Wolman, M., et al. (1989). Erosion control and reservoir deposition. Proceedings of Bilateral
Seminar on Problems in the Lower Reaches of the Yellow River, China. pp. 139-163. Kluwer
Academic Publishers, The Netherlands.
148
p
Appendix I.
Reservoir data
!
!
This appendix provides data which ICOLD has obtained from member
countries. Table A1.1 provides, on a country by country basis, data on
reservoir capacity development in the twentieth [century. Table A1.2
provides sedimentation data for individual dams.
lSi
m
Table AI.I. [COLD data - world total storage volumes (M.m 3 )
~
()
Region Country Notes Totals <1900 1900--9 1910--19 1920--29 1930--39 1940-49 i950-59 1960-69 1970-79 1980-89 1990-1998 ute
C
No. Volume Av. vol No. Volume No. Volume No. Volume No. Volume No. Volume No. Volume No. Volunic No. Volume No. Volume No. Volume No. Volume No. Volumt!
~
Africa
Africa
Angola
Benin r
IS
2
9446
1734
630
867
3 142 4 44 3
I
413(>
24
5
I
5124
1710 o
Africa Botswana 3 231 77 2 146 I 85 z
Africa
Africa
Cameroon
Congo
9
2
14325
44
1592
22
0 0 0 0 0 0 0 0 0 0 0 0 I
I
0
8
0
I
0
36
4 3410 4 10915 0 0 0 ()
o
Aldca
Al'rica
Congo (ORC)
Ethiopia
12
8
5319
2624
443
328
0 0 0 0 0 0 0 0 0 1267 I 0 5
I
2417
7
I
3
35
1925
3
4
0
692
I 1600 I 0 0 0 "
V1
m
Africa
Africa
Gabon
Ghana
I
4
220 220
150279 37570 0 0 0 0 0 0 0 0 ' 0 0 0 0 0 0 I ISO 000 2 192
I
I
220
0 0 0 0 87 o
Africa Guinca 2 237 119 2 237
I 1100 2 14 28444
3:
Africa Ivory Coast 22
15
37917
2428
1724
162
0 0 0 0 0 0 0 0 0 0 0
I
0
I I 16 I
9
16 2 170
3
9
8339
2155
0
I
0
70
2 25 m
Africa Kenya
Africa Lesotho 7 1970 281 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 3 0 0 I 4 I 1950 3 n Z
Africa Liberia I 0 0 I 0 -f
V1
Africa Madagascar 10 431 43 0 0 0 0 0 0 0 0 I 125 0 0 7 293 <> 0 2 13 0 0 0 0 0 0
Africa Mal awi 4 4 I 2 2 I I I I
Afiica Mali 2 13 440 6720 I 2170 I I 1270
Africa Mlluritius 9 61 7 I 6 2 7 2 8 3 36 I 4
Africa Mozambiljue 6 57 102 9517 I 1536 I 33 3 54260 I 1273
ArriclI Namibia 13 662 51 I 2 I 4 I 295 6 222 l' 98 I 41
AfriclI Nigeria 45 40414 898 0 0 0 0 0 0 0 0 0 0 2 14 0 () 6 15079 23 14856 8 9331 0 () (, 1134
Afril:a Scnegal 2 11520 5760 I 250 1 I 1270
Africa Seychelles 2 1 I 1 0 I I
Africa Sierra Leone I 22 22 I 22
Africa South AI'rica 539 30583 57 2 144 5 12 11 164 26 679 22 3534 15 173 61 707 104 1824 125 16053 124 2194 44 5099 0 ()
Africa Sudan 5 5587 1117 0 0 0 0 0 0 I 931 I 0 0 0 0 0 2 4656 0 0 0 0 0 0 0 0
Africa Swaziland 6 250 42 I 50 3 23 2 177
Africa Tanzania I 1135 1135 I 1135
Africa Togo f 2 I I I I I 0
Africa Uganda I 200 200 I 200
Africa Zambia c 3 47 16 0 0 0 0 0 0 I 41 0 0 0 0 0 0 0 0 2 6 () 0 0 0 0 ()
Africa Zimbabwe c 213 187117 878 0 0 I 4 0 0 2 39 2 39 5 56 8 0 21 183266 36 1431 51 1016 82 1201.1 5 66
Africa Total 967 575352 595 2 144 6 16 11 164 31 1696 27 4967 26 251 96 6440 164 358850 235 124018 219 46580 132 30900 17 1326
China China 1851 649322 351 4 2 1 1 2 8 1 6 I 15 3 24057 177 26113 479 256549 583 103536 220 58090 120 57278 260 123667
China Total 1851 649322 351 4 2 1 I 2 8 I 6 1 15 3 24057 177 26113 479 256549 583 103 536 220 58090 120 57278 260 123667
S. Asia Total 4131 318602 77 41 2368 41 983 85 2560 46 3388 40 6658 36 557 224 30088 446 71020 1152 72 812 1200 41956 137 1093 683 85119
- )
I
.-
Mid. E"sl Aruh.'nill Iii I7H9 112 I I 4 131 7 787 3 593 I 277
Mid. Easl Azerbaidj'lIl 17 183K5 1081 I 160011 7 190 8 2182 I 13
Mid. Easl CYI'I1IS 5'2 2B6 6 0 U 0 0 0 () II () (I 0 0 0 5 I 16 29 4 17 21 231 6 8 0 0
Mid . 1;",1 (iL."Ofgia 14 2S!!1 20ti I 312 '2 120 '2 350 3 9 4 1206 2 885
Mid.E'lsl 11';111 66 32935 49<) 0 0 0 0 I) I) II 1/ () 0 0 n I 5'/ 5 5568 8 8551 /0 1756 28 4371 14 12 632
(vlid.E:ISI Inil( n ':>7183 4399 I 6800 3 3053 I 26000 8 21330
Mid. Easl JII.-.I:1I1 8 203 25
Mill. Hasl
3 II I 52 I 80 3 60
l.d>:IIIOII 5 161 32 0 0 0 0 0 0 0 0 0 II 0 0 2 I 3 160 0 0 0 U 0 0 I) 0
Ivlid. ['ast Saudi Ambia 311 2114 7 2 1 I I 12 76 15 165 II 39
Mid. Fasl Syri'l 41 342 a 0 0 0 0 () 0 II 0 () 0 0 0 0 0 13 327 6 14 7 U 6 0 9 0
Mid. E'lSI Turl,ey 1i25 109233 175 () () n II 0 () 0 (I 2 18 1 2 6 0 29
---'-'- - - - - - - - ---- I- -- -- ---
12241 72 3490 151 18437 295 64 050 69 \0 995
Mid. Easl
- - --
'lblal
-----.-
895 22:161B 2511 n 0 0 0 0 0 0 n 2 18 2 314 21 2298'2 86 22061 122 41 178 221 43811 335 68429 \06 248SS
-- -- -. --
N. "f,-il' a 1\lgcri;1 99 4732 48 3 5 I I 0 II (I 0 0 '/31 6 106 t--
3 165 9 75.) 6 259 40 2238 31 473 tl n
N. Arri.:a EUypl 6 Ifl71100 27833 0 0 'l
0 t) II I) 0 '2 5000 0 II I) t) II 0 I 1620nO 0 0 I 0 () 0
N. Africa Lihya 12 3]6 28 I) 0 0 tl n 0 0 0 (I 0 0 0 n II 0 0 II 290 4 41i 0 0 0 [I
N. A!'rica t\'tllruccCl 92 I~ 54\) 15~ t) 0 n IJ 0 0 I 2 5 373 I 27 6 1414 5 595 8 6423 37 563 29 5143
N. Afrim "ruilisia 8tl l11fi5 II I) (I II 0
0 I) () 0 II 0 0 0
- ---
N. Afri':3
.- - - - -
'lllial
-- ' -'
289 18~ 47:1 652
- -I 0 0 I 331 6 197 0 0 12 1137 61 199
"I -0 0
3 5 3 0 0 I '2 7 6104 7 133 1911
_- _·--0_._.- --._0 \0 20 15H 23 168 972 93 3984 122 5815 II
N. AIII('ri~a Call:llla 1111 I 3~6 860 1597 9 6115 10 48 ·12 12247 85 17679 82 8176 55 32428 1-17 211691 128
(- -
432092 142 455767
----
N.A llwric'l USA 116 155495 15 1i32 0 II
6T!4 517 670 81 665 47.'i~ 132 1638 246 III 747 296 19284 332 61520 294 52495 Tl8 125221 1544 160663 1251
--- 51 \07 641 16 MIl 195 )-125 0 0
N. AlIler;":1 'lhlal 7205 IM'14530 t--
._-
256 674 5364 142 36R6 28K 30994
r---'t - - 1-- - --. - . -
381 36~63 414 6')696 349
--
84923 925 336912 1672
~ - --
592755 1393
-~--.--
r--
506874 757 172 306 2111 4057 0 0
N. ElIr<lp" lkigilll\l 16 252 Iii 0 0 () 0 0 0 I II I II I 25 I 3 2 5 Y 132 1 68 0 0 0 0
N. Ellrllp" [klllll;1rk 111 1929 19:1 0 (J 0 0 0 0 0 0 0 0 0 0 '2 5 4 7 0 0 3 17 I 1900 0 0
N. Ellmp~ I'illiand 55 19 Inn 3'16 () 0 0 II Il 0 I 2 2 21 6 5726 14 7639 16 2953 II 2539 2 77 3 72 0 0
N. EllfllpC Fr:mc.: 569 160SY 28 35 141 17 56 9 CiK 26 158 35 826 38 906 75 2755 90 3998 93 2054 96 1322 54 3133 I 672
N.Ell rll!,c Gl'I"IIHIIIY 311 <In!) 14 15 15 14 117 14 403 14 92 27 60S 9 2\J 34 850 49 SKI 76 639 48 463 6 181 5 100
N. Ellrupe kdand IS 2679 149 0 I) 0 0 0 0 () 0 0 II I 31 Il 0 2 \3 4 1575 7 640 4 420 0 0
N. ElIwpe Ireland 16 994 62 3 14 0 tl (I 0 I 6 3 654 '2 j 4 307 0 0 2 3 0 () I 5 0 0
N. EUI"llI'" Lllvi" 5 H66 173 I 16 ')
5\0 2 340
N.l'lIl"1lp" Lillluania 20 6"1 31 3 462 12 65 4 4K I 46
N. Europe Luscmhllrg 3 79 26 0 0 0 0 () 0 I) 0 0 ()
0 II 0 0 3 79 0 0 0 0 0 0 0 0
N. Eul'Opc Mllld'lI'ia ~ 1220 6\0 I 485 I 735
N. Eurup~ Nl'lherlalld,; 9 9259 1029 () 0 0 0 0 II 0 0 I 51:2() 0 II :2 35 3 204 2 1I2U 1 2780 0 0 n 0
N. Eurllpc Nlww:IY 321l 53496 163 J ' :\ .5 1221 \I 543 14 1532 6 162 14 954 37 2994 101 11 765 75 9154 51 2U944 6 4044 5 175
N. Eurnpl! Polalld 69 3598 52 I 5 I 15 5 83 8 54 9 201 2 276 3 1811 9 883 10 724 8 2RO ,9 653 236
N. Europe
N. EUIll(1c
Rllssia
Slnwnia
91
28
731 637
206
R040
7
0
3
0
(I
0
(I
II
0
0
II
0
/I
I
2
30110
5
15
I
1859
.5
S
3
10281l
27
20 225700
47
19 173475 IJ 162265 5 10493 1. 327liU "
\I 111805
N. EllrClfX'
N. Europe
Sw'-'lIcn
Ukr:linc
19[1
20
3:l216
46 B8'1
201
234 '1
0 0 5 482 III 612 3 177 9
1
10929
3320
18 1006 "
0\7
3
5329
18216
4
53
7
41
7927
19 7~('
4
30
6
20
11028
2510
6
12
2
52
631
92
I
3
<)
95
()
0
I
0
JOUO
II
N. Europe UK 517 6864 13 188 52Y 56 281 22 9') 19 168 43 74J 17 166
---- [ - . _ - - .--- - - 68 0 58 38!li! 32 541 II 4011 3 69 II 0
N. Europe TOlal '1277 9JS 168 7(1
t----. ._-
412 248 98 214·2 71 1S1.S 90 5202 153 24457 117 19631 315 264 553
---.--- - -- --- . - - - - - - - --- I - : - - - - - -- (--
425 226517 382 195444
-
257 38307 93 43341 28 116034
Pacilic l~illl Auslralia 484 89855 186 33 240 14 627 21 144 21 5102 21 3526 16 1187 S'S\()
51 93 121'13 R6 45850 92 9808 36 271H 0 0
j'al'ifiL: \(jill Chik H7 10 775 124 2 10 I 95 6 30 II 24 20 459 5 71 7 1608 12 5901 II 64 14 1222 7 291 2 10011
P:tdlkRil\l Fiji 2 153 17 2 153
Pacific Ilim Indonesia 9ti 74348 774 0 0 I 2050 1 17 5 35 5 106 4 3(1'l3 2 128 I 2556 10. 574 27 9277 27 14515 \I 41997
Parili.: Rim Japan 1077 21 OKS :!I) 6 J2 2 2 7 10 18 151 ,II 255 37 832 137 3504 208 6444 202 3586 165 1579 198 3004 56 16119
1"lcilkRill1 J{orl~il 7(-,5 161161 22 ·0 0 I I 0 0 13 55 22 86 62 1361
P;ll'ilic Rim
Pacilie Rim
I<llr~:llN)
N. Z"'II:md
70
86
34312
16515
490
192 2 0 4 2 2 52
3
II
114
161
9
9
2332
322
7 2112
102
12
191
16350
140
18
1138
4316
170
12
5453
1053
'187
6
63911
3139
68
3
217K
6726
0 II
»
-0
3 94 \0 nll5 12 2907 14 8978 19 614 0 0 II 0
Pa.:ilic Rilll I'. N.Guillca 3 333 III -U
I 0 2 33) m
Pa~ilic Rim l'elll 43 3301 77 .6 51 0 0 I II I 12 2 31 3 31 4 322 8 875 5 1069 4 690 3 157 6 52
P;wilic Rim Phjlipllincs 15 6714 44H I 12 3 424 I 1!50 4 3003 6 2425
Z
Pacilic Ililll
- --
'Ihiw;ln
._-50 2808
- - -- .
57 4 160
--
I I 6 332 4 II 8 198 6 367 I) 1)95 9 505 3 299 o
Ln PHeili. Rim 'lillal :ms 277 124 100 53 493 23 2777 40 264 7J 5655 135 ~449 142 6(i74 336 341)20 5(11) 37,197 525 70958 531 3581U 3115 29888 75 X
W --- 4'1738
m
~
n
c
Table Ai.i. continued ~
Region Country NOles TOlals <1900 1900-9 1910-19 1920-29 1930-39 1940-49 1950-59 I 1960--69 1970-79 1980--89 1990-1998 utC o
Av. vol No. Volume No. Volume No. Volume No. Volume No. Volume No. Volume No. Volume No. Volume No. Volume No. Volume No. Volume No. Volume
z
No. Volume
------ -~~~~~~ ~ ~ ~ ----
---- o
..."
S. America Antigua I 5 5 I 5
(.I')
S. America Argentina a, d 100 130380 1304 0 0 0 0 0 0 3 32 3 615 14 1057 14 447 17 4204 27 79705 18 9530 3 34593 I 197
Bolivia 7 372 53 0 0 0 0 0 0 0 0 0 0 3 161 0 0 2 151 0 0 I 6 () 0 I 54 m
S. America
S. America Brazil 586 540386 922 77 1874 3 179 16 540 35 2029 35 3856 16 1665 100 11825 103 98590 91 135220 60 161247 21 85063 29 38298 o
S. America Colombia 49 10913 223 0 0 0 0 0 0 0 0 I 3 0 0 9 250 10 1593 10 3821 14 3162 2 85 3 1999
S.America Costa Rica 9 2288 254 0 0 0 0 0 0 0 0 0 0 0 0 I 0 2 51 I 2220 2 4 3 13 0 0 3:
S. America Cuba 49 3755 77 I 7 I 4 I 5 15 782 31 2957 m
S.Americ:n Dominican R. II 2373 216 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6 1141 3 1057 2 175 0 0 Z
S.America Ecuador 31 70()3 226 I 6 I 12 22 190 4 6330 3 465 --i
(.I')
S. America EI Salvador 5 2430 486 2 620 I 0 I 1430 I 380
S. America Guatemala 4 460 115 0 0 0 0 0 0 1 0 0 0 0 0 0 0 I 0 0 0 2 460 0 0 0 0
S.America Guyana 2 44 22 2 44
S. America Haiti I I I I I
S.America Honduras 8 9035 1129 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 1940 4 10 1 7085 0 0 0 (I
S. Europe Total 3220 145162 45 85 182 30 84 76 852 138 2344 136 3910 143 3922 424 21629 629 42764 613 29163 559 18395 278 11330 109 10587
----
,.
S.E. A~ia Brunei '2 45 22 I I I 44
S.B. A$iu C:unhmlia '2 23'1 119 I 14 I 223
S.E.Asia Lnos I 7030 7(nn I 7030
::i.E. Asia Millays; .. 59 28 ')(,0 491 4 I 0 0 () 0 0 0 2 96 () 0 2 44 18 2442 4 6657 19 19720 n 0 0 0
S.E.Asi:1 M,YIIIIHlur 5 2325 465 I 75 I B27 3 1423
S.r:. Asia Sil1gllllure :I 75 15 1 28 I )0 I 17
S.E.l\siil Thailand 204 78534 385 2 '11 0 0 0 Il 0 0 0 0 0 0 0 0 II 18536 20 30976 135 27 ROI 3ri 1150 0 0
S.n.Asia VkWiltll I 16.5 165 I 165
-1- ----
~
~~
S.I1.Asia 1blal 277 117 311 424 6 0 II (J 0 0 0 2 96 I 75 '2 44 33 21 18(i 27 49005 37 1J73 0 0
.,'--
Worltllllllli 25432 6464730 254 1208 11308 1355 9986 600 411117 808 57579 979 131 139 911 162241 2734 781847 4793 I 838603 5425 I 673793 '1426 872998 1867 J9478K I32S 488566
NOles:
:\ AI'IlClllill:t ~.,c1udesYm·yr~la 21 OIlO M.llll underconslruclion - indudcd in Paraguuy
b Paraguuy ex<:1udt'sllllil)U 24 (Jon M.II1) (1983) - included in Bl1Izil
c Zmllbiu .:xclmJes Kuritm 1806()O M.m' (1959)- included In Zimbubwil
d Ul1IglIUY .:xllld.:s SUlll~ Grande 5500 M.ll1) (1979) - included ill Argenlinn
e Yugu~lavia excludes Djcrdall 1 255 M.m' (1972) and Djerdap 11868 M.m' (1987) - included ill Homania
r 'lllgU exdudes Nangheln 1710 M.nl) (1988) - il1c1uded io lIel1in
[tala cXlracl.::t1 from ICOLD World ,.esister of ifillll.r (1998)
»
'1J
-0
m
Z
o
X
EVACUATION OF SEDIMENTS
China Sanmenxia - Yel10w River 688000 16200 1960 1978 5450 33·6% 440·1 302·78 1·87%
China Sanmenxia II 9640 1960 1989 5690 59·0% 196·21 2·04%
China Hongshan 2560 1960 1987 670 26·2% 24·81 0·97%
China Guating - Yongding River 43402 2270 1953 1994 630 27·8% 354·0 15·37 0·68%
China Fenhe 721 1960 1989 330 45·8% 11·38 1·58%
China Liujiaxia - Yellow River 181 800 5720 1968 1989 1410 24·7% 369·3 67·14 1·17%
China Yanqouxia - Yellow River 182700 216 1958 1968 J61 74·5% 88·1 16·10 7-45%
China Qingtongxia - Yellow River 285000 606 1967 1980 566 93-4% 152·8 43·54 7·18%
China Bapanxia - Yel10w River 215900 49 1975 N/A 15 30·6% N/A
China Danjian kou 16000 1968 1986 1130 7·1% 62·78 0·39%
China Celian 200 1960 1983 205 102·5% 8·91 4·46%
China ZhenziJiang 36 1959 1973 29 80·6%. 2·07 5·75%
China Naodehai 168 1963 1986 2 1·2% 0·09 0·05%
China Gondzui 357 1967 1987 206 57·7% 10·30 2·89%
China Bikou 521 1976 1986 218 41·8% 21·80 4·18%
China Shimen 105 1973 1988 28 26·7% )·87 1·78%
China Hongsiba 34 1960 1986 7 20·6% 0·27 0·79%
China Wangyao 203 1972 1990 77 37·9% 4·28 2·11%
China Fengjiashan 389 1971 1990 63 16·2% 3·32 0·85%
China Miugong 183 1960 1989 97 53·0% 3·34 1·83%
China Dongxia 77 1959 1983 41 53·2% 1·71 2·22%
China Shixiakoll 175 1959 1988 35 20·0% 1·21 0·69%
China Chang shan tao 348 1960 1986 47 13·5% I·S1 0·52%
China WenYllhe 105 1959 1988 20 19·0% 0·69 0·66%
China Bajiazui 496 1958 1990 249 50·2% 7·78 1·57%
China Yang maowan 120 1970 1990 17 14·2% 0·85 0·71%
China Zhaikou 185 1970 1990 8 4·3% 0·40 0·22%
j
China Luhun 1320 1960 1983 62 4·7% 2·70 0·20'11.
i
156
J
J
APPENDIX I
Start
I End M.m" rn"lkm!/yr M.mJ/a,.'l11 Mtlann % vol.
China
I
I Total I 42804 I 12013
I 28·1% 510·71 1·/9% I
ChilUl Total-on3!) reservoirs 80986 I
1981 ! 1500 14·2% 800·CO 0·99%
China 83 357 reservoirs with 460000 M.m' storage willI an aver.lge annual loss af 2·3%
Netherlands HlIringvUet
I
1000 \970 1999 70 1·0%
I 2·41 0·24%
Spain Jucar - Alarcon 1112 1955 1984 -92·742 -8·3% -3·20 -0·29%
Souin Guadatquillir - EI Pint:ldo 202 1942 1995 -10·835 -5-4% -0·20 -0·[0%
Spain S~ra - Sllntomera 26 1965 1993 --4·962 -IS·8% -0·18 -0·67%
Spain Sur - Guadalteba 149 1973 199[ -4·03 ~·7* -Q~:!2 -0·15%
Soain Guadalquivir - Ameen!! 123 1969 1977 -3·724 -3·0% -0·47 -0·38%
Spain Guadalquivir - NegrJ.tin 570 1984 1990 -3·478 -0·6% -O·5S -0·10%
SOllin Douro - Barrio. de Luna 308 1956 1985 -2·014 -0·7% -0·07 -0·02%
Spain Guudaiquivir - Los Hurones 138 1962 1%9 -2 -1·4% -0·19 -0·21%
Sp:!in l'ugus - i'lorbollon 86 1954 1990 -I·n'::: -2~1% -0·05 -0·06%
Spain Guadian:l - Gasset 23 1909 1983 -0·504 -2·2% -0·01 -0·03%
0·00 0·00%
I
Spain , T:lb'lls - Buendia 1520 1957 198J 0 0·0%
Spain ; ucar - Alcon 2 1958 1976 0·()75 3·8% 0·00 O':H%
Spain Ebro - La Tranqu:::ra 1870 84 1960 1994 0·121 O·I%- 0·00 0-00%
Spain Sur - R~(Ie"'o.do 2 1974 1992 0·154 8-6% 0·01 0·480/"
Spain CaL:1.lonia -'"Riudecanll5 3 19i5 1981 o·un 5·6% 0·00 0·09%
Spain S<!gura - Camarillas 579 35 1960 1993 0·189 0-5% 0-01 0·02%
Spain Norte - Rioseco 4 1973 1994 0·285 6-7o/c (}O2 042%
Spain 7 1959 1979 0·326 4·7% 0·02 0·23%
Spain I Tagus 4 1948 1968 0·561 12-8% 0·03 0·64%
Spain Ta!<u$ - Palmaces 275 32 1954 1984 0·629 2·0% 0·02 0·07%
Sp;z.in Tagus - El Vado 379 57 1972 1979 0·691 1·2% Q.1O 0·17%
Spain Jlicar - Amadorio 17 1960 1991 0·723 44% 0·02 0·14%
Spain Ebro - Arguis 3 19:.8 1980 0·7:58 25~3~c ()·O2 0-60%
Jucar - .~eIlQS 1311 1979 1994 0. 793 0·6% 0·05 0-04%
Saain
Iuco.r - Bus(!o 220 8 1912 19S0 0-807 \0·-1% 0·01 0·15%
Spain
4- 1927 1979 0-S4t 21·0% 0·02 0-40%
Spain Ebro G:tl!iu~uen
330 H,) 1973 [981 e·9! 9-1% 0-1 I 1-14%
Spain
luco.r - A:ouillo de Dan Bias 690 2:2 1960 1988 0·965 4·4% 0-03 0·:6%
Spain
9 1990 1994 0·994 10·6% 0·25 2·~%
Spain NOlie -Alfilorios
SP:1!tl Gu:!d:llqtlivi~ - G'.lado.lmena
I 1350 3J..7 1969 1 1989
H), 0·3%
I 0·05 0·02%
157
J
EVACUATION OF SEDIMENTS
]
Table Al.2.
Country
continued
Reservoir Catchment: Capacity: Survey dates Total Vol. lost: Annual sedimentation Notes
J
km 2 M.m; sedimentation: %
Start End M.mJ m"/km 2/yr M.mJ/ann Mtlann % vol. - -J
Spain
Spain
Catalonia - Foix
Guadalquivir - Gergal 36
6 1928
1979
1983
1985
1·158
1·305
19·3%
3·6%
0·02
0·22
0·35%
0·60%
j
Spain Guadalquivir - Bennejales 300 104 1958 1978 1·395 1·3% 0·07 0·07%
Spain Jucar - Forata 1058 39 1969 1983 1·542 4·0% 0·11 0·28%
Spain Segura - Argos 12 1970 1991 1·666 14·2% 0·08 0·68%
Spain
Spain
Ebro - Santa Maria de Belsue
Segura - Anchuricas
190 13
8
1931
1957
1980
1979
1·742
1·759
13·4%
22·0%
0·04
0·08
0·27%
1·00%
'-1
Spain Sur - La Vinuela 117 170 1986 1994 1·799 1·1% 0·22 0· 13%
Spain £bro - Santolea 1221 49 1932 1993 1·851 3·8% 0·03 0·06%
Spain Ebro - Moneva 10 1939 1984 1·991 19·9% 0·04 0·44%
Spain
Spain
Guadiana - Valuengo
Jucar - La Toba
20
II
1959
1935
1985
1980
2·131
2·174
10·7%
19·8%
0·08
0·05
0·41%
0·44%
-1
Spain
Spain
Guadalquivir - CubiJIas
Segura - La Cierva
737
167
21
8
1956
1929
1990
1987
2·299
2·429
10·9%
32-4%
0·07
0·04
0·32%
0·56%
1
Spain Catalonia - San Pons 25 1957 1968 2·591 10·4% 0·24 0·94%
Spain Douro - Burgomillodo 800 15 1953 1989 2·603 17·4% 0·07 0·48%
Spain Jucar - Sichar 52 1960 1976 2·729 5·2% 0·17 0·33%
Spain
Spain
Spain
,
Guadalquivir - La Bolera
Douro - Linares del Arroyo
Jucar - Guadalesl
187
760
60
56
58
16
1967
1951
1965
1979
1980
1989
2'828
2·954
3·008
5·1%
5·1%
18·8%
0·24
0·10
0·]3
0·42%
0·18%
0·78%
1
Spain Guadiana - Torre de Abraham 761 60 1974 1988 3·15 5·3% 0·23 0·38%
Spain Ebro - Las Torcas 9 1946 1979 3·195 35·5% 0·10 1·08%
Spain Guadalquivir - Cala 59 1927 1984 3·603 6·1% 0·06 0·11% 1
Spain
Spain
Ebro- Pena
Guadalquivir - La Minilla
64
965
22
60
1930
1956
1989
1984
3·619
3·64
16·8%
6·1%
0·06
0·13
0·29%
0·22%
j
Spain Guadalquivir - El Tranco de Beas 500 1945 1990 3·675 0·7% 0·08 0·02%
Spain Jucar - Maria Cristina 1334 23 1920 1991 3·75 16·1% 0·05 0·23%
Spain Norte - Penarubia 12 1961 1994 3·788 31·6% 0·11 0·96%
Spain
Spain
Spain
Jucar - Beniarres
Sur - Guadalhorce
Tagus - CazaJegas
469
965
1955
31
134
11
1971
1972
1949
1991
1991
1990
3-831
3·97
4·052
12·4%
3·0%
36·8%
0·19
0·21
0·10
0·62%
0·16%
0·90%
]
Spain Guadalquivir - Bernbezar 1665 347 1963 1994 4·899 1·4% 0· 16 0·05%
Spain Guadalquivir - Puente Nuevo 559 289 1972 1994 4 ·97 ].7% 0·23 0·08%
Spain Tagus - Guajarez 375 25 1971 1982 5·391 21 ·6% 0·49 1·96%
Spain
Spain
Spain
Spain
Tagus - Entrepenas
Guadalquivir - Torre del Aguilla
Segura - Cenajo
DOUTO - Agueda
439
1060
8911
70
472
22
1956
1947
1960
1931
1979
1992
1992
1980
5·611
5·643
6·403
6·582
0·1%
8·1%
1-4%
29·9%
0·24
0·13
0·20
0·13
0·00%
0·18%
0·04%
0·61%
J
Spain Ebro - Cueva Foradada 644 29 1926 1992 6·617 23·1% 0·10 0·35%
1
Spain Jucar - Benageber 3495 228 1955 1992 6·663 2·9% 0·18 0·08%
Spain Guadiana - Zujar 309 1989 1994 7·122 2-3% 1·42 0·46%
Spain Ebro - La Estanca de AJcaniz 14 1944 1971 7·133 51·0% 0·26 1·89%
Spain Ebro - Sotonera 323 189 1963 1986 7·288 3·9% 0·32 0·17%
Spain Jucar - Ernbarcaderos 16952 9 1952 1983 7·539 83·8% 0·24 2·70%
Spain Catalonia - Sau 177 1963 1979 8-495 4·8% 0·53 0·30%
J
Spain Guadalquivir - Guadelen 1281 173 1954 1977 9·718 5·6% 0·42 0·24%
Spain Tagus - EI Burguillo 209 1931 1991 10·936 5·2% 0·18 0·09%
Spain Sur - Conde del Guadalhorce 268 78 1921 1991 11·051 14·2% 0·16 0·20%
Spain Guadalquivir - Guadalcacin 77 1917 1969 11·972 15·5% 0·23 0·30%
Spain Ebro - Ribaroja 219 1969 1982 12·224 5·6% 0·94 0·43%
Spain Segura - Talave 763 47 1918 1993 12·344 26·2% 0·16 0·35%
Spain Segura - Valdeinlierno 31 I 25 1897 1984 12·473 49·9% 0·14 0·57%
Spain Tagus - Gru;el y Galan 1846 924 1961 1990 12·84 1·4% 0·44 0·05%
Spain Douro - Santa Teresa 1858 496 1960 1989 13·387 2·7% 0·46 0·09%
Spain Tagus - Riosequillo 49 1956 1970 14·024 28·9% 1·00 2·07%
Spain Guadalquivir- Bornos 1361 215 1961 1990 14·815 6·9% 0·51 0·24%
Spain Ebro-Oliena 2694 101 1959 1985 15·18 15·0% 0·58 0·58%
Spain
Spain
Spain
Spain
Guadalquivir - La Brena
Guadalquivir - Guadalmellato
Guadalquivir - Pedro Marin
. Segura - Alfonso XII
465
1195
420
852
116
163
19
42
1935
1965
1954
1916
1991
1992
1977
1985
15·869
16·323
17-893
18·184
13·7%
10·0%
94·2%
43-3%
0·28
0·60
0·78
0·26
0·24%
0·37%
4 ·09%
0·63%
1
Spain Segura - Puentes 1042 32 1884 1985 18·726 59·3% 0·19 0·59%
Spain Jucar - Contreras 2344 872 1975 1994 19·595· 2·2% 1·03 0 ·12%
Spain
Spain
Spain
Spain
Ebro- Yesa
Guadalquivir - Dona Aldonza
Tagus - San Juan
Ebro - Barasona
2181
3766
863
1250
471
23-
162
71
1960
1955
1955
1932
1986
1977
1992
1993
20·78
220439
24·258
24·764
4·4%
97' 6%
15·0%
34·9%
0·80
1·02
0·66
0·41
0·17%
4 ·43%
0·40%
0·57%
J
Spain Segura - La Fuensanta 1201 235 1933 1991 25 ·273 10·8% 0·44 0· 19%
J
Spain Ebro - Talarn - Tremp 258 1916 1990 69·592 27·0% 0·94 0·36%
Spain Ebro - Mequinenza 1530 1966 1982 92-822 6· 1% 5·80 0·38%
Spain Guadiana - Cijara 7456 1670 1956 1983 138· 111 8·3% 5· 12 0·31%
Turkey Caygoren 1510 130 \1971 1999 24·6 19·0% 583 0·88 0·68%
1·77%
4
]
I ::~
Turkey Demirkopru 6590 814 1999 563·6 69·2% 2193 14·45
Turkey Buldan 180 46 1999 31·5 68·5% 5467 0·98 2·14%
Turkey Kerner 2500 373 1954 1999 209·8 56·3% 1865 4·66 1·25%
Turkey Yalvac 133 9 1973 1999 5·9 68·4% 1697 0·23 2·63%
Turkey Kararnanli 164 25 )973 1999 9·9 39·7% 2311 0·38 1·53%
J
158
1
1
APPENDlX I
j
Turkq Sdc:vir 721 61 1965 1999 33-9 55-8% 1379 1·00 1-64%
Turk~y Cubuk- I 660 6 1936 1999 12-6 215·0% 303 0·10 3-57')I>
Turkey Bayi ndir 70 7 1965 1999 3·5 50-0% 1464 0-10 147%
Turkey Hilfanli 261 70 5980 1959 _1999 3936-0 65-8% 3760 98·40 1·65%
Turk~y Kcssikkopru 360 95 1966 1999 28-1 29·6% 2364 0-85 0-90%
Turkey Alcinapa 589 3:! 1967 1999 14-2 43-9% 752 0·44 1·37%
Turkey Sevhan 19 :!54 120U 1956 1999 429·7 35·8% 519 9-99 0·83%
Turkey Kmalkaya ~~ 11 30 148 1972 1999 42-1 28·4% 1379 1·56 1·05 %
Turkey Cip 236 10 1965 1999 J.6 37·1 % 3i O 0·1 1 1·09 %
Turkey Surgu 2i5 71 1969 1999 28·0 39·4% 3388 0·93 1-31%
I""
Germeny Bal c!eney 8 770 0-06
G~r:nany
Germolny
Germany
Ba\'aria - Forggensee
Bavaria - Saalachsee (flushed)
B :lv:ui~ - Sylven5rei nsee
14;;
940
I 150
4 1913
1965
:960
2408
3-644
1-6%
100-L%
154
93-6
0·22
0·09
1 ' 15%
2-5t %
I US 32 1 i959 1966 1· 1 i·3% I ;49 G' ll Q·4[0/0
I I I I
Germany Torn I
I 236
I 7 3·0% 043 0·20%
I Ring let I
1 1990
19~
Malaysia 183 165 0·030 6
Malaysia Ringlet 183 1980 275 0·050
Malaysia Ringlet 183 550 0· 101
M~ lolysia R i ng! ~l 183 I 1996 1360 0·249
Malaysia To~1
I I
Tunisia Kasseb 101 82 1969 I 12-38 15· 1% 4228 0-427 0·42%
Tunisia EI Kebir 271 26 1925 23·74 91 ·3'70 1200 0·3252 0·1 2%
Tunisia Mellegut: 10 300 332 1954 232-;6 70·1 % 514 5·29 ()'05%
Tunisia Nebhana 855 86 1965 54· 12 62·6% 1918 1·64 0·19%
Tunisia Bezlk 84 6 1960 6· 23 96·5% 1952 0· 164 0·20%
T~~ ~si:\ Chiba 64 8 1'765 7·43 94·5% 3516 0·225 0-35%
Tunisi a L:tkmess 127 8 1966 10·02 125-2% 2465 0·313 0·25%
Tunisia ~rasri 53 7 1968 6·06 83-9% 3811 0·202 0·38%
Tlmisia .sou Her1.'na 390 11 8 1976 11 ·00 9·4% 1282 0-5 0-13%
Tunlsia Iuumine 418 130 1983 i ·50 5-8% 1196 0·5 0· L2%
Tunisia Lebna 99 29 1986 6·00 20·7% 5051 0·5 0·5 1%
Tu nisia Sisi Saad 8950 209 1981 9 1·80 43·9% 603 5-4 0·06%
Tu nisia Sidi Salem 18 250 ~55 1981 56·1 0 10·1'70 18L J.3 0·02%
Tanisin Silianu 1040 iO 1987 38-50 55·0% 3365 3·5 0·34%
Tunisia Marg'Jellil 1110 110 1990 44·00 40·')% 4911 5·5 0-49%
UK 95 reservoi rs surveyed
I I 0· 10% 7
r~l y To~i
I 1
I
I O - I~
USA 190 reservo irs 0 co 12 -33 m' 425 I 1 I! ·8 0 25·7% 0·02 3·56%
USA 257 reservoirs 12·33 to L23·3 mJ 1132 11 14·3 2 19·3% 0· 15 2-00%
US A 283 reservoirs 123·3 to L233 mJ 94 12 138 23·5 23 16·5% 0·97 1·02%
USA 176 reservoirs L233 co 12 330 m~ 39 345 700 20·9 94 1304% 4·49 0·81%
US A 107 reser/oir5 12 330 to 123 300 mJ 2265 i6 5331 23-6 4SB 9·1% 20·69 0·43%
USA 69 reservoi rs 123 300 co I 233 (JOO m' 454897 29249 l804 1066 3·6% 57-92 0·23%
USA I 23 reservoirs> I 233 000 m' 477 14B 74499 19· 1 2563 3-4% 134-44- 0·1 6%
USA Total
I 1203 875 109 980
I 18·6 424 1 3-')% 218-69
I 0-20%
159
EVACUATION OF SEDIMENTS
Mormeo
Morroeo
ldriss ler
I Abde1moumen
I
3680
1300
1217
216
1972
1981
1986
1987
30·94
1·38
I
2·5%
0·6%
2·21
0·23
I 1 0. 18 %
0·11% 1
MOITOCO Total 135195 10351 771·08 7·4% 43·27 I 0-42%
Algeria
Algeria
Algeria
Algeria
Ghrib
DjorfTorba
Bough7..0ul
Bouhanifa
9099
34118
28571
169
1120
55
73
1974
1948
1986
1986
1986
1986
112·6
33·6
22-3
21·4
66·7%
3·0%
40·9%
29·3%
)]1
85
21
59
J·OI
2·90
0·60
0·56
0·60%
0·26%
1·10%
0·77%
1
Algeria 5MBA 235 1978 1986 9·4 4·0% 213 1·18 0·50%
Algeria Chcurfas 1986 12 41
Algeria
Algeria
Algeria
Algeria
K'Sob
Bakkhadda
Foum EI Gharza
Beni Bahdels
1414
246
1132
31
203
43
63
1939
1974
1952
1938
1986
1986
1986
1986
13·16
8·1
20·5
6·5
42·5%
4·0%
47·7%
10-3%
198
122
433
i36
0·28
0·03
0·49
0·14
0·90%
0·01%
1·14%
0·21%
1
Algeria Oued Fodda 443 145 1974 1986 92·3 63·6% 2098 0·93 0·64%
Algeria Ighil Emda 1277 100 1954 1986 34·9 34·9% 2498 3·19 3·19% I
I
Algeria La Chaffia 873 171 1965 1986 7·3 4·3% 573 0·50 0·29%
Algeria
Algeria
Zardesas
Sarno
288
1453
111
22 1953
1986
1986
16·5
0·7
14·8%
3·2%
938
234
0·27
0·34
0·24%
J·55%
1
Algeria Foum E1 Gheiss 3 1939 1986 1·4 43·6% 187 0·03 0·93%
Algeria Hamiz 1& 2 19 1972 1986 6·65 34·8% 640 0-48 2·49%
Algeria
Algeria
Algeria
Mef\'rouch
Reservoirs on the Atlas Mountains
Reservoirs all flatter land
70 1937 1986 J·4 I
2·0% 648
1100
2800
0-03
I
0·04%
1
Algeria Tolal 2633 409 15·5% 13 !
0-49%
TaiwlIn
Taiwan
Taiwan
Taiwan
Shihmen
TenKee
WuSheh
Tsengwen
763
592
219
481
309
255
150
713
1963 2700
1900
4900
8800
2·06
1·12
j·07
4·23
0·67%
0·44%
0·72%
0·59%
8
1
Taiwan Pai ho 27 25 14300 0·38 1·51%
Taiwan Gen Shai Pei 11 7 1938 9400 0·10 1·44% 9
Taiwan
Taiwan
Taiwan
Taiwan
A Kung Ticn
Lu Liao
Tapu
Ku Kuan
32
8
104
708
36
4
9
17
16300
30
2200
600
0·52
0·00
0·23
0-42
1·43%
0·01%
2·54%
2-48%
1
Taiwan Total 2944 1525 10·14 0·67%
1
I
Indonesia Karangkates 2050 343 1977 1982 10·22 3·0% 2·04 0·60% I
Ind{lIlesia Selorejo 238 62 1970 1982 I·J6 1·9% 0·10 0·16%
Indonesia Wonogiti 1262 730 1981 1985 9·07 1·2% 2·27 I 0·31%
Sudan Total
i
5584
, 2427 1 43·5% 96·96
I 1·74% i 1
Egypt Aswan High Dam 167000 1964 1990 I 2100 1·3% 80·77 0-05%
CYPl1ls
Cyprus
Cyprus
Galini
Petra
Kalokhorio
26
37
24
0·023
0·031
0·069
I
0·00192
0·0024J
0·00628
8·2%
7·8%
9·1%
4·10%
3·90%
3·03%
1
Cyprus Lynlhroclhonda 9 0·041 0·00232 5·7% 1·14%
7,13%
1
Cyprus Lymbia 33 0·026 0·00739 28·5%
Cyprus Kophinou 12 0·004 0-00159 43·0% 14·33%
Cyprus Akrounda 26 0·022 0·00370 16·7% 5·55%
Venezuela Sanlo Domingo 427 1976 1978 0·575 19·2% I 0·29 9·58% II
" I
160
r"
.1
1
i APPENDIX I
Austria Avisio 956 2·()O ! 18~2 , 189() 100·0% 261 0·25 12·:50%
Austria P"rnegg 6150 0·50 1925 1927 0·35 70·0% 3704 0·23 46·70%
New Zealand
I Roxburgh 8826 IOH 1961 1979 26·28
~:~: 165 [4600 I 1·39%
g:~~~~ I
New Z:!\]and MutOlhina 28+1 49·7 [966 1980 2-982 1 75 0·43%
0·80%
~~: ~~::~~~ I ~~:~~~
5740 47·5 193.;1. 1959 9·5 20·0% 66
34 0·1 i 1939 1979 0·0296 21·0% 22 0·0007 0·53%
i7:
1
1577 0·34%
SQuth Africa To!:!!
161
EVACUATION OF SEDIMENTS
Kenya 1 11737 1
I
Lake Kariba 650000 116000 M.m' - no re-survey data available. but estimates = 1600 to 16000 yrs to !ill dead storage
Lake Cabora Bassa 1000000 Sedimentation may in
(inc. Kariba) I
I I I I I I I
NOles:
1. Reservoir is dis of Lake Wular
2. Catchment area not permanantly covered by snow only
3. 1·2 M.mJ of sediments dredged
4. Values calculated from hydrographic survey results
5. Approlt 50% of sediment is polluted
6. Example of the effect of deforestation
7. Typical losses from 95 reservoirs surveyed
8. dis of cascade of 3 dams & 100's of check dams - dredged at a rate of 600 000 m'Janl1 since J985
9. Flushed since 1955 at an average of 328 000 m·l/ann
10. Reservoir will not be able to function after 2010 unless rules changed
11. Reservoir Hushed in 1978 for 3 weeks clearing 620000 m3 of sediments
162
Appendix'2o
Numerical model case study
A2.1. TARBELA DAM, PAKISTAN
The Tarbela dam, a key component of the Indus basin scheme in Pakistan, was
completed in 1974 for the purpose of irrigation and hydropower. With an annual
sediment inflow into the reservoir of over 200 million tonnes, the live storage is
being rapidly depleted and liPless action is taken hydropower generation could
cease within a decade, with irrigation releases declining over the next 30 years.
The overall purpose of the feasibility study ca..rried out in 1998 by HR
vVallingford and TAMS UK, was to determine a strategy for the economical
preservation of h1e assets at Tarbela on a more sustainable basis.
A2.1.I. History
Tarbela dam was constructed in the 1970s to help regulate the seasonal flows of
the upper Indus both for irrigation of the Indus plains downstream and for the
generation of hydropower. It is still, 30 years on, the only major storage reservoir
on the Indus and, as such, plays a key role in the provision of dry season releases
of water for irrigation. Tarbela irrigation releases amount to 11 600 M.m3 , or 50%
of the WAPDA (Water and Power Development Authority of Pakistan) total, with
a corresponding agricultural revenue of Rs 2·8 billion. In addition, with an
installed capacity of 3478 MW and a finn electrical energy of 14·8 GWh/yr the
Tarbela dam provides 32% of both Pa.1dstan's total power and energy needs with
a corresponding annual revenue of Rs 6 billion. It is, therefore, a strategic
national resource whose continuing future efficient operation is of paramount
national interest.
A2.1.2. Sediment
The Tarbela dam impounds the waters of the Indus, which carry a heavy
sediment load. This is the case particularly in the spring and summer when the
rnelting snows cause heavy erosion of the uplal'1d catchment. I\t1ost of the
sediments brought down by the Indus are trapped in the Tarbela reservoir. Thus,
with an average annual sediment inflow into the reservoir of approximately
240 Mt per year, the live and dead storages of the reservoir have diminished by
16% and 21 % to 9000 1V1.m3 and 1360 M.m3 respectively in 1997.
163
EVACUATION OF SEDIMENTS
• A loss of live storage which results in a gradual reduction in the regulated yield
of the reservoir. This in tum results in a reduction in the water available for
agriculture and a reduction in the firm energy available from the project.
• The physical effect of sediment, which includes the risk of blocking the
outlets, particularly in the event of an earthquake, and erosive action of
sediment laden water on the dam's outlet works and turbines, which will result
in increasing maintenance costs to the point when the scheme will eventually
become inoperative.
Unless remedial action is taken, the reservoir will be largely filled up with
.sediment oy the year 2030, giving the project a useful life equal to that estimated
at the time of the original design. However, in view of the size of the investment
already made in the Tarbela project, and of its critical national impol1ance
outlined above, there is clearly a need for a programme of future actions to
maximize the economic returns from this resource.
A2.2. ENVIRONMENT
A2.2.1. Hydrology/climate
The source of the river Indus is situated in the Tibetan Plateau, at an elevation of
5500 metres above sea level. From there it flows across some of the highest
mountain ranges in the world before emerging onto rain-fed lower-lying country.
Downstream of Tarbela, the Indus flows along a broad valley until it reaches
Attock Gorge, some 51 km downstream. On leaving the gorge the Indus flows
onwards for a further 1600 km to its mouth on the Arabian Sea.
The Indus basin upstream of Tarbela Dam, an area of 169 650 km2, consists of
two distinct hydrological regions. Over 90% of this basin lies between the
Karakoram and Himalayan mountain ranges; the meltwaters from the snow and
ice that cover approximately one quarter of this mountainous portion of the basin
cont1ibute a major part of the annual flow reaching the Tarbela Dam. Seven of the
ten highest lnountains in the world reside within the catchment. The remainder
of the basin, about 11 700 km2, lying immediately upstream of the dam, is subject
to monsoon rainfall, primarily dllling the months of July, August and September,
the run-off from which causes sharp floods of short duration that are
superimposed on the slower responding snowmelt run-off.
Climate in the Indus basin is subtropical and semi-arid in the headwaters. It is
divided to form two distinct seasons: kharif (summer), extending from April to
September; and rabi (winter), coveling the remaining months. The annual rainfall
averages around 900 mm of which two-thirds fall between June and October.
164
:f
r
]
-l
:]
APPENDIX 2
1
t
l 14 000
I
j
12 000
IJl
~ 8000
E
iti
3:
0
E 6000
4000
2000 I-
J J J A s o N o
Month
165
EVACUATION OF SEDIMENTS
maintained. During the initial years of operation, until the year 1988, the
reservoir was drawn down close to the minimum operating level of 1300 ft every
year and, consequently, the delta advanced towards the dam. As the rate of
advance of the delta downstream is related to the extent to which the reservoir is
drawn down, the policy since this time has been to operate the reservoir with a
higher minimum water level. This, however, has encouraged the deposition of
sediment further upstream, in the middle reaches and within the live storage.
A2.2.6. Verification
The model was verified by simulating the observed sediment deposition from the
time the reservoir was impounded to 1996 and comparing the profiles predicted
by the model in 1996 with those observed. The model gave excellent predictions
in the 20 km immediately upstream of the dam, see Figure A2.2.
166
APPENDIX 2
455
435
E
c: 415
.2
ca>
Q)
m 395
"0
Q) --1974observed
co
375 _. - 1996 observed ·
- - - - 1984 mode!
••••••• 1996 model
355
335
0 10 000 20 000 30 000 40 000 50 000 60 000 70 000 80 000 90 000 100 000
Distance from dam: m
500
c::::J Silt 1 c:::J Silt 2 ~ Slit 3 ~ Silt.:;. :::::::: Siit 5 ~ Sand 1 ~ Sand 2 _ DiSChaigel120 000
450
100000
r
i'! ::~
,..,
E
c::
80 009 ~
'E
(Ii
60 000 ~
250[ ~
(3
CIl
] 200r '6
40000 ca::l
g 150 C
c:
c
c -<
-< 100
50
OL-~ __-L__~__~__L-~__~__~__~~~~__- L_ _~____~~_ _~I 0
1975 1985 1995 2005 2015 2025 2035 2045 2055
Year
Figure A2.3. Case study - sediment throughput without flushing, Tarbela reservai,;
Pakistan
167
EVACUATION OF SEDIMENTS
500 120000
c:::::J Silt 1 c:::J Silt 2 E::El Silt 3 Silt 4 c=J Silt 5 Sand 1 _ Sand 2 - Discharge
450
-c 400
100000
'"E
~
'E 350
80 000 ~
-0 'E
Qi 300
':;" a.;
C 250
OJ
60000 . ~
~
E u
C/)
'5 200 '5
OJ
C/)
(i'l
40000 (i'l
150 ::!
::! C
c C
c
« 100
«
20000
50
o~~--~--~--~~--~--~~~~--~--~~--~--~--~--Uo
1975 1985 1995 2005 2015 2025 2035 2045 2055
Year
Figure A2.4. Case study - sediment throughputs with flushing, Tarbela reservoir,
Pakistan
~ .......................... . . :. . . . :. . : i
4:·ir.···~·~.·.-;-;-:-F:.~-·~·-.~.-.o-.
~
0- - 0 - 7
u.
~
,I
6000
.....~
~ -
.() - -0
- -
0-
- 0 - -0- -0- -
4 ~
::J
-1
4000
-----~-----x------x-~~~..~--~--~---..~ ~
-&'--A-=--
3
2
2000
1
O~ __ ~ __ ~ ____ ~ __ ~ ____ ~ __ ~ __ ~ ____ ~ __ ~ __ ~ ____ ~ __ ~O
1996 2001 2006 2011 2016 2021 2026 2031 2036 2041 2046 2051 2056
II
Year ..-1
Figure A2.S. Case study - prediction oj Juture live storage, Tarbela reservoir,
Pakistan
I
j
~
168
j
J
APPENDIX 2
Figure A2.5 shows the live storage volumes attainable with the chosen flushing
regime, compared with some of the other operational policies that could be
pursued.
These results were subsequently analysed along with the costs and the benefits
associated with each scenario in order to determine the most appropriate course
of action at Tarbela.
169
Appendix 30
Flushing case studies
A3.I.I. Sedimentation
By 1958, the original live storage capacity of the reservoir had-reduced by 59%
and the intake structure and screens were aLTJ10st buried. In addition, the high-
level outlets of the upstream sediment trapping reservoir had failed, with the
result that water and sediments were being discharged through the low-level
diversion tunnel. The problem became increasingly serious by the mid 1960s,
tr-l1'eatening the continued operation of the power station.
A3.I.2. Hushing
In 1969, it was decided to attempt sediment- flushing at Mangahao reservoir
through the low-level diversion tunnel. The tunnel had not been used for 25
years, owing to problems with the gate, but had been used routinely between
1925 and 1944. For a period of 24 hours 'nothing happened, then on the second
day silt began to extrude from the tunnel and the reservoir emptied, leaving a
crater-like depression in the 13 metres of sediment which had overlain the tunnel
entrance' (Jowett, 1984).
A month of flushing, using water released from the upstream reservoir,
resulted in 0·88 M.m3 of sediment being scoured from the reservoir basin,
equating to 75% of the sediment that had accumulated since 1924. It is reported
that large logs and tree roots had to be cleared from the entrance, but that much
debris passed through the tunne1.
Since the initial flushing operation the reservoir has been emptied and flushed
annually (Jowett, 1984). The power station is closed for three weeks for the
flushing operation, reducing the annual energy produced by 4%. The annual
EVACUATION OF SEDIMENTS
flushing of the reservoir has resulted in the removal of most of the sediment from
-Mangahao Reservoir and a considerable amount from the upper reservoir.
No information is available regarding the flushing discharges used, although
the fact that -the original diversion tunnel is used suggests that the available
discharge capacity is generous.
A3.2.1. Sedimentation
From the time of construction, about 66% of the total catchment fell within the
catchment of the Pathfinder dam, reducing sediment inflows to Guernsey
reservoir substantially. After the constluction of Glendo dam in 1957, only 40/0 of
the catchment was expected to be contributing significant sedilnent inflows to
Guernsey reservoir. Until 1957, the reservoir was subject to a high sedimentation
rate, losing 39% of its oliginal capacity over a period of 30 years, with deposits
cOlnprising 170/0 sand, 61 % silt and 220/0 clay. The maximuln depth of deposit
repol1ed was about 12 m.
A3.2.2. Flushing
Partial draw down (by 12-13 m) was carried out at Guernsey reservoir annually
between 1959 and 1962 and data were collected to determine inflow and outflow
rates and sediment movement within the reservoir. No definitive details are
available in the references concelning the ·bottom outlet or other flushing
facilities, although Morris and Fan (1997) suggest that the 'overflow spillway'
was used, which is possible if there are large spillway gates. The elevations of the
power intake and the sediment deposits in the vicinity of the dam are about 10 In
above the original bed and the aInount of draw down would be consistent with an
outlet at about this elevation. Flushing discharges were typically in the range of
120-140 In3/s, cOlTesponding to about double the nlean annual inflow.
172
J
APPENDIX 3
Although sediment was scoured from the upper portions of the reservoir
during the four years of drawdowil, n10st of this was apparently redeposited in
the lower part of the basin nearer to the dam, and the suspended solids
concentration in the water discharge from the reservoir never exceeded 0·8 gil.
From the inflow and outflow data during the period 1957-62, it was estimated
that only 144 000 m 3 of accumulated sediment was relTIoved from the reservoir
basin, the equivalent of less than 0·2% of the original capacity. The long time
before first flushing was probably a factor in reducing the erosion of the deposits
(Atkinson, 1996).
It was concluded (Jarecki and Murphy, 1963) that, with future annual
drawdowns following a similar pattern, only about 2% of the original capacity
could eventually be recovered. As sediment inflows had been severely reduced by
t.1}e construction of the upstream dams, this appeared to represent a satisfactory
state of affairs. .
A3.3. I. Sedimentation
During the first two years of operation, the storage capacity reduced by 22% a
year and during the following eight years a further 32% of the capacity was lost.
Only 4% more was losfduring the next 18 years (1937-54), suggesting that an
equilibrium had been reached.
A3.3.2. Flushing
No details of the flushing facilities are given, but they are sufficient to pass over
double the Inean annual flow when the reservoir is emptied.
Plior to 1939, the reservoir apparently operated with a lilnited annual
draw down of 2·3 ill, but this was not effective. Between 1939 and 1966, it is
reported that 38 flushing operations (between one and four per year) were
undertaken with full drawdown. A wide range of results were obtained, due to
influences such as the duration of flushing and the amount of accumulation
173
EVACUATION OF SEDIMENTS 1 1
l
between flushing operations (Morris and Fan, 1997). Peak concentrations were
reported in only two events, but were very high at 270 and 370 gil.
Each flushing event comprised two stages: partial drawdown, as the reservoir
JI
j
was being emptied, followed by total drawdown, with flows along the bottom of !
the scoured channel through the basin. The duration of flushing varied from 8·5 ~
to 65·5 hours, with a mean of 18·5 hours, and was canied out mainly in the
i
month of April, Mayor November.
The volumes scoured each year ranged between about 0·5 and 2 M.m3, with an J
average of approximately I M.m3 • This is substantially less than the reported
mean annual sediment inflow, suggesting that most of the annual sediment load
passes through during routine operation.
The data presented by UNESCO (1985) suggests a reduction in the total
volume of sediment contained in the reservoir since full draw down flushing
began, although there is some ambiguity in the plotted data. J
General conclusions drawn (UNESCO, 1985) from the flushing operation at
the ZelTIo-Afchar reservoir were:
• an optimum flushing discharge between 400 and 500 m3/s produces the most
effective evacuation of sediment (higher discharges cause a greater water depth
at the dam, reducing the effectiveness of flushing)
• during the process of sediment flushing, the most active erosion occurs in a
period of 8 to 10 hours after effective erosion starts
• when the effectiveness of flushing starts to fall, it can be restored by
temporarily raising the water level for a short period.
J
A3.3.3. Downstream impacts - I
No information available. J
A3.4.2. Flushing
L., 1955 a 1·5 m diameter flushing tunnel was built through the base of the darn,
as a result of the sedimentation problems, and annual flushing commenced.
Flushing is arranged by emptying the reservoir between May and July and
allowing free flow through the reservoir. This suits the water'demands, as sugar
mills do not use water between May and October.
Between 1955 and 1980, the sediment volume contained in the Jensanpei
reservoir remained almost constant, showing that the adopted flushing regime
was highly effective, albeit retaining only about 45% of the enlarged capacity.
A3.5. f. Sedimentation
The heavily silt-laden Liuhe River has an annual average sediment concentration
of 77 gil, so that detention floods resulted in the deposition of sedilnent
deposits on the floodplain wit.~in the basin. High volumes of deposition were
reported during floods in 1949 and 1963. Table A3.1below gives some data for
flood peaks which occurred over a few successive days in 1963.
A3.5.2. Fiushinoo
The flushing wpich occurs in a reservoir of this sort is essentially uncontrolled.
Although there is some scope for control since the installation of gates on the
outlets, no information is available to judge if their use has had any on
sedimentation in this case.
UNESCO (1985) shows the reported variation in available storage capacity in
N aodehai reservoir, reducinQ: from 168 M.m3 1942 to a minimum of 97 M.m3
3
in 1950 and varying up to about 134l\1.m in 1972. Overall, the storage loss has
ranged between 20% and 42% and it appears to be dominated by massive
deposition in the largest floods, followed by a period of progressive erosion.
175
J
EVACUATION OF SEDIMENTS
J
Table A3.1.
Date
Quantity of deposits during flood peaks in 1963, Naodehai reservoir ·
July 20-22 July 23-27 July 28-31
J
Max. water level (m) 84·57 88·62 83·08
-,
Between 1950 and 1958, for example, there was progressive erosion along the
J
valley bottom, vi11ually reaching the original 1940 thalweg. However, this did not
influence the levels of deposition over the flood plain, which continued to lise.
It is doubtful whether the uncontrolled erosion of sediment deposited in the
basin that occurs duling operation could be significantly enhanced without
mechanical intervention. j
A3.6.I. Sedimentation
vVith the bypassing of bedload, settlement of suspended-sedilnent still posed a
problem for the GmUnd reservoir. After the Durlassboden reservoir was
commissioned, the annual sediment load entering Gmund reservoir reduced from
O· 2 Mt to 0·07 Mt, comprising mainly sands and gravels. Since the 1960s, the
annual sediment load has been estimated to be equivalent to about 16% of the
reservoir's volume, illustrating the need for effective flushing measures.
Aggradation of sediment in the Gmund reservoir increased between 1948 and
1960, reaching a maximum of about 0·2 M.n13 (over 20%) in the early 1960s.
From then until 1981, as a result of annual flushing and the construction of
Durlassboden reservoir, the total sediment volume was generally less, with a
typical value of about 0·15 M.m3 •
A3.6.2. Flushing
The bottom outlet passes around the right abutment of the dam in a curved
tunnel, with an inlet elevation 28 m below the crest. (It presumably occupied the
diversion tunnel used during construction.) A second outlet was added in the
middle of the dam during the reinforcement of the dam in 1963, with an inlet
elevation 27 m below the crest.
During the first period of flushing, from 1948 to 1960, flushing was not
executed every year, but was carried out depending on the amount of sediments
accuTIlulated. However, from 1960 flushing was carried out every year. Initially,
flushing proved to be difficult due to the low flow from the GerIos stream which
was the only flow available. There was also a problem with the positioning of the
entrance to the bottom outlet, apparently some distance upstream of the dam. In
most instances flushing was carried out for a week. The efficacy of flush1."1g
improved from 1964, after t."1e addition of the second outlet.
From 1967 and the beginning of the operation of Durlassboden reservoir, tIle
period required for flushing reduced to a one day. Flushing efficiency was also
increased by the increased flow, by using the turbine water released from
Durlassboden reservoir as well as the natural flow from the Gerios stream.
The flushing operation for the reservoir since 1967 is as follows:
~ flushing ·is then performed through the night with the natural flow from the
GerIos stream
III the following day flushing continues with the turbine water released from
177
EVACUATION OF SEDIMENTS
-r
J
-- f
The construction of Durlassboden reservoir upstream of Gmund reservoir has
-_been a key factor, reducing the incoming sediment load and aiding the flushing J
process. The idea of separating the bedload from the suspended load has not
proved successful in the long run (Rienossl and Schnelle, 1982), due to the
-,
maintenance costs resulting from wear on the long bypass tunnel.
No information available. f
~,
i
impounded in the reservoir by a 72 m high concrete arch dam, which also forms
the free overflow spillway. The impounded depth at the dam is 55 m and the I
length of the reservoir is 2·6 km, giving the reservoir an original storage capacity J
of 5·5 M.m3 , which is less than 3% of the average annual run-off.
I
I
A 1760 m long diversion tunnel, with a discharge capacity of 225 m3/s, was
constructed in 1974, from upstream of the basin to downstream of the dam, with
the _main purpose ot allowing sediment-laden flows to be bypassed.
J.
The materials deposited during the flood ranged from silt, sand and gravel to
1·4% wood.
In the period 1953-78 a total of 3·79 M.m3 had been deposited, representing
69% of the original volume and an average annual accretion rate of 0·15 M.m 3,
or almost 3 % of the capacity per annum.
A3.7.2. Flushing
The flushing of the alluvium following the 1978 flood was accomplished in two
different phases and required a period of four and a half mont.1.s, commencing in
mid-November 1978 and finishing by the end of March 1979. Phase '1 was
accomplished by flushing th.rough the upper of the outlets, about 44 m below the
spillway crest, with low flow of approximately 300 lis, coming from a lateral
valley. phase ran until the end of December and evacuated about 0·3 M.m3
of materiaL
Phase 2 started with the opening of the lower bottom outlet, at the base of the
impoundment, at the beginning of January. The water used for flushing was
successively increased from 1 to 1·5 m3/s by using part of the river run-off.
During this three-month phase it was estimated that approximately 2·1 M.m3 of
material was flushed from the reservoir. Flushing was assisted by the use of
bulldozers and shovels to remove the wood buried in the alluvium and to push the
fflaterial into the eroded channels.
During the fiusping operation it is understood that L1}e balance· of the inflows
was passed through the sediment diversion tunnel, which bypasses the reservoir
basin.
No information is available on subsequent flushing.
179
EVACUATION OF SEDIMENTS
arm about 30 km to its confluence, with the Yongding about 5 km upstream of the
dam. The original surface area at the flood storage level was 229 kln2 •
The catchment area is 43 400 km2, which has a semi-arid continental climate,
with dry and very cold winters and 750/0 of the annual rainfall occurring between
June and September. The average annual rainfall (1951-84) is 420 mm, with a
range of 278-545 mm. The mean annual run-off in the period 1925-85 was
1250 M.m3 (Binnie and Partners, 1986), although it was also noted that there had
been a progressive decrease in flow between the 1950s and the 1970s, probably
due to the construction of about 300 reservoirs in the catchment and the increased
use of water for irrigation. Only about 2% of the total catchment drains to the
Guishui arm of the reservoir.
The original storage capacity was 2270 M.m3, which is about 80% greater than
mean annual run-off. This original storage capacity comprised 600 M.m3 of dead
storage, 660 M.m3 of 'benefit' storage (for river regulation and hydropower) and
1010 M.m3 of flood storage. Of the total original storage, about 60% lay in the
Guishui arm.
The dam has a gated spillway, which was under reconstruction in 1986 (Binnie
and Partners, 1986), to give a discharge capacity of 2950 n13/s.
There is an 8 m diameter bottom outlet tunnel, through which flows are
controlled by two sets of four sluice gates, with invert levels 27 In and 39 In
below the flood storage level. The maximum discharge capacity is about
560 m3/s, which is about 14 times greater than the mean run-off from the
catchment.
A3.8.1. Sedimentation
About 40% of the total catchment is classified as loess areas, with friable soils
and sparse vegetation cover.
There has been a wide range in annual sediment loads in the rivers entering the
reservoir, with a nlaximum of 132 Mt and minimum of 1·3 Mt in the period
1951-84 (Binnie and Partners, 1986). As shown in Table A3.2, there was a
1971-80 11 96 85 588
180
APPENDIX 3
181
EVACUATION OF SEDIMENTS
accretion raising bed levels. Proposals were made for a two-phase approach
comprising:
(a) divelting sediment-laden flows from the Yongding arm of the basin into the
Guishui arm, to store sediment in its dead storage zone for a period of about
eight years
(b) subsequently raising the flood storage elevation and developing a system of
polders to contain the bulk of the sediment inflows at the upstream end of the
Yongding arm.
514 M.m3 • 1,
A3.9.1. Sedimentation
The mean annual sediment inflow was estimated at 0·63 Mt (UNESCO, 1985).
With no bottom outlet for sediment flushing available, sediment deposition was
severe, amounting to,85% of the original storage capacity of the reservoir by
1981. The remaining 1-4 M.m3 of volume was insufficient for flow regulation,
which required 3·6 M.m3 ,
182
APPENDIX 3
The general procedure is that, after drawing down the reservoir, about 50 m 3/s
is released from the upstrealll reservoir fOl~ flushing. The duration of effective
flushing is generally about one day, during which time typically about 0·2 M.IU3
of sediment can be removed by between 1 and 3% of the annual flovv. This is only
about one third of the average annual sediment inflow. It was noted that, in the
vicinity of the dam, the amount of erosion was about 2 m, but 4 krn upstream
there was none.
The quantity of sediment flushed is reported to be limited by a combination of
factors, including consolidation of the fine silts and the deposition of bedload in
the upper part of the reservoir basin, but an inlportant factor must be the
relatively high elevation of the spillway through which the flushed discharge
must pass.
IWHR's (1983) characterisation of the flushing in Shuicaozi reservoir can be
interpreted as follows.
In the light of these observations and theoretical studies into the behaviour of
the sediments, nVHR (1983) proposed the following refinements to the flushing
procedure.
183
J
EVACUATION OF SEDIf'lIENTS
}
A3.IO. HEISONGLIN RESERVOIR (China, 1959)
I
Heisonglin is a small reservoir, with an initial storage capacity of 8·6 M.m3 , J
which occupies a valley with an average slope of 1% located in a hilly region on
the upstream_reach of the Yeyu River, a tributary of the Yellow River. The 45 m
high earth dam was constructed in 1959, for the purpose of impounding water for
irrigation and flood protection. The average annual inflow to the reservoir is
14.2 M.m3 , which is 65% greater than the original basin capacity. i
. Following severe sedimentation problelns at numerous reservoirs built in J
China in the 1950s, a decision was made in 1961 to use Heisonglin reservoir as
an experimental site, to study sedimentation behaviour in detail and to develop I
appropriate sediment management techniques (MOITis and Fan, 1997). J
!I
A3.! 0.1. Sedimentation ,1
The mean annual sediment load is reported as 0·71 Mt, and this has the potential
to cause accretion of the order of 8% per year if it is all trapped in the reservoir.
The annual amount is both irregular and seasonal, with 87% of the mean annual J
sediment load entering the reservoir during July and August, which represents
about the first half of the flood season. In terms of discharge, however, these two ,
!
months account for only about 25% of the annual inflow. (This pattern is pretty J
typical of most reservoirs with the potential for significant sedimentation
problems.) The sediment is lnainly silt and derives largely from high rates of II
gully erosion in the catchment. J
For the first three years of operation, up to June 1962, the reservoir was
operated purely as an impounding reservoir, with no flushing, resulting in serious I
siltation of 1·62 M.m3 , representing an average rate of 6% per annum. J
J
APPENDIX 3
Drawdown flushing
The lowering of the reservoir pool in July initiates the erosional processes
associated with flushing. A large and highly concentrated discharge of sediment
occurs at the transition from drawdown to flushing. Smaller amounts of sediment
are removed thereafter, by base flow and by the smaller inflow events that can be
released through the bottom outlet. Due to the operational objectives and
constraints at Heisonglin, in particular the supply of irrigation water at
mal1ageable rates and the limited discharge capacity of the bottom outlet,
flushing ,"vith large flows to widen the. main channel cannot be undertaken.
Detention flushing
vVhen floods entering the reservoir storage during the drawdown period exceed
the release rate to irrigators, the pool retains the silt-laden water for petiods of
hours or days. As the silts settle slowly, if the flow can be discharged within a
couple of days, approximately 70% of the sediment can normally be evacuated
with the water releases. In addition, as the reservoir is emptied at the end of each
flood and riverine flow is again established, deposits of silt are scoured in t..l-J.e
same manner as for the initial draw down ..
Lateral erosion
Attempts were made to create a longitudinal channel running parallel to the main
flushing channel for hydraulic scouring of the floodplain deposits. Accidental
overflow from this channel initiated lateral erosion, which formed a gully across
the deposits. From this, it was recognised that erosion of the deposits would
proceed much faster by directing the flow along the high lateral slopes from the
sides of the reservoir towards the main channel. Because of the high gradient that
can be achieved by lateral drainage, even small discharges were highly effective
in eroding deposits of non-cohesive silt. This technique was applied for a total of
6·8 months between 1980 and 1985. Using a flow of only 0·2 m 3/s, 816000 m 3
of deposits were eroded into the main channel. This equated to a remarkably
higher sediment/water ratio of 0·23.
185
j,
EVACUATION OF SEDIMENTS
1,
the irrigation canals, are designed to accept very high concentrations of sediment,
in order to enhance soil fertility. ,
Environmental penalties are clearly either non-existent or very minor.
1
A3.11. SANMENXIA RESERVOIR (China, 1960)
Sanmenxia dam was completed in September 1960 and was the first to be built
1
on the middle reaches of the silt-laden Yellow River. The 96 m high concrete
gravity danl, which controls a drainage area of almost 700 000 km2, was planned
as a multiple-use project for flood control, hydropower, inigation, navigation and
ice jam control. The maximum historic flood at the site was 36000 m 3/s in 1843
(Morris and Fan, 1997).
It was designed originally with a full reservoir level of 360 m, giving
1
65 000 M.m3 of storage capacity, inundating 3500 kn12 of floodplain and
requiring the relocation of 870 000 people. To reduce these impacts during the
first stage of construction, the dam was built to an elevation of 350 m with a
1
maximum operating level of 340 m and an original storage capacity of
9640 M.m3 • Because of the high sediment loads in the Yellow River, the original
plans included two sediment control measures:
1
• the reservoir was to impound water continuously, but release 350/0 of the I
,..l
sediment inflow as turbidity currents, through 12 outlets at an elevation of
300 m (40 m below top operating level) "
t
• total sediment inflow was to be reduced by 3% annually by soil conservation
works in the catchment, resulting in a 60% reduction over 20 years. 1
, These figures proved to be extremely optimistic (Monis and Fan, 1997).
J
, -'
A3.II.!. Sedimentation 1
The estimated sediment load prior to construction of the dam was 1600 Mt per
annum, with a median diameter of 0·03 mm, representing a mean concentration
of 38 gIl. Extreme concentrations of up to 940 gil had been measured in the
1
Yellow River in the vicinity of the dam site. About 60% of the annual sediment
load occuned with about 300/0 of the annual run-off in July and August.
Iffilnediately after impounding began, severe sediment problems became 1
evident. During the first 18 months of operation, 1800 Mt of sediment had
accumulated in the reservoir, representing 93 % trapping and the loss of about
20% of the storage capacity. In the next four years, 3400 Mt was deposited and
the total loss of storage reached 3700 M.m3 , or 400/0 of the original capacity
(Monis and Fan, 1997). This rate of accretion threatened to elilninate all the
project benefits, in addition to sedinlent deposits which were raising the bed
1
elevation and flood levels in the Yellow River as fat as 260 kin upstream of the
daln.
186
1
1
APPENDIX 3
Stage 1 (1960-62)
Impounding of the reservoir began in September 1960 and seli.ous deposition
was occurring by the time the water level reached 335·5 m (4·5 m below the
planned impounding level). Sediment accumulation raised bed levels by 4·5 m
. near the upstream end of the reservoir and caused backwater effects 250 kID
upstream of dam.
This was endangering agricultural land in the floodplain of the Wei River,
which joins the Yellow River near the upstream end of the impoundment, and
industrial developments in Xian City.
Total deposition up to the start of the 1962 flood season was estimated as
1800 Mt.
Stage 2 (1962-66)
The reservoir operation was changed from April 1962 to maintain a lower water
level throughout the year, by using the 12 outlets at an elevation of 300 m.
However, the outlet capacity proved insufficient and water levels during periods
of high discharge were too high for efficient sediment release during the flood
season, resulting in the trapping of a further 3400 l'vlt over four flood seasons.
It was also noted that sediment trapping in the reservoir was having an effect
downstream of the dam, where degradation was occurring.
It became clear that additional sluicing capacity would be required.
187
EVACUATION OF SEDIMENTS
[
Stage 3 (1966-70)
Prior to the start of the 1966 flood season, additional sluicing capacity was
[
provided by the excavation of two 11 m diameter bypass tunnels, with an invert
elevation of 290 m and controlled by 8 x 8 m radial gates, around the · left
abutment of the dam. Four of the eight power intakes were convelted to sediment
[
sluices and the pool level was lowered duling the flood season. This had the
effect of reducing the trap efficiency to 17·5%, but did not lower the bed
elevation at the upstream end of the reservoir basin, and deposition in the Wei
[
River was still a problem. Consequently, it was decided that still more low-level
sluicing capacity would be needed. [
Stage 4 (1970-73) r
In time for the start of the 1970 flood season, 8 of the original 12 river diversion L
outlets, filled with concrete, were reopened at an elevation of 280 m. Reservoir
operation was changed to flood detention and sediment sluicing, with all outlets
constantly open. Sediment release efficiency reached 1050/0, representing a year- [
on-year reduction in the amount of accumulated sedin1ent in the reservoir basin.
The bed elevation at the upstream end of the basin fell by nearly 2 m. In 1973
five generating sets of 50 MW each were installed, replacing the original eight L
r
125 MW sets.
r
Stage 5 (1973-78) i
Once an overall sediment balance had been achieved and bed levels at the
upstream end of the reservoir were under control, it was decided that reservoir
operation could be modified after the 1973 flood season, to increase the project i
benefits.
This regin1e, which is understood to have continued to the present day,
provides water for irrigation, hydropower, and ice jam control during the non- ·
flood months. At the start of the flood season, in July, all the outlets are opened
I
and the high-capacity bottom outlets allow a low pool level to be maintained. The
high discharges can-ying the .sediment load also prevent excessive deposition in
the Yellow River downstream of the dam.
I
I
188
L
r
I
APPENDIX 3
A3./2.1. Sedimentation
During the period 1961-70, the average measured suspended-sediment concen-
tration was 727 mg/l, giving an average sediment inflow of 15·3 Mt, suggesting
a maximum accretion potential of the order of 80/0 per annum. The particle size
distribution of measured suspended load comprised 12% sand, 60% silt and 28%
clay (Mahmood, 1987). In addition, the Kabul River carries a bedload of gravel
and cobbles, which were not included in the rneasuredconcentrations.
After the first year's operation, 30 M.m3 of sediment had deposited in the
reservoir, increasing to 70 M.m3 five years. By 1980, after 20 years of
operation, reservoir had completely silted to the conservation pool elevation,
except for a 60 m wide by 6 rn deep channel on right bank, where the po\ver
and ilTigation intakes are located. The reservoir deposits had an accumulation of
cobbles and boulders on the surface and, in 1983, NIahmood observed
189
EVACUATION OFSEDIMENTS
gravels and cobbles up to 75 nun being passed from the reservoir with the
irrigation supplies.
A3.12.2. Flushing
Five flushing operations were performed during the period 1976-79. The
flushing was carried out by lowering the water level to the spillway crest level.
The total duration of flushing was about 20 days and these operations removed
an estimated 4·2 M.m3 of sedimel1t deposits from the reservoir, amounting to
about 6% of the probable sediment inflow over the same period.
No information is available on any later attempts at flushing, but it appears that
the reservoir has essentially reached an equilibrium condition with virtually no J
residual live storage capacity, and that it will not be practicable to increase the
live storage unless deeper high-capacity outlets are provided at the dam.
J
A3.I3. J. Sedimentation
The annual sedinlent discharge into the reservoir is reported as between 12 and
14 Mt. The volume of deposited sediment reached about 30 M.m3 by 1968 and
was reasonably stable at 50-55% of the original storage capacity up to 1970,
after which no further data are available.
190
J
J
APPENDIX 3
A3.13.2. Flushing
The dam has eight bottom outlets, 35 m below impounding level and 21 m below
the elevation of the power intake. These are reported to have a discharge capacity
of about 350 m 3/s at maximum impounding lev~l. From other data available, it
appears that this is the discharge capacity for each, giving a maximum discharge
capacity of 2800 m 3/s when the reservoir is full. This is much larger than the
mean inflow of about 500 m3/s.
Since 1963, drawdown flushing of the reservoir has been operated, which was
achieved by lowering the water level by 4-5 m during the May to August flood
season. The fact that the lowering is so modest suggests that it may have
depended on the use of a gated spillway, as well as the bottom outlets. The
available data are not entirely consistent, as a plot of sediment concentrations
entering and leaving the reservoir during the 1964 flood season suggests that
there would be net accretion, rather than an approximate equilibrium.
A3.14.1. Sedimentation
The vast majority of the annual sediment inflows occur with high discharges in
the months of March to June.
Sedimentation was a serious problem in the first 17 years of operation and
caused an average storage loss of 36·5 M.m3 per annum, equivalent to an annual
rate of 2·1 %. The trap efficiency during this period was estimated as 730/0, with
19/
EVACUATION OF SEDIMENTS
A3./4.2. Flushing
The reservoir was built with three bottom outlets on the right-hand side (total
discharge capacity 430 m3/s, elevation 191·3 m) and two bOttOlTI outlets on the
left-hand side (550 m 3/s, 193·8 m). These are near the bottom of the reservoir,
close to the original river-bed level and their total discharge capacity of 980 m3/s
compares well with the mean annual flow of 160 m3/s. (With the reservoir level
drawn down to 25%, the bottom outlet discharge capacity would be reduced by
about a factor of two, so would still be three times the nlean annual discharge.)
Because of the rapid and continuing reduction in storage capacity, a decision
was made in 1980 that the operating regime should be changed to incorporate
more pro-active sediment removal. A number of alternatives were considered,
before selecting the option of annually elnptying and flushing the reservoir.
The peak annual inflows to Sefid-Rud reservoir occur in the months of March
to June and the inigation period is from May to September. Accordingly, the
flushing programrne was designed to occur from October to February, virtually
emptying the reservoir down to an elevation of 197 m, then allowing the reservoir
to fill in time for the start of the inigation season. (During the first two years of
flushing, the reservoir was not completely drawn down because of fears that
unstable sediment would block the bottom outlets.) It would have been desirable
to flush the reservoir for a further period, to take advantage of the higher flows
that would further SCOUT the deposits and widen the main channel, but the
uncertainty of filling the reservoir for the irrigation period dictates the closure of
the outlets in February. Furthennore, the bottom outlets may not have sufficient
discharge capacity to allow this to be done.
The consequence of this operating regime is that the majority of the annual
sediment load enters the reservoir when it is either filling or full, so is likely to
be deposited until it can be subjected to erosion during the subsequent flushing
period.
The initial years of flushing provided very high sediment outflows and a rapid
recovery of storage capacity up to about 75% of the original in 1992. This was
expected to be approximately sustained with a continuing flushing regime, with
the possibility of an increase to about 900/0 with supplementary measures (as
described later).
The gross benefits of the flushing operations, which comprise the volume of
sediment removed plus the vollune of deposition averted, amounted to 320 M.ln3
over the first 10 years of flushing, equivalent to 1·9% of the storage capacity per
annum. The average suspended-sedinlent content in the flushing flow was 48 gil,
with a peak of up to 670 gIl. Selected statistics on the annual flushing period up
to 1990 are summarised in Table A3.3.
92
: •...
; '.r-
·. . r
.:~
APPENDIX 3
1980-81 61 0 536 10 24
1981-82 6S 0 390 11 12
I
1982-83 117 10 1513 26 52
1983-84 16 80 795 23 68
1986-87
I 17 85
!
942 I 26 27
1987-88 24 86 ]812 22 57
I
1988-89 9 113 1057 31 54
I
1989-90 5 103 681 22 32
I
Total 351 744 10667 21 514
I
It was observed that sediInents were eroded during draw down and flushing by
three processes:
• sheet erosion
• channel erosion
• bank failures.
Sheet erosion was the most important type of erosion during the first
draw down operation, comprising sheet flow and scour of recentiy deposited fine
sediments in the lower reaches of the reservoir. After the first couple of years of
drawdown flushing, the amount of sediment removed by this process was less
important, removing only a small part of the deposition which had occurred onto
the submerged floodplain in the previous period when the reservoir was filling or
full.
Channel erosion was the most important process from the third drawdown
period onwards, when the first full draw down was undertaken. The channel
banks would be near-vertical initially, but would then fail. The rate of sediment
removal was found to be sensitive to changes in discharge, which would trigger
off accelerated erosion.
It \vas concluded that the long-term storage recovery would be limited by the
narrow width of the Inain channel in relation to the overall width of the reservoir
basin. Furthermore, sediment would continue to be deposited on the submerged
floodplains during impounding periods, suggesting that, after the initial period of
recovery, the storage capacity would begin to decrease again in the longer term.
193
EVACUATION OF SEDIMENTS
Two novel methods were investigated to promote the removal of deposits from
the floodplains and to prevent a progressive long-term loss of storage, as
described below.
Diversion channel
Longitudinal erosion is achieved by constructing a pilot channel parallel to the
main channel. The channel is fed by water, either from a tributary or by diverting
flow from the main river using a temporary diversion dam. This concept was
initially tested in the 1987-88 flushing season, in which a 5 km diversion channel
was formed and flows of between 1·2 and 2·2 m 3Is were diverted from a small
tributary. A substantial increase in the outflow sediment concentration was
achieved, so, following this success, a longer channel was built in the following
flushing season, along the smaller of the two main river valleys forming the
reservoir basin. The sequence of operation was: .
• construct a pilot channel defining the route and connected with the stream at
its upstream end
• form an earth dam to divert the streaInflow to the. diversion channel 1
-. ~
i
• divert water to the pilot channel at a rate which is high enough to avoid
overtopping of the diversion dam, but low enough to avoid the pilot channel
overtopping and sholt-circuiting back to the main channel previously formed
I
by drawdown flushing alone. I
.--l
The pilot channel was 7·6 km long, with an average slope of about 11200, ~d
that generally followed the edge of the floodplain deposits, to allow access by
earthmoving plant. Initially, starting in mid-January 1989, a flow of 1 m 3/s was
passed down the pilot channel, which was increased progressively in line with
the erosion of the channel to the full 12 m 3/s flow in the tributary. Retrogressive
erosion was the principal means of channel development.
Erosion continued until the channel was submerged by rising water levels ._- /
during ·i mpounding in February, then resumed when the diversion channel was !
:--I
£. -
APPENDIX 3
re-exposed for the next flushing season. By December 1989, after a total" of 95
days of operation, the diversion channel had reached an essentially stable
condition. The eventual channel top width ranged between 50 m and 200 m, but
no information is given on the depth or total volUlne eroded.
Long-term predictions
Studies by Tolouie (1993) estimated that, by creating a new diversion channel
each year and by deploying 75% of the total flushing season inflow, it would be
_I
possible to recover lost storage and maintain a long-telID storage capacity of
about 90%, compared with about 75% by flushing alone.
A3.15.1. Sedimentation
The capacity of the reservoir was seriously depleted by an average annual
sediment inflow of about 84 Mt (UNESCO, 1985). Morris and Fan (1997) show
a photograph of a water supply intake in the delta upstream of the dam,
apparently completely surrounded by sediment, but no further details are
available.
A3.15.2. Flushing
Little infom1ation is available on the flushing operations. Table A3.4 lists some
data for flushing operations carried out in July 1971 and July 1973.
195
EVACUATION OF SEDIMENTS
A3.16.1. Sedimentation
From 1966 to 1973, the first eight years of the reservoir's operation, 3·19 M.m3
of sediment had deposited in the reservoir, representing 24% of the original
storage, with the height of the deposits behind the dam reaching 27 m. Deposits
near the dam were described as fine, with a Dso of 0·02 mm, becoming coarser
at a distance of 350 m to 800 m from the dam.
A. 3. 16.2. Flushing
The dam has a small outlet, 2·6 m above the base of the danl, with a discharge
capacity (at full impounding level) of 17 m 3/s, but there is also an outlet for flood
196
APPENDIX 3
discharge 14·5 m above the river bed, capable of passing a maximum discharge
of 1260 m 3/s.
Flushing was first can-ied out in July 1974, when the reservoir was emptied
and flushed for 37 days. During this flushing period, 0·8 M.m 3 of sediInent was
I
removed from the reservoir. The reservoir was then impounded for five years to
Ij
June 1979, before flushing for the second time for a period of 52 days during the
i
i
I
flood season. The second flushing period removed 1·03 M.m3 of sedilnent,
reducing the volume of sediment in the reservoir to 2·62 M.m3 (20% of the
I original storage capacity). Emptying and flushing were subsequently undertaken
i in 1982 and 1986.
Ij
I
During emptying and flushing, it was reported that strong retrogressive erosion
occurred as a result of lowering the water level. A channel was rapidly fonned in
the floodplain deposits regressing upstream and deepening continuously. In the
first 350 m from the datu, deposits on the floodplain collapsed and slid into the
main channel. In the upstream reaches, where the sediments were coarser, the
cross-section eroded was initially rectangular in form and was followed by the
collapse of the floodplain deposits into the main channel. Outflow concentrations
were reported to reach about 1000 gIl, irrespective of the flushing discharge
(UNESCO, 1985).
Experience at Hengshan reservoir suggests that flushing every few years is
sufficient in this case, which is probably aided significantly by the high gradient
of u~e original stream bed and the steepness of the valley sides. The efficiency of
the flushing was high when the main channel, which had been eroded in the
previous flushing, had been silted up by deposited sediments during a period of
several years. It was thought that greater recovery of storage capacity could be
achieved if the reservoir was to be emptied prior to the start of the flood.
197
EVACUATION OF SEDiMENTS
J
,J
A3./7./. Sedimentation
The average annual sediment inflow is 0·81 Mt, which would have a deposited
J
volume of the order of rather over 1% of the original storage volume. The annual
load is estimated to be distributed as follows: J
18 % throughflow from normal hydropower and gate operations;
21 % deposited on telTaces;
7% bedload trapped in reservoir; and
54% thalweg deposits, removed by flushing.
A nalTOW section of the reservoir 4 km upstream of the dam divides the basin
into upper and. lower parts. The upper basin is being progressively filled with
sand and coarse matelial, which is generally not removed by flushing, whereas
'the lower basin consists of a deep river channellnaintained by flushing, between
a series of relatively flat river terraces, onto which fine sediment is deposited.
For the first seven years it was apparently operated without flushing, with the
reservoir trapping 82% of the incoming sediment. Part of the suspended load was
transported by turbidity CUlTents to the area of the dam and, after several years,
were starting to interfere with hydropower production.
A3./7.2. Flushing
The dam has a single bottom outlet located near the thalweg of the original river
channel and immediately adjacent to the intake screen, a location that facilitates
flushing of sediment from in front of the intake.
Flushing operations at eachi reservoir have been well documented and have
been considered successful in preserving the storage capacity of the reseIVoir
(MolTis and Fan, 1997).
The first flushing operation was canied out in October 1973, to flush
sediments that had accumulated near the power intake. Owing to the success of
this operation, it was decided to carry out flushing every year during the wet
. season. During the 18 years from 1973 to 1990 the reseIVoir was flushed 14
times. Flushing was cmTied out in three stages:
• Slow drawdown: the reseIVoir level was lowered from 990 m (full impounding
level) to 965 m at a rate of 1 mlday, with the turbines operating at full capacity
and supplemented by opening the spillway gates and the bottom outlet as
necessary
• Rapid drawdown: the turbines were stopped and the bottom outlet opened to
evacuate the remaining water from the reservoir, which typically took between
5 and 10 hours
• Free flow: this typically lasted 2 to 3 days and occulTed once the reseIVoir was
empty and the river was flowing freely along the original river channel.
At the end of the flushing operation, the outlet was closed up and the reser-
voir allowed to refill, typically taking between 16 and 21 days. The amount of
198
APPENDIX 3
sediment released during each stage varied considerably from one event to
another, reflecting variations in the rates of sediment inflow and different
intervals between flushing operations, as illustrated in Table A3.5.
Little erosion of the sandy or gravely material in the upper part of the reservoir
basin was observed during flushing operations. In the lower basin, minor gullies
developed across the terraces during the slow draw down period and on the
terrace slopes there was a tendency for deposited sediments to be eroded by wave
action. However, there was no general erosion of sediment from the surface of
the terraces.
The zone of ma"'{imum erosion was along the main channel, which is also
where most of the incoming sediment was deposited. The slow draw down
exposed channel sediments to scouring action, and finer sediments were
transported nearer the dam.
The erosion and release of sediment during the rapid draw down phase was
reported to be spectacular (Monis and Fan, 1997). During the last few metres of
rapid draw down hyper-concentrated flows were observed. A major part of the
Oct. 1973 - - - -
Aug. 1974 186200 I! 225200 - 411400
Oct. 1975 - - -
Oct. 1977 - 40700 44000 84700
June 1985 - - -
I
-
July 1987 - - -
Sept. 1988 61600 627 000 577100 1265700
Oct. 1990
r---moo 278 700
199
EVACUATION OF SEDIMENTS
total sediment flushed on each occasion occun-ed during the final few hours of
the rapid drawdown phase and the first few hours of free flow conditions.
A3.IB.I. Sedimentation
As a result of glacial activity the sediment inflow to Gebidem reservoir is very
high, with an annual average of about 0-4 M.m3 , equivalent to over 4% of the
storage capacity. This is mainly granular material, ranging from very fine sand to
gravel, of which about 20% is between 1 mnl and 100 nlm in diameter. The
sediment load is strongly correlated with the flow hydro graph during the summer
months.
A.3.IB.2. Flushing
Because of the high sediment load in relation to the reservoir capacity, sediment
managelnent was planned for in the initial design. Consideration was given to the
alternatives of sedilnent bypassing and dredging before selecting flushing as the
. most practicable and economic option. Venting of turbidity cun-ents was also
considered, but was rejected because the sediments would be too coarse for it to
be effective:
The danl was designed with two flushing tunnels located directly beneath the
. power intakes and close to the original streanl-bed level. Originally, the low-level
200
APPENDIX 3
outlets each contained two gates; a radial service gate at the downstream end, and
flap gate at the upstream end that could be closed in emergencies or for
maintenance of the service gate and outlet tunneL To resist erosion, the entire
sutface of the outlet tunnel was lined with steel plate.
After 25 years of operation, erosion of the service gate seal on the bottom
outlet had become a problem, preventing an effective watertight seal to be
maintained. In .1995-96 a third gate was added to each outlet, for use as the
discharge control during flushing operations, allowing the original service gates
to be used only fully open or fully closed, without significant wear on the
replaced seals.
The reservoir is flushed between May and July every year, for 2 or 3 days.
Owing to the gorge-type geolnetry of the impoundment, flushing has resulted in
the entire reservoir basin being kept virtually sediment free. Flushing is carried
out prior to late SUnL.T..er floods, \vhen conditions favourable to flushing occur:
I)the flow of the Massa River is low enough (less than 20 rn3/s) to allow full
drawdovvn
• the flow of the Rhone is large enough (greater than 40 rn3/s) to dilute and
transport the sediment-laden flows, but not sufficient to pose excessive flood
risks downstream
• the 0° isotherm is located around 3000 m, which corresponds to stable
meteorological conditions (presumably indicating that a summer storm would
not interfere with the flushing operation).
201
EVACUATION OF SEDIMENTS
202
APPENDIX 3
The release of sediment into the fast flowing Rhone has appeared not to have
had any significant adverse effect on river morphology, but has helped maintain
sediment loads in the face of a long history of gravel extraction. However,
temporary high suspended-sediment loads and deposition on the river bed have
caused some problems at water supply intakes, and have also been linked to fish
kills in the Rhone.
Al/9.1. Sedimentation
There was scant information available to predict likely sediment loads at the time
of design, but the adopted design suspended load of 160 000 m 3 per year (based .
on lirnited actual data) was considered to be conservative. The bedload was
estimated at half the suspended load for design purposes. Thus, there was
expected to be the potential (with 100% trap efficiency) for the loss of up to 8%
of the storage capacity per annum. ., ..
During the first four years of operation, from 1974 to 1978, the scheme
operator was required to continue generation without any interruptions for
sedii'1lent flushing. Generally, the reservoir was held at the highest level possible
at the time, as this minimised the passage of the highly abrasive sediment through
the turbines, maximised the generating head and allowed for the easy release of
floodwater over the spillway.
Surveys of the reservoir bed carried out in February 1976 and April 1978
indicated that 0·58 M.m3 of deposition had occurred over two flood seasons,
suggesting a rate of accretion of 0·29 M.m3 per annum, which was rather greater
tt~an the design estimate of 0·24 M.m3 • This was attributed, at least partly, to road
building and deforestation in the catchment. On the other hand, the estimated
accretion over the first four years amounted to about 25 % of original storage
capacity, so was a little less than might have been expected.
203
EVACUATION OF SEDIMENTS
A3.19.2. Flushing
The dam is provided with three bOttOITI outlets for flushing sediment froin the
reservoir, two in the deeper San Domingo river valley and one in the Aracay river
valley. The outlets are each equipped with a 3 x 5 rn radial gate and a 3·2 x 2·5 m
sliding gate. At normal top reservoir level (1585 m), the total discharge capacity
of the three bottom outlets is 170 m 3/s. They were also considered large enough
to deal satisfactorily with major obstructions, such as tree trunks.
Hydraulic model studies during the design stage had indicated that sediment
could be effectively flushed from the reservoir both under pressurised and free-
flowing conditions, which_ are described below.
From the first flushing operation, the following conclusions were drawn
(Krumdieck and Chan10t, 1979) specifically for the Santo Domingo reservoir, but
also having application for other small reservoirs with heavy sedin1ent loads:
204
APPENDIX 3
e flushing should take place annually, during, and preferably towards the end of
the high-flow period
'I even under low-flow conditions, hydraulic flushing can be effective
e flushing operations should begin when the sediment deposits are not less than
100-200 m from the face of the dam
• free-flow flushing is generally more effective than pressure flushing, but free-
flow flushing should be intenupted occasionally, to catTY out pressure flushing
of deposits around the entrances and exits of bottom outlets (for up to 10
minutes at a time, eroding up to 5000 m 3).
A3.20.1. Sedimentation
The mean annual suspended-sediment load inflow is given as 0·53 Mt, the vast
majority of which enters during the flood season, July to September. The annual
average covers a wide annual variation, of between 0·12 Mt and 1·34 Mt in the
period 1974-83 (Chen and Zhao, 1992).
In 1975, the second year of operation, a major flood occurred which deposited
gravel 3-4 ill thick, 1 krn upstream of the oliginal impoundment.
By the end of 1983, 53% storage capacity in the Nanqin reservoir was
reported to occupied by deposited sediment and it was estimated that the
span of the reservoir would end by the 2000 (Chen and Zhao, 1992). (From
the quoted data, this percentage loss apparently to an intermediate
205
EVACUATION OF SEDIMENTS
impoundment level of 118 m.) The Inaximum depth of deposition near the dam
was of the order of 12 m.
A3.20.Z. Flushing
A 3 m diameter tunnel, 3 m above the original river-bed level, was built into
the dam for the purpose of sediment flushing. This has a discharge capacity of
14 m 3/s when the pool level reaches the soffit, rising to 110 m 3/s at maximum
impounding level.
Drawdown flushing
Although the removal achieved by density CUITent venting was considered good,
it was realised that more effective methods would be needed to deal with bedload
deposition and to recover and preserve storage in the longer term. At the end of
the 1984 flood season, an experimental flushing operation by emptying the
reservoir was carried out. Flushing was carried out for a period of four days, in
which all the sediment deposited in the current year was flushed out, along with
0·72 M.m3 of sediment that had been deposited in earlier years. The effective
storage capacity was restored to the value that applied in 1980 (Chen and Zhao,
1992) and the maximum thickness of deposits reduced to about 6 m.
Conclusions
From the experience gained in the 1984 flushing test, the following operational
rules were drawn up for N anqin reservoir:
• the pool level should be kept high during the flood season to prevent bedload
from advancing too far downstream and armouring the more erodible
deposits
• density current venting should be practised and the level of the turbid water
reservoir kept below the floodplains at about 114 m
• drawdown flushing should be undertaken at the end of the flood season once
every 3-4 years, triggered by a storage depletion criterion.
206
APPENDIX 3
A3.21.1. Sedimentation
Suspended sediment inflows to the reservoir have been estimated by one or two
daily samples, supported by more detailed sediment concentration profiles. There
are no direct measurements of bedload entering the reservoir, but values have
apparently been inferred from outflow sediment measurements and surveys of the
reservoir basin. Between 1976 and 1984 the estimated total annual amounts of
sediment inflow have ranged between 0·49 an.d 29 M.m3, with a median value of
2·2l\1.m3, which is about 20% of the original storage (Bhargava et al., 1987).
The reservoir began impounding in March 1975 and the sediment deposited
was surveyed after one year of operation, by which time it had reached the crest
of the spillway, which is 16 m below the full reservoir level, reducing the storage
capacity by 23%. Between then and 1981, the sediment level throughout the
reservoir basin rose progressively, reaching a total storage loss of 60%. Table
A3.8 summarises some key data regarding the sedimentation of Ichari reservoir
(Bhargava et al., 1987). It is notable that over 90% of the very high sediment load
in 1978-79 was passed downstream.
In an inspection of the roller bucket of the gated spillway in 1984, severe
damage of the teeth was found, including exposed concrete surfaces (which'Ii'ad
been eroded sufficiently to expose the reinforcement) and steel plate armouring
(some of which had been totally removed and washed away). The damage was
att.ibuted to the impact of cobbles and pebbles passed through the spillway after
the reservoir had silted up (Bhargava et aI., 1987).
PJ.21.2. Flushing
The po'wer intake incorporates facilities for sediment exclusion (with the
excluded sediment discharged downstream of dam), but no details are given
of any facilities for flushing sediment from the dead storage of the reservoir
basin. It is understood that the only facility for sediment flushing from the
reservoir basin is by opening the spillway, which is done during the rainy
207
EVACUATION OF SEDIMENTS
Table A3.8. Annual inflow and sedimentation datafor the Ichari reservoir
Year Total I Total Sediment capacity: M.m3
(June to water sediment trapped:
May) inflow: inflow: M.m3 Dead Live Total
M.m3 M.m3
1975-76 - 2·62 I
3·93 5·00 8·93
season, whenever the powerhouse is closed. The spillway gates are fully raised,
to allow free flow through the reservoir along the top of the deposits.
Measurements are made during these periods, froin which the quantities of
sediment flushed can be calculated. It appears that flushing by this method has
been undertaken annually since 1976-77 and accounts for between about 30% of
the annual sediment discharge in years with low sediment loads, increasing to
70% or more in years with high sediment loads.
It appears from the information given in Table A3.8 that the regin1e of annual
flushing is likely to result in a fairly stable residual storage capacity of the order
of 4 M.m3 (Atkinson, 1996), but no more recent data are available in the
Iiterature.
208
APPENDIX 3
high elnbankment dam (earth core, rockfill shoulders) diverts ·the flow of the
Baira River to a network of tunnels leading to the powerhouse. The oliginal
storage capacity of the reservoir was 2-4 M.m3 (Paul and Dhillon, 1988)
representing only about 0·1 % of the annual inflow (which is variously reported
as 1900 M.m3 or 3500 M.m 3).
A3.22.1. Sedimentation
The reservoir is subject to both monsoon and winter floods carrying high silt
loads of up to 100 gil (Jaeggi and Kashyap, 1984). A mean annual rate of siltation
had been estimated at 0·092 M.m3 , but in the first 18 months of operation, a silt
volume of 0·45 M.m3 had accumulated, representing allnost 20% of the original
capacity and suggesting an annual sediment load of at least 0·3 Mt (Atkinson,
1996).
A122.2. Flushing
The 5 x 7 m diversion tunnel for the construction of the dam, with an upstream
invert level of 1088 Ill, which is believed to be at least 35 m below the maximum
impoundment level, was equipped with a service gate and an emergency gate to
facilitate flushing. Model studies carried out during the design stage (albeit with
a different design of flushing tunnel) had indicated that almost the entire silt
content upstream of the tunnel could be flushed out.
The first flushing operation. was undertaken in August 1983, with the objective
of achieving the maximum possible volume of silt removal, adopting the
following sequence:
e the reservoir was drawn down to the minimum normal operational water level
(1113 m), following which flows to the powerhouse were stopped
e the diversion tunnel was opened to allow a discharge of 150 m 3/s until
reservoir was almost empty
1/ the diversion service gate was opened fuHy
CD water was fed to t.~e reservoir from,the Suil and Bhaledh to clear silt from
around the associated structures
e flushing ceased when the concentration in the discharge had decreased to
about 10 gil
e the diversion tunnel gate was closed and the reservoir refilled.
209
EVACUATION OF SEDIMENTS
Kashyap, 1984). Bearing in mind the steepness of the valley sides and the small
size of the reservoir basin in relation to the annual inflow, it appears that the
recommended flushing regime should be capable of maintaining a high
proportion of the original storage capacity in the Baira reservoir.
210
Appendix 4.
Erosion
A4.1. FACTORS THAT AFFECT EROSION
A4.1.1. Definition
A fundamental definition of erosion is the detachment and removal of rock
particles by water and other geological agents such as wind, waves and ice
(Mahmood, 1987). A broader definition would include the subsequent removal of
material deposited temporarily at another point in the catchment. The rate of
erosion is generally expressed as the mass of sediment removed from a given
area per year (tlkm2/yr). It vruies with climatic, geological, topographic and
human factors. Sediment yield expresses the quantity of material that reaches a
defined point on a river draining the catchment and therefore the quantity
entering a reservoir created by the construction of a dam at this point. The
quantity will depend on the effectiveness of sediment transport in the basin. The
majority of sediment yield studies consider only the suspended part of the total
load. Bedload is generally assumed to be a minor part, representing about 10%
of the total, even though in extreme cases it can vary between 4% and 60%
(Jansson, 1988).
A4.1.2. Climate
Precipitation
The rate of erosion depends on the erosive power of the rainfall which is related
to the intensity, droplet size and total quantity. High intensity, short duration
events produce more erosion than long duration staTInS of low intensity. Storms
with large rain drops are more erosive than drizzle with small droplets (Goldman
et at., 1986). Tne effect of rainfall intensity is illustrated in Table A4.1 by data
for 183 events which caused erosion at Zanesville, Ohio, between 1934 and
1942. They show that the average soil loss per rain event increases with the
intensity of the storm (Fournier, 1972, reported by Morgan and Davidson,
1986).
Seasonal variations in erosion rates are influenced by previous meteorological
conditions. The moisture content of the soil and hence the infiltration capacity
will depend on previous rainfall and this will affect the amount of run-off which
in turn has a direct on erosion rates (rv10rgan and Davidson, 1986). Highest
erosion rates are likely after a long dry period when there will be a supply of
readily erodible material (Morris and Fan, 1997).
211
J
EVACUATION OF SEDIMENTS
]
Table A4.1. Relationship between rainfall intensity and soil loss
]
Maximum 5 min intensity Number of falls of rain Average erosion: kg/m2
per rainfall: mm/hr
0-25·4 40 0·37
]
I
25·5-50·8 6] 0·6
50·9-76·2 40 1·18
]
76·3-101·6 19 1·14
101·7-127·0 l3 3·42
127·1-152-4 4 3·63
],
152·5-177·8 5 3·87
212
APPENDIX 4
L
800 r
L >-
>- C)' 600
<'t' E
.:s::
E
~
-
~
~
Qi
';:;" 400
Qi
's:, "E
tlJ
"E E
CD
E '5 200
CD
'6 U)
w
en
0
250 500 750 1000 1250 0 200 400 600 800 1000 1200
Effective precipitation: mm Mean annual run-off: mm
(i) (ii)
(a)
10 000
1250 I
L -L
>- 1000
C\I
I ;;'1000
E l-
-
oX
~
a:; 100
-s:,
-
~
~
c::::
a:;
750
';:;"
1:: 1:: 500
tlJ
tlJ
E E
'6 10 '5
tlJ
en tlJ
(j)
25:
J 1' 10 100
I
1000 10000
I
0 400 800
(ii)
1200 1600
-L (i)
>-
L 1200
<'t'
150 >-
~ t
~ 800
E
-c
100 -
-c
Q.i
a:; ':;'
'5-
C 50 'E 400
<ll
<ll
E E
'6 '0
Q)
w 0 (j)
en 400
0 200 600 800 1000 400 800 1200 1600 2000 2400
Mean annual run-off: mm Mean annual precipitation: mm
(iii) (iv)
(b)
Figure A4.1. Sedirnent yield and annual precipitation: (a) USA; (b) world
213
EVACUATION OF SEDIMENTS
and shows the first peak at a preCIpItation of 450 mm, which reflects the
relationship proposed by Langbein and Schumm (1958). The two initial peaks
are similar to the curve produced by Wilson (1969) but the magnitudes of the
peaks are 450 mm compared with 750 mnl, and 1350 mm compared with 1750
mm, which conesponds to the troughs in these curves. The relationship between
sediment yield and Inean annual run-off is similar to that produced by Douglas
(1967) although the peak in semi-arid areas is less pronounced, with maximum
yields occurring in areas of high annual run-off.
Other factors such as relief, geology and human impact may be more
important controls at a global scale than precipitation. The seasonality, intensity
and type of rainfall and its effect on vegetation cover are also important measures
of the effect of precipitation on erosion rates .
10000
'0 •
I .... .. .
. ~:
00 •
>- :. : ..
~
E
-
~
i:i
Q)
1000
o •
':;"
c Q)
E
:0
(!)
C/)
"C
Q)
100
"C
c:
Q)
c..
C/)
:J
C/) :
(ij
:J
c:
c:
ns
c: 10
ns
Q)
:2
o 200 400 600 800 1000 1200 1400 1600 1800 2000 3000 4000 6000 8000
Figure A4.2. Mean annual suspended-sediment yield versus mean annual precipita-
tion
214
APPENDIX4
similar general relationship will exist between run-off and erosion as for
precipitation and erosion; with maximum erosion "levels at intelmediate
values.
o Temperature. The temperature will affect vegetation growth and evapo-
transpiration rates. Where there are high temperatures, higher rates of
evapotranspiration occur and therefore larger amounts of rainfall are required
to cause erosion. High temperatures will also cause more rapid rates of
vegetation growth which will reduce run-off rates and erosion. The ilnportance
of temperature will depend on the quantity of precipitation. In regions with
high precipitation quantities the relative importance of temperature is likely to
be reduced.
• Wind speed and direction. The wind speed and direction will affect the
movement of soil particles. In areas where the wind speed is high and there is
a lack of vegetation to hold the soil pa.'1:icles togell-J.er high rates of wind
erosion are likely to occur. Wind erosion is inlportant in arid or semi -arid
regions as an agent that can transport sediment from ridges into depressions
which can then be transported by run-off.
1200 r-
o~
~
~
800
4DO
;/0 0 \./
0/\ ./
j\ .. · .. \. //
•
\.of
~ 1 I I I I ,
Qi
's;. o 400 800 1200 1600 2000 2400 2800 3200 3600 4000
c
Q)
E Mean annual precipitation:mm .
'6
<ll
en
ca::l 1200
c::
§
c::
ctI
Q)
~
800 0
Figure A4.3. Mean annual suspended sediment yield versus mean annual run-off
215
EVACUATION OF SEDIMENTS
J
Table A4.2. Effect of geology type on erosion rates (Jolly, 1982)
Lithology Sediment loss, Utah: Sediment loss, New Mexico/
m3/km 2/yr Arizona: m3/km 2/yr
A4./J. Geology
Rock type
.J
The geology of the area has a major impact on the rate of erosion as it determines
the susceptibility of the rock to the effect of other factors. Table A4.2 shows that
under the same climatic conditions geology can produce a variation of ten times
J
in the sediment loss.
Generally, suspended-sediment loads are greater in areas of sedin1entary rocks
compared with crystalline rocks by a factor of 2-4 and .compared with areas of
mixed rocks by a factor of 1·4 (Dedkov and Mozherin, 1992).
A4.1.4. Soils
The soils in arid and semi-arid environments with sparse vegetation cover are
:
very different from soils in more humid regions. The key soil ·characteristics i
influencing erosion rates are the texture, structure, organic matter content, shear .,.....L
Texture
The texture describes the sizes and proportions of the particles making up the
soil. Soils with high sand contents are coarse textured with high infiltration rates,
low run-off and relatively low erosion potential. Soils with a high content of silts
and clays are fine textured, the clay binds the soil and makes it resistant to
erosion. Soils high in silt and fine sand and low in clay and organic matter are the
nlost erodible. Well drained sandy and rocky soils are the least erodible as they
have large particles which require large forces to transport theln (Goldman et ai.,
1986).
216
APPENDIX 4
Soil structure
Soil stnlcture is the anangement of particles into The soil structure
affects the soil's ability to absorb water. When the soil is compacted or
crusted, water tends to run off rather than infiltrate. Granular structure is the most
desirable to minimise erosion as it absorbs and retains water, reduces run-off and
encourages plant growth (Goldman et al., 1986).
Organic content
Organic matter improves the soil structure and increases permeability, water
holding capacity and soil fertility (Goldman et al., 1986). Clay content can be
used as an indicator of erodibility as it combines with organic matter to form soil
aggregates and it is the stability of these particles which determines the resistance
of the soil. Soils with an organic content of less than 3 ·5% are highly erodible.
Shear stren~J1
This is a measure of the cohesiveness of a soil and its resistance to shearing
forces exerted by gravity, moving fluids and mechanical loads. Its strength is
derived from frictional resistance met by its constituent particles when they are
forced to slide past one another or to move out of interlocking positions. The
higher the shear strength of a soil the more resistant it is to erosion (Morgan and,
Davidson, 1986).
Infiltration ratelpermeabiliv/
infiltration capacity is the maximum sustained rate at which soil can absorb
water and is influenced by pore size, pore stability and the form of the soil
profile. Soils with stable aggregates maintain their pore spaces better while soils
with swelling clays or minerals that are unstable in water tend to have low
infiltration capacities. Where infiltration varies with depth, the horizon with the
lowest infiltration capacity is critical. Texture, structure and organic matter all
contribute to the pelweability of a soil. High erosion rates occur where
rates are low and large volumes of run-off are ....,"". . . ""............ ' -' .
...l ... .l,u, .............. v ,...
Orientation of catchment
Southern facing slopes in the northern hemisphere are eroded more rapidly than
·~~"',o ... facing slopes as they are hotter and drier with less dense vegetation and
1"\ ..... 't"I
217
EVACUATION OF SEDIMENTS
they experience greater fluctuations in air and soil temperature. North facing
slopes are cooler and more moist with less sun.
105
x
104 x
o o
E 103 +
o
-
~
i::i o
o
o
05
";;:.. +OX x
C
o +
~ 102 1::.. q.
0 +
'5
Q)
C/) + o 1::..0 -k
M
o 0 0 1::..
10 1 ~
8
+
10oL-~--~---L--~--~~~~--~--~--~--~~
218
APPENDfX4
relationspip between suspended sediment yield and drainage basin area (Walling
and Webb, 1996),
Drainage density
Drainage density is all expression of the distribution of streams in the drainage
system. It is a crude indicator of run-off and is often used as an index of the
severity of erosion - areas of high drainage density being associated with
elevated erosion rates. Broadvariations in drainage density on a macro-scale are
associated with differences in climate. At the meso-scale, regional variations can
be related to differences in rainfall volume but are complicated by lithology and
relief. Micro-scale differences in soil type and frequency and intensity of
individual climatic events are important (Morgan and Davidson, 1986).
Vegetation
Vegetation is the most important erosion control factor. It dissipates the energy
of rainfall, prevents rain impact on the ground, reduces splash erosion, increases
infiltration, decreases surface nln-off volumes and velocity, holds soil particles in
place and maintains the soil's capacity to absorb water.
The type of vegetation cover is dependent on the raiilfall, temperature, soils
and topography of a region. These factors interact to produce distinctive zones
called biomes. Climates with relatively mild year-round temperatures and
frequent~ regular rainfall·are favourable to plant growth. Cold and dry climates
are less favourable to growth and therefore more susceptible to erosion (Goldman
et ai., 1986).
Land use
The land use of an area is influenced by the topography, geology, soils and
climate of a region. These factors determine the use to 'which land is put by
humans. Cultivation may decrease the erodibility of clay soils but increase that
of sandy soils (Morgan and Davidson, 1986).
Soil loss from hillslopes in West Africa between a gradient of O·J:) and 4°
experienced mean annual erosion rates of 0·015, 0·02 and 0·003 kglm2 under
natural conditions of open savanna grassland, dense savanna grassland and
tropical rain forest respectively. Clearance of the land for agriculture increased
rates to 0·8, 2·6 and 9·0 kg/m2 while leaving the land as bare soil produced rates
of 2, 3 and 17 . The removal of rain forest produces greater rises in erosion
rates than the removal of savanna grassland (Morgan and Davidson, 1986).
Areas of low precipitation are more vulnerable to land use changes. Changes
in one part of the ecosystem may produce changes in the basin condition and
response and recovery may take a long period of In semi -arid regions the
recovery time is four times that of humid areas (Walling and Kleo, 1979).
The effect of cultivation or lack of soil cover by on erosion rates can
be seen in Table A4.3.
The increase erosion rates from natural to bare ground can also
be seen graphically in Figure A4.S which shows the results of soil erosion tests
under different vegetation cover at Mpwapwa, Tanzania.
219
EVACUATION OF SEDIMENTS
Table A4.3. Rates of erosion in selected countries in kg/m2/yr (after Morgan and
Davidson, 1986)
Natural Cultivated Bare soil
Human impact
It is estimated that human activities have degraded 15% (200 million ha) of the
land between 72°N and 57°S. Around half of this is due to hUlnan-induced water
erosion, a third due to wind erosion with most of the balance due to chemical and
physical deterioration (US Global Change Research Information Office, 1999).
Activities such as deforestation, urbanisation and agriculture all increase the
erodibility of soil. Present rates of erosion are approximately two and a half times
historic rates mainly due to human influences. With the conversion of forest to
agricultural land there Inay be an increase in sediment yield at the basin mouth
by three and a half times (Mahmood, 1987) - see Table A4.4.
Explanation:
Ot
1·9%
rI
Water lost by run-off
per cent of rainfall
26·0%
50-4%
220
APPENDIX 4
Table A4.4. Erosion rates for different land use categories (Morris and Fan, 1997)
Land Llse
Forest 8
Under natural conditions erosion rates in mountain zones are 27 times greater
than in lowland areas. The influence of man has increased sediment yield from
mountainous regions by 1·4 times; however, larger increases in lowland areas
have reduced the difference between mountainous and low land regions to 3·2
. times. For eX3J.I1ple, sediment yields in the sout.h and middle Urals are up to
~ 30 tlk..rn2/yr, less than the neighbouring eastern part of the Russian plain where
rates of up to 200 tlkn12/yr occur (Dedkov and Moszherin, 1992).
Increases in sediment yield caused by human activity are demonstrated in
Table A4.S.
221
I'
EVACUATION OF SEDIMENTS
Table A4.S. Increases in sediment yield due to land use changes (Walling and Webb,
1983)
Region Land use change I Factors for Source
increase in
sediment yields
Texas, USA Forest clearance and cultivation 340 Chang et al. (1982)
Mississippi, USA Forest clearance and cultivation 10-100 Ursic and Dendy (1965)
Southern Brazil Forest clearance and cultivation 4500 Bordas and Canali (1980)
The altitude has a distinct effect on the climate causing a knock-on effect on
the vegetation of each zone, as shown in Table A4.6. There are two rainy seasons:
from March to May and from October to December.
Sediment yields from the forested areas of Mount Kenya are around
20 t/km2/yr rising to 1000 t/km2/yr on grazed areas and lnore than 3000 t/km2/yr
on steep cultivated areas of the basin. This shows the strong influence of land use
on sediment yield. Under natural conditions cultivated areas have ground cover
of crops for around eight months of the year. Soil losses from grazing lands are
therefore generally higher and increase markedly as basal cover declines. In
cultivated areas, rural roads yield 5-20% of the sediment yield (Ongweny,1979).
Table A4.6. Variations in rainfall, vegetation and soils with altitude (Ongweny, 1979)
Altitude: m Rainfall: mm Vegetation Soil
222
APPENDIX4
Comparisons between the data for the Upper Tana basin and that produced by
Dunne, relating mean annual suspended-sediment yield to annual run-off
depending on land-use, showed good agreement. This can be seen in Figure
A4.6, which shows the relationship between mean· annual suspended-sediment
yield and mean annual run-off.
-- --
1~~~ __~-L__~~~~~~--~~--~--~~--~
a 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500
Mean annual run-off: mm
1 2
III 0 Forest
r:. 0 Forest> Agriculture
• • Agriculture> Forest
o 0 Grazing
A Scrub lorest
Figure A4.6. Mean annual suspended-sediment yield plotted against annual run-off
223
EVACUATION OF SEDIMENTS
. - :!
In all three regions the slopes are longer than 500 m with only a small portion
of gradients over 0·1. The n1aximum rates of erosion occur at the central, steepest
part of the hillslopes. The sediment yield on the slopes is 2500 t/km 2/yr to
3150 tlkm2/yr, on the basement rocks 11 300 tlkm2/yr and on Kilimanjaro lavas
17 600 t/lGl1 2/yr (Dunne et a!., 1979).
NOlthern 12-2045
East 27-356
South 30-1787
224
-~. APPENDIX 4
Table A4;8. Sediment contribution for sections of the Yellow River (Tal Wei Soong and
Yean Zhao, 1994)
I Area: km'
I
Length of
channel: km
Average
slope: m/m
I Contribution:
% water
Contribution:
% sediment
I
Upper 385000 3472 111000 48-7 9-0
1 1 I
Middle 345000 1206 111400 36-6 89·3
1 I !
Lower
I 22000 786
I
1/8000 I
I
11
I 1·7
Sanmenxia reservoir
The drainage basin to the reservoir is 688 400 km2, constituting 92% of the
Yellow River basin. The Yellow River drains China's semi-arid loess plateau
composed of thick aeolian deposits of silty soils. Due to the high erodibility of
this soil, intensive land use, inadequate soil conservation practices and virtually
limitless supply of sediment, the load through the valley is high. Rates of
sediment trarlsport are especially high in July and August which account for 60%
of the total annual sediment yield and 30% of the annual run-off. Sediment
discharge averages 1·6 billion tlyr, equivalent to an annual sediment yield of
2300 t!k..rn2 and a mean suspended-sediment concentration of 38 gil ~Aorris and
Fan, 1997).
The relationship was based on selected reservoir sites in various areas of India
(Lagwanker et aI., 1985). '
225
EVACUATION OF SEDIMENTS
226
APPENDIX 4
tough and easily erodible siltstone or sandstone with nlore granitic rock above
Caonillas reservoir.
Loiza reservoir
The region has an average rainfall of 1900 mmJyr (841 mrn run-off) and
teluperature of 25°C. The dam impounds 534 kn12 of the Loiza catchment of
which over half the area has slopes greater than 35°. The sediment yield of the
region is high, between 1000 tfKm2/yr and 2000 ttlCIl12/yr (Monis and Fan,
1997).
227
EVACUATION OF SEDIMENTS
Group I
Specific sediment yield less than 150 t/km2fyr. Twenty catchments in zones
where erosion processes are slow, or where carbonate lithologies predominate.
Group 2
150-1000 t/km 2/yr. This group contains over 500/0 of the reservoirs.
Group 3
Over 1000 t/km2/yr. This group includes 7 basins which are less than 420 km2 so
that eroded sediment is likely to be transported into the reservoir.
These relationships describe the sediment yield at the reservoirs, which is not all
the material eroded. The sediment delivery ratio varied from 0·8% to 67-470/0
with most ratios less than 25%. Those catchments with the lowest specific
sediment yields are not always those with the lowest erosion rates. The surface
area and location of the sediment source, relief, slope, transpo11 system and
vegetation cover all affect the delivery ratio. The predominant factor is the
drainage basin area (Salas et aI., 1997a and 1997b)
228
APPENDIX 4
A4.3.2. Thornthwaite
.il..n alternative climatic classification was put forward by Thornthwaite (see Table
A4.9), based on the relationship between precipitation and potential evapo-
tra...l1spiration.
The classification calculates an index I which expresses the relationship
between surplus moisture, 5, moisture deficiency, d,and potential evapotranspira-
tion, pe.
1= (5 - d)lpe x 100
The zero value separates moist (positive) from dry (negative) climates. To
incorporate a thermal parameter in the classification Thomthwaite used potential
evapotranspiration as this parameter expresses water need as a function of
temperature and length of day (UNESCO, 1990).
A4.3.3. Alisov
A classification by Soviet climatologist Alisov is based on the conditions of
circulation of the atmosphere. He identified seven main climatic zones which can
be seen in Figure A4.8. Each zone is characterised by the predominance of the
air mass corresponding to the name of the zone (UNESCO, 1990).
229
IV m
w <
(:)
»
()
c
»
--I
0
Z
0
-n
l/)
m
0
-
3:
m
Z
--I
l/)
C. Humid
nwtOlMtm.1 D
'0
Wafm wilh dry win'., lewl
Imon_ • upllllld .....Mll
W..m whh dry ILl"""" IC.I
cllmal"
o IMtdil..,_enl
Humid I....".,.'. ICII
O. Humid
mlcrotMfma'
f11M Cold ",ilh main w'nter lOt)
I·.,·-t-J~~
ctifNl" lmmCoId with lily win.er 10wi
~ ImonJOOf\ IYpe)
• . f'ollf
clime..,
_1 J .J J J ' ]
APPENDIX4
B4 80-100
I
B3 Humid Meso thermal 99·8-114 60-80
I
Bz 85·6-99·8 40-60
B, 71·3-85·5 20-40
5
60~----~----~~~-------4------~~~------r---'------+--------~
231
EVACUATION OF SEDIMENTS
cold winter. Consideration was also given to the time of the rainy periods and to
the length of the dry peliod (UNESCO, 1990).
A4.3.5. Budyko
Budyko (1956) devised a climatic classification based on a 'radiational index of
dryness' as a means of rating the efficiency of the moisture supply. This is
illustrated in Figure A4.9.
The index is given by:
1= RnlLn
where Rn is the annual average net radiation in cal/cm2
Ln is the energy (in calories) required to vaporise 1,ocal precipitation
This index expresses the relative values of the heat and water balances (see
Table A4.10).
For each index RnlLn there is a corresponding value of the run-off
coefficient.
232
APPENDIX4
0
:!
0
~
~
0
~
0
:!
0
~
~
i
=
0
..
<0
0 0
'Of
~
0
0
~
..-."
Cl
~
0
CI ~.
e
~
~
:....
~
0 ¢:..
~
0
~
--Co
c..::
c..::
~
~
0
:
-e
~
C)"
)...,>
~
~
.::;
~
'';::
.~
-
'\::
C
,..~
0\
~
~
~
::::
CJ:
t;:
233
EVACUATION OF SEDIMENTS
30~----~--------~~~~-----------
25
20
15
Warm temperature
o Cold temperature
-5
-10
-15L-----~~-----L------~------~----~------~------~------~----~
o 50 100 150 200 250 300 350 400 450
Total annual rainfall: em
234
APPENDIX 4
Rain forests
Abundant ai110unts of rainfalL Olympic Rainforest: located on the west coast of
USA, on the OlYlnpic peninsula in \Vashington State. Warm climate.
Tropical rainforest
Abundant amounts of rainfall. Located near the Equator. Wann climate.
Grasslands
The typical rainfall is 25 cmJyr to 30 crrJyr. The lowest rainfall occurs in a desert
biome. The grass has roots that can penetrate into the soil to obtain some water,
but the amount of water available is not enough to sustain much tree life.
Biomes are explained more fully at:
o http://sheepshead.usl. edufT......aCEPTIrainforest.ht!pl
235
m
Table A4.11. Country data for Asia ~
n
Climatic Rates of sediment Example rates: No. of Total c
Country
classification yield: tlkm 2/yr tlkm 2/yr dams capacity
% Li:lP,lL Il Y
lost to
Annual
loss of
»
--I
(Koppen) (Walling and Webb, (from literature) data for of dams: sediments storage o
1983) M.m} Z
o
."
Afghanistan BwklBsk 100-250/250-500 28 0 800·0% 0·0%
Vl
m
Armenia DfiDs 100-250 CJ
----
3:
Azerbaijan DflDs 100-250 m
Z
--I
Bahrain Bwh <50 Vl
-----
Brunei Af <50
Cambodia Aw 250-500
Kyrgy:-:tan Df 100-250
Island
-
Maldives Csb
- -
Mongolia Bwk/Bsk 50-100
Singapore Af 250-500/500-750
Tajikistan Ds 250-500
Congo, Aw <50
Democratic
RepUblic
Ghana Aw 50-100
Guinea Aw 50-100
---
Liberia Am 50-100
--------
L
Mauritania DwhlBsh <50
Nigeria Aw 50-100 33
~-~.
Country Climatic Rates of sediment Example rates: No. of Total % capacity Annual
classification yield: tlkm2/yr tlkm2/yr dams capacity lost to loss of
(Koppen) (Walling and Webb, (from Ii terature) data for of dams: sediments storage
1983) M.m3
.1
'j
Fiji Af Island
Kiribati Af Island
Micronesia Af Ishmd'
Tuvnla Af Island
»
iJ
Vanuatu At" Island iJ
---- I----- . ITI
Z
Western Samoa Af Island
o
~
w
x
~
m
Table A 4.14. Country data for Europe
~
()
Country Climatic Rates of sediment Example rates: No. of ~
Total % capacity Annual C
t/km2/yr
classification
(Koppen)
yield: tlkI1l2/yr
(Walling and Webb, (from literature)
darns
data for
capacity
of dams:
lost to
sediments
loss of
storage
~
1983) M.m3 o
z
Albania Ctb 500-750 4150/3590 o
"m
Vl
"J "\
! 'I J
Hungary Ctb 100-250
Moldova Of <50
-
Monaco Cfb 250-500
1983) M.m3
) , -)
Table A4. J5. Country data for North America
Country Climatic Rates of sediment Example rates: No. of Total % capacity Annual
c1nssi1ication yield: tlkm2/yr t/km2/yr dams capacity lost to loss of
(Koppen) (Walling amI Webb, (from literature) data for of dams: sediments storage
1983) M.m3
- _..__.
Antiglla amI At" 1.00-250
Barhuda :
-
Bahamas Af 100-250
Belize Af lOO-250
Cuba Af 100-250
Dominica Af 100-250
- -
Dominican Af 100-250
Republic
-------.--- --
EI Salvador Aw 100-250
-
Greenland Hf 50-100
Grenada Af 100-250
-
Guatemala Aw 100-2501250-500
Hniti Af 100-250
----
m
~
()
C
Table A4.15. continued ~
o
Country Climatic Rates of sediment Example rates: No. of Total % capacity Annual z
classification yield: tlkm 2/yr tlkm2/yr dams capacity lost to loss of o
(Koppen) (Walling and Webb, (from literature) data for of dams: sediments storage "
(/)
1983) M.m 3 m
o
Honduras Aw ]00--250 3:
m
Z
Jamaica Af 100-250 -I
(/)
Nicaragua Aw 100-250
--------_.
Panama Af 100-250
USA DfblDfaiB w/Cfal < 50/50-1 001 407212374122921 1105 109980 3·9% 0·2%
CsblBs/Df 100-2501250--5001 1167/500114511 071
500-750/7 50--10001 71150112
>1000
1
Table A4.16. Count}")1 data for South America
Country Climatic Rates of sediment Example rates: No. of Total % capacity Annual
classification yield: tlkm 2/yr tlkm 2/yr dams capacity lost to lo~s of
(Koppen) (Walljng and Webb, (from literature) data for of dams: sediments storage
1983) M.m3
Suriname Af <50
---- -
UllIgtluy Cra <50 »
-0
--- -0
Venezuela Aw/Af <50/100-250 212 1 3 19·2% 9·6% m
Z
o
~ x
~ ~
Index
251
EVACUATION OF SEDIMENTS
252
INDEX
253
EVACUATION OF SEDIMENTS
254
INDEX
255
EVACUATION OF SEDIMENTS
256
INDEX
257
EVACUATION OF SEDII'1ENTS
258
INDEX
259
EVACUATION OF SEDIMENTS
260