Вы находитесь на странице: 1из 280

,

" ~

Evacuation of sediments from reservoirs

Rodney White

I ThomasTelford
I

\
I

..,, :-

'\

Published by Thomas Telford Publishing, Thomas Telford Ltd, 1 Heron Quay, London E14 4JD.
URL: http://www.thomastelford.com
Distributors for Thomas Telford books are
USA: ASCE Press, 1801 Alexander Bell Drive, Reston, VA 20191-4400, USA
Japan: Maruzen Co. Ltd, Book Department, 3-10 Nihonbashi 2-chome, Chuo-ku, Tokyo 103
Australia: DA Books and Journals, 648 Whitehorse Road, Mitcham 3132, Victoria
First published 2001

Front cover shows reservoir sedimentation in Zimbabwe

- j
i

--I
,i

:-1
~I

-j

-I

A catalogue record for this book is available from the BIitish Library
ISBN: 07277 2953 5
Rodney White and Thomas Telford Limited, 2001

All rights, including translation, reserved. Except as pennitted by the Copyright, Designs and Patents Act
1988, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form
or by any means, electronic, mechanical, photocopying or otherwise, without the prior written permission of
the Publishing Director, Thomas Telford Publishing, Thomas Telford Ltd, 1 Heron Quay, London E14 4ID.

II
:--1

-I-~

This book is published on the understanding that the author is solely responsible for the statements made and
opinions expressed in it and that its publication does not necessarily imply that such statements and/or
opinions are or reflect the views or opinions of the publishers. While every effort has been made to ensure that
the statements made and the opinions expressed in this publication provide a safe and accurate guide, no
liability or responsibility can be accepted in this respect by the authors or publishers ..
Typeset by Apek Digital Imaging, Bristol, UK
Printed and bound in Great Britain by MPG Books, Bodmin, Cornwall

-I

Preface
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.

I. Minimising sediment loads entering reservoirs


There are three common ways of achieving this objective:
catchment conservation prograrIh'TIes to minimise sediment yields. Land use
practices, agriCUltural methods and engineering measures to control erosion all
feature in this category.
o upstream trapping of sediments. Check dams and veget~tion screens can be
used to intercept sediments on their way to downstream reservoirs .
bypassing of high sediment loads. The principle here is to fill the reservoir at
low to medium flows when sediment concentrations are low and to bypass
high flows, with their high sediment content, around reservoirs. This can be
achieved using bypass channels or tunnels or by having the reservoir 'off
line'.

2. Minimising deposition

of sediments in reservoirs

There are two main ways of passing sediments through reservoirs without
deposition:
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.
a

3. Removing accumulated sediments from reservoirs


Hydraulic and mechanical methods are available for removing sediment which
has already accumulated in reservoirs:

iii

EVACUATION OF SEDIMENTS

flushing - the process of re-entraining deposited sediments and passing the


sediment laden flow through low level outlets in the dam. This method
involves the reduction of water levels in the reservoir, it consumes significant
quantities of water but is capable, under certain circumstances, of removing
coarser sediments (mainly sand sizes).
It removal of sediments using dredging or mechanical means
this method is
feasible but usually requires reservoir levels to be maintained at low levels for
extended periods of time. It is expensive in itself, quite apart from the loss of
benefits from the reservoir during dredging operations. The disposal of large
quantities of sediment often presents problems.
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:
It

It

the availability of suitable engineering facilities at the dam to control water


levels and outflows
the availability of 'surplus' water and its value if used for other purposes
the predictability of dver flows, including seasonal variations
the characteristics of the sediments entering, and within, the reservoir
the availability of disposal sites for dredged sediments
the effects on the downstream reach of evacuating sediments through the dam
th~ effects of sediment management on the normal operation of the scheme
and the financial and social consequences of the measures taken
the effects of sediment management on other reservoirs within the region
institutional and political problems among the affected stake-holders.

The objective of making reservoirs more sustainable using sediment


management techniques is clearly laudable. However, the techniques are not
applicable to all reservoirs and some dams will inevitably need to be either raised
to regain storage or decommissioned and possibly replaced elsewhere. However,
there are fewer and fewer good dam sites available and new dams can have
sedous environmental and social consequences.
This book is concerned principally with one of the methods of removing
previously deposited sediments from reservoirs, namely the flushing of
sediments through purpose-built outlet works within the dam. This technique can
be applied to existing dams (with adaptation of the engineering works) and to
new dams. However, the technique is only effective under certain favourable
conditions and is not applicable universally.
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 book is to give 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

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.

Rodney lVhite developed his interest in hydraulics at Leeds University from


where he gained his first degree in 1962 and his PhD in 1965. He joined the
Hydraulics Research Station, as it was, in 1965 and specialised in flo1;V
measurement and sediment transport during his early career. He led the River
Engineering Department of HR Wallingford before becolning the Research
Director in 1990. In more recent years he has been a consultant to the firm with
a remit to develop and apply new technologies, particularly with regard to
sediment related issues.
His research has resulted in internationally accepted theories that explain the
movement of sediment in rivers, the resistance of naturally fonned alluvial
channels, the equilibrium, size of natural channels and their plan form
characteristics. He has extensive practical experience of sedim.entation in rivers .,
and reservoirs worldwide. He has written several books and many scientific
papers based on his research and on his specialist consultancy assignments and
he is currently the editor of the International Association of Hydraulic
EngineeringResearch (IAHR) lournal of Hydraulic Research.

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

groundwater
urban drainage
rivers
tidal waters
ports and harbours
coastal waters
offshore.

Address:
Internet:

Howbery Park, Wallingford, Oxon, OXlO 8BA, UK


http://www.hrwallingford.co.uk

vii

Notation
DDR

DSOT
D50B
F~VR'

H fiush

Hmax

LTCR
lv!in
LV

P sand

Qf
Qm
Qs
S

SBR
SBR d
SSR
TE

Tf
TIVR
\tV

Ibed

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 (nondimensional)
50 percentile size of sediment in transport (IILrn)
50 percentile size of river-bed material (mm)
the ratio of the natural width of the flushing channel and a
representative bed width for the reservoir (non-dimensional)
the height of water at the datu during flushing, measured above
original river-bed level (m)
the maximum height of water at the dam (to reservoir retention
level), measured above original river-bed level (m)
the sustainable storage capacity divided by the original storage
capacity of the reservoir (non-dimensional)
the average sediment inflow rate (t/yr)
the interval between flushing operations (yrs)
proportion of total sediments in motion which exceed 0-06 rom
in size (sand and coarser material)
the flushing discharge (m3/s)
mean annual flow (m 3/s)
the sediment transporting capacity of the flow in the incised
flushing channel (tis)
the longitudinal energy gradient trliough the reservoir (nondimensional)
the ratio of sediment flushed to sediment depositing (nondimensional)
the specific value of SBR related to flushing with maximum
reservoir drawdown (non-dimensional)
.~
sand-size ratio, DsoTIDsoB
the trapping efficiency of the reservoir, i.e.
ratio of sediments
retained within the reservoir to sediments~.;,entering (nondimensional)
.
the duration of flushing (days)
the ratio of the natural top width of the flushing channel and a
representative top width for the reservoir
the bed 'width of the incised flushing channel (m)
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
Wmin

a
p

"

the representative top width of the reservoir, taken as the top


width of the reservoir just upstream of the dam (m)
the lesser of Wand Wbed (m)
the angle of the side slope of the incised channel formed during
flushing (zero is horizontal) (degrees)
the density of the deposits expressed as weight of dry material
per unit volume (tlm3 )
constant related to the sediment type (non-dimensional)

Contents

Dlustrations

1.

Executive summary
1.1.
1.2.

2.

xv

Review of sedimentation
2.1.
2.2.

2.4.

2.5.

2.6.

Introduction, 3
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

reservoirs

15

Summary, 17
World total reservoir storage, 17
2.2.1. ICOLD World register of dams, 17
2.2.2. Other sources, 18
2.2.3.
Conclusion, 18
vVorldwide distribution of existing storage, 18
2.3.1.
Global water resources, 18
2.3.2.
Geographical distribution, 19
World demand for more storage, 19
2.4.1. Population, 19
2.4.2.
Irrigation, 21
2.4.3.
Hydropower, 21
2.4.4.
Conclusion, 23
Distribution of demand for more storage, 23
2.5.1.
Europe, 23
2.5.2.
North America, 24
2.5.3.
South and Central America, 25
2.5.4.
Aflica, 26
2.5.5.
Asia and Oceania, 27
2.5.6.
Summary, 28
Rate and distribution construction of new

xi

EVACUATION OF SEDIMENTS

2.7.
2.8.
2.9.
2.10.

xii

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
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
Trends in the rate of loss of storage, 34
Reservoir size and rate of loss of storage, 35
Requirements for new storage, 36

3.

Research into factors which influence flushing


3.1.
Introduction, 39
The mechanism of flushing, 40
3.2.
The development of criteria for successful
3.3.
flushing, 42
3.3.1. Sediment balance, 42
3.3.2. Sustainable reservoir capacity, 47
3.3.3. Evaluation of flushing criteria, 50
3.3.4. Practical criteria for successful flushing, 50
Summary of the requirements for effective
3.4.
flushing, 58
3.4.1. Hydraulic conditions required for efficient
flushing, 58
3.4.2. Quantity of water available for flushing, 59
3.4.3. Mobility of reservoir sediments, 59
3.4.4. Site-specific factors, 60
3.4.5. Constraints on the ultimate capacity
achievable by sediment flushing, 60
3.4.6. Economic assessment, 60
3.4.7. Summary, 60
Numerical models, 61
3.5.

37

4.

Worldwide experience of sediment flushing


Introduction, 65
4.1.
Flushing,
66
4.2.
Worldwide experience of flushing, 67
4.3.
4.3.1. Overview, 67
4.3.2. Flushing teclmiques, 67
4.3.3. Sediments flushed, 70
Case studies of reservoir flushing, 71
4.4.
4.4.1. Summary, 71
4.4.2. Findings, 81
4.4.3. Summary of findings, 88

63

CONTENTS

5.

Geographical areas suited to flushing


5.1.
Worldwide variation in erosion rates, 93
5.1.1. Factors that affect erosion, 93
5.1.2. Estimates 9.f global sediment yield, 93
5.1.3. Maps of global variation in sediment
yields, 100
5.2.
Climatic zones of the world, 101
5.2.1. Introduction, 101
5.2.2. Precipitation regimes and their seasonal
variation, 101
5.2.3. Koppen classification, 108
5.2.4. Relationship between climate zone and
erosion rates, 115
5.3.
Geographical areas suitable for flushing, 120
5.3.1. Introduction, 120
5.3.2. Factors affecting erosion rates, 120
5.3.3. Sediment delivery ratio, 122
5.3.4. Hydrological characteristics, 123
5.3.5. Areas of the world which are best suited to
reservoir flushing, 123

6.

Site-specific investigations and design considerations

125

7.

References

131

8.

Bibliography

141

Appendices
Appendix 1.
Appendix 2.
.8..Dlue:nWlX 3.

91

149
Reservoir data, 151
Numerical model case study, 163
Flushing case studies, 171
Erosion, 211

251

xiii

I!' ..

I
Illustrations

Tables

Table 1.L
Table 1.2.
Table 1.3.
Table 1.4.
Table 1.5.
Table 2.1.
Table 2.2.
Table 2.3.
Table 2.4.
Table 2.5.
Table 2.6.
Table 2.7.
Table 2.8.
Table 2.9.
Table 2.10.
Table 2.11.
Table 2.12.
Table 2.13.
Table 2.14.
Table 2.15.
Table 2.16.
Table 2.17.
Table 2.18.
Table 2.19.
Table 2.20.
Table 2.21.
Table 2.22.
Table 3.1.

Demand for storage


Geographical demand for new storage
Demand for new storage, South and Central
A.merica
Demand for new storage, Africa
Gross storage requirements to 2010
Distribution of reservoir storage volume
Growth in 'world population
Prediction of global demand (after Shiklamanov)
Growth in irrigation area
Growth in energy generated by hydropower
Comparison of actual andeconolTlJcal potential
energy
Annual gro\tvth rates and increase in storage
European growth in irrigation and hydropower
North American growth in irrigation and
hydropower
South American growth in irrigation and
hydropower
Annual growth rates, South and Central America
African growth in irrigation and hydropower
Asian and Oc~anian growth in irrigation and
hydropower
Annual growth rates, Asia and Oceania
Regional demand for new storage
of construction of new storage
Distribution of storage increase
Regional sedimentation rates
Extent of sediment data
Distribution of sediment rate and storage loss
Ringlet reservoir, sedimentation
Gross requirement for new storage
Application of sediment balance and long-term
capacity ratios to existing reservoirs

5
5

6
8

19
20
21
21
22
22
23
24
25
26

26
27

28
29
29

29
30
31

32
33
34
35

46

xv

EVACUATION OF SEDIMENTS

Table 3.2.

Table 3.3.
Table 3.4.
Table 3.5.
Table 3.6.
Table 4.1.
Table 4.2.
Table 4.3.
Table 4.4.
Table 4.5.
Table 4.6.
Table 5.1.
Table 5.2.
Table 5.3.

Table 5.4.
Table 5.5.
Table 5.6.
Table 5.7.
Table 5.8.
Table 5.9.
Table 5.10.
Table 5.1l.
Table 5.12.
Table 5.13.
Table 5.14.
Table 5.15.

Figures
Figure 2.1.
Figure 2.2.
Figure 2.3.

xvi

The relative importance of the discharge used for


flushing and the duration of flushing for a
particular volume of flushing water
The influence of sediment size on the amount of
sediment removed from reservoirs
Effect of the sediment size ratio on sediment
balance
Effect of the proportion of sand and coarser
material on extending the life of reservoirs
Application of constraint criteria to existing
reservoirs.
Summary of reservoirs flushed
Summary of experience in flushing
Distribution of flushing experience by purpose
Summary of flushing techniques
Detailed list of reservoirs subject to flushing
Summary of key flushing parameters
Continental variations in sediment yield
(Mahmood, 1987)
Continental variations in sediment yield (Jolly,
1982, taken from Gregory and Walling, 1973)
World maximum recorded suspended-sediment
yields greater than 2000 tlkm2/yr (Jolly, 1982,
from Gregory and Walling, 1973)
Rates of sediment yield for the world's maj or
rivers at ocean level, excluding basins with an area
less than 10000 km2 (Mahmood, 1987)
Values of sediment yield in excess of
10000 t/km2/yr (Walling and Webb, 1983)
Colombia, SON, elevation 65 m
India, 13 oN, elevation 22 m
Wadi HaIfa, Sudan, 22N, elevation 160 m
England, 515N, elevation 5 m
Calgary, Canada 51N, elevation 329 m
Italy, 42N, elevation 131 m
Greenland, 815N, elevation 35 m
Antarctica, 665S, elevation 30 m
Reasons for combining climates into
homogeneous climatic groups (Jansson, 1988)
Countries classified into climatic zones showing
number of river basins in each zone (modified
from Jansson, 1988)
Growth in world population
Comparison of growth rates
Historic growth in reservoir storage

52
54
55
56
58
68
69
70
70
72
74
94
94

96
97
99
112
112
113
113
114
114
115
115
117
119
20
23
30

ILLUSTRATIONS

Figure 2.4.
Figure 2.5.
Figure 3.1.

Figure 3.2.
Figure 3.3.
Figure 3.4.

Figure 5.1.
Figure 5.2.
Figure 5.3.
Figure 5.4.
Figure 5.5.
Figure 5.6.
Figure 5.7.
Figure 5.8.
Figure 5.9.

Storage lost to sedimentation


Reservoir size and rate of loss of storage
Longitudinal profiles during flushing: (a) flushing
with full drawdown; (b) flushing with insufficient
drawdown; (c) final conditions after a long period
of flushing with insufficient drawdown
Channel widths formed in reservoir deposits
during flushing
Cross-sections of flushing channels: (a) Heisonglin
reservoir, China; (b) Sanmenxia reservoir, China
Simplified reservoir geometry for application of
capacity criterion: (a) actual reservoir plan;
(b) fitted reservoir plan; (c) simplified reservoir
plan and sections; (d) simplified reservoir
elevation; (e) enlarged section immediately
upstream of dam
Global patterns of sediment yield: (a) after
Strakhov (1967); (b)
Fournier (1960)
Global patterns of suspended sediment yield: (a)
from Lvovich (1991) in \Valling and Webb (1996);
(b) from Walling and Webb (1983)
Annual precipitation for 1998 in mm per month
Precipitation distribution during winter 1998
(December to February)
Precipitation distribution during spring 1998
(March to May)
Precipitation dist.ribution during summer 1998
(June to August)
Precipitation distlibution during autuIllll 1998
(September to November)
. Climates of the world according to the Koppen
classification
Number of basins within sediment yield classes in
climatic groups

33
35

41
44

48

49
102
104
106
107
109
110
111
116
118

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.
Flood control and storage
Many dams have been built with flood control and storage as the main motivator,
e.g. the Hoover dam, the Tennessee Valley dams and some of the more recent
dams in China.
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 purposebuilt 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

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 sitespecific. 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.

1.2.

SUMMARY OF CONCLUSIONS

-.i

1.2.1. Review of sedimentation in reservoirs


World storage
The best estimate of world storage in reservoirs (excluding natural lakes used as
storage for power and irrigation) is 6815 km3
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.
Demand for more storage
The world population in 1990 is estimated to have been 5286 million, growing
at an annual rate of 15%. 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. Much of this demand will be satisfied by

-~

EXECUTIVE SUMMARY

Table 1.1.

Demand/or storage

Period

Increase in net storage:


km3

Annual growth rate:


%

increased surface and groundwater abstraction; water re-use and no direct


linkage between overall demand and water storage can be assumed.
An estimate of the growth in total water demands by Shlklamanov is given in
Gleich (1993), which shows that t.h.e rate of growth in demand is consistently
higher than population growth rate and that the contribution of storage to the total
supply is greater still, as shown in Table 2.3.
From the rates of growth for population, water consumption, irrigation area
and hydropower, the following growth rates for demand of storage are postulated,
and are shown in Table 1.1.
The forecast future demand for storage is shown in Table 1.2.

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
Table 1.2.

Geographical demand/or nevv storage


Demand for new storage: km3

Region

49

51

54

South and Central America

467

495

424

Africa

167

203

248

EVACUATION OF SEDIMENTS

Table 1.3.

Demand for new storage, South and Central America

there remains a large undeveloped potential resource in Canada, environmental


pressures will probably preclude any further development.
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.

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.

Africa.

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.

Asia and Oceania.

New reservoirs under construction


Worldwide. The historic rate of construction of storage worldwide is shown in
Table 2.13 and Figure 2.3. The overall growth rate for the century as a whole has
been 65%. It is interesting to note that neither of the World Wars nor the
Depression made any serious impact on the rate of growth: in this context the
Table 1.4.
Period
2000-2010

')0')0..')010
~

Demand for new storage, Africa


Annual growth rate: %
2-0

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.

Rate and distribution of loss of storage


In order to assess the variation in the rate of loss of 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 2.15.
The estimates of annual loss of storage owing to sedimentation have been used
in conjunction with the gross storage volume data available in the ICOLD World
Register of Dams 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

Rate of loss of storage


The rate of loss of storage for a given reservoir is dependent on the rate of erosion
of the catchment. In regions where the catchments have remained stable, e.g.
N orthem Europe and North America, the rate of loss of storage is constant. In
regions where deforestation has occurred the rate of catchment erosion and
consequently the rate of loss of storage increases.

Reservoir size and rate of loss of storage


The highest rates of loss of storage are found in the smallest reservoirs and the
lowest rates in the largest. Of the 1105 reservoirs studied, 730 have a storage
volume of less than 1233 M.m3 and an average rate of loss of storage in excess
of 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 016%
per annum.

Requirement for new storage


New storage will be required in the future both to satisfy increasing demarld
generated by the growing world population and to replace the storage lost owing
the next decade
to sedimentation. The estimate of the gross storage required
is shown in Table 1.5.

1.2.2. Research into factors which influence sediment flushing


For effective flushing the following factors need to be considered and satisfied.

Hydraulic conditions required for efficient flushing


Riverine conditions must be created in the resel'loir for a significant length of
time. The reservoir level must be held low throughout the flushing period,
possibly with minor fluctuations in level to activate sediment movement. To
achieve this:

EVACUATION OF SED1MENTS

Table 1.5.

Gross storage requirements to 2010


Storage volume: km3

Continent
Gross requirement

New demand
2000-2010

Loss to sedimentation
up to 2000

49

54

103

North America

112

112

South America

467

17

484

Aftica

167

35

202

Asia and Oceania

315

349

664

998

567

1565

Europe
I.

Total

I
i

the hydraulic capacity of the bypass must be sufficient to maintain the


reservoir at a constant level during the flushing period
flushing discharges of at least twice the mean annual flow are required
flushing volumes of at least 10% of the mean annual run-off should be
anticipated.

Quantity of water available for flushing


There must be enough water available to transport the required volume of
sediment. This has the following implications.
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.

Mobility of reservoir sediments


The nature and quantity of river sediments are important factors in determining
whether the quantity of water available for flushing is adequate to remove the
desired quantity of sediment from the reservoir.
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, 0001 and 0002. In
smaller rivers the equivalent range nlay be 0002 to 0005.

EXECUTIVE SUMMARY

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

fII

II

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.
The longitudinal slope available for the flushing channel is relatively small,
thus limiting the amount of sediment transport.
Reservoir volumes in the lower reaches are likely to be larger compared with
.
run-off and hence water availability becomes a restraint on
the mean ap. .11ual
sediment flushing.
.-

1.2.3. \j\/orldwide experience

of sediment flushing

The findings from the review of worldwide experience of flushing can be


sum..rnarised as fonows.

The hydrology and sedimentology of the catchment


The hydrology and sedimentology of the catchment need to be understood fully
in the planning of flushing facilities for new or existing reservoirs and to provide
the background for analyses of past sedimentation and flushing perfonnance.

The storage capacity of the reservoir


Successful hydraulic flushing is more likely to be practicable in reservoirs which
are small hydrologically, with a storage capacity less than 30% of the mean
annual inflow. The smaller the reservoir, the greater the chance of it being
successfully flushed and the
the likely residual storage capacity.
The sediment deposition potential
Flushing is vital for the preservation of long-term storage in reservoirs
the
sediment deposition potential is greater than 1 to 2% of the original capacity.
Even in
reservoirs with a potentially long life, consideration should be
given to possible eventual decommissioning problems when deciding whether or
not to flush.

EVACUATION OF SEDIMENTS

The shape of the reservoir basin


The shape of the reservoir basin can have a large impact on the practicability of
effective flushing and the residual storage capacity. Narrow steep-sided
reservoirs in valleys with a steep longitudinal slope are the easiest to flush. Wide
valleys, where the impoundment covers fonner floodplains, can be flushed less
effectively, because the deposits tend to consolidate and are remote from the
flushing channel.
The low-level outlet facilities provided
For effective empty flushing with full draw down , the low-level outlets must be
both low enough and of sufficient capacity to allow the drawdown to be
controlled during the'time of year when flushing is undertaken. Proportionately
- larger outlets are required for flood-season flushing than for flushing outside the
.
flood season.
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 longterm preservation of an important resource.
The deployment of full or partial drawdown
Full drawdown and empty flushing have been found to be much more effective
than partial drawdown~
The scope for enhancements to flushing
Fluctuations in water level and discharge during flushing are beneficial to the
promotion of bank slumping and increasing the rate of sediment discharge. Also,
the deployment of lateral and longitudinal diversion channels has been successful
in promoting flushing in reservoirs which are large hydrologically or contain
significant proportions of deposition in areas remote from the main flushing
channel.
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.
Value of sediment flushing
The degree of success in flUShing should be judged by whether it makes a
worthwhile difference to the beneficial uses of the reservoir, rather than simply
by whether it meets numerate objectives, such as a long-term balance between
inflows and outflows, or the retention of a certain percentage of the original
storage volume.

10

!"

EXECUTIVE SUM MARY

1.2.4. Geographical areas suited to flushing


E.rosion rate
The erosion rate depends on a complex interaction of the following factors.
Climate: precipitation and run-off, temperature, wind speed and direction .
., Geotechnics: geology, volcanic and tectonic activity, soils.
e Topography: slope, catchment orientation, drainage basin area, drainage
density.
Vegetation.
Land use and human impact.
factors are discussed in Appendix A4.1. It is not easy to generalise
between areas of high and low erosion rates depending on their geographical
location.
Estimates of average global rates of denudation have ranged from 006 to
016 mm/yr (Morris and Fan, 1997). This is equivalent to estimates of between
15 and 20 x 109 trKTn?/yr (vValling and vVebb, 1996). Areas with sediment yield
over 1000 t/km2/yr are 88% of the total land area and account for 69% of the
total sediment load. Regions with less than 50 t/km2/yr account for about half of
the land area and 21 % of the sediment yield. Case studies of erosion rates are
presented in Appendix A4.2.

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.
Tropical wet: classification Af.
Tropical wet and dry: classification Aw, Am and BS.
e Tropical desert: classification BW.
e Mid-latitude wet: classification Cf and Df.
Mid-latitude winter dry: classification Cw and Dw.
Mid-latitude summer dry (Mediterranean Climate): classification Cs and Ds.
Polar wet and dry: classification ET.
Polar desert: classification EF.

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.
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 those sites where conditions are ideal to those sites which are quite
unsuited to sediment flushing.
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:
.
parts of Central America extending into South America
areas in North and South America where the rivers are fed by the Rockies and
the Andes
parts of Central Africa from the Ivory Coast in the west to Sudan in the east

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.

1.2.5. Site-specific investigations and design considerations


There are many detailed factors which need to be evaluated on a site-specific
basis before the technical viability and economic soundness of sediment flushing
can be confirmed. Chapter 6 provides details of the nature of these site-specific
investigations, including design considerations for the sediment bypass itself.
There are numerous stages for such investigations, as follows.

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

of looking at the effectiveness of various aspects which affect reservoir


sustainability,over periods of up to 50 or 60 years, including:
measures to reduce the amount of sediments entering reservoirs, such as
catchment conservation or upstream storage
measures to manage the sedimentation process within reservoirs, such as
variations in the operating rule curves for the reservoir
measures to evacuate sediment from the reservoir, including dredging and
sediment flushing.

System simulation modelling


System simulation modelling is required to evaluate the conflicting demands of
hydropower production, irrigation and other requirements, and must be able to
assess the impacts of the various reservoir operating strategies.' The simulation
model must be able to replicate the outputs of water and power under a range of
operating strategies so that an optimal economic and technical solution may be
identified. In addition, it nlust be possible to take account of the effects of other
reservoirs upstream and downstream of the one under consideration.
Economic and financial analysis
The main aim of economic and financial analyses is to assist in the identification
and selection of the most favourable sediment management option. For each
option the most important factor, from the economic viewpoint, is to define the
'with' and 'without' project cases. These-will illustrate the net economic impact
of the availability of water resources over time, including any seasonal variations.
Evaluation of the impact of alternative investment phasing is also important.

14

atio

er\f Irs

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.
'

2.2.

WORLD TOTAL RESERVOIR STORAGE

2.2. I. ICOLD World register of dams


The most recent ICOLD World register of dams was published in 1998 and was
compiled from data collected from member, and some non-member, states in
1996. ICOLD required, in their circular instruction for reporting dam data,that
respondents should include all dams with a height greater than 15 m and dams
between 5 ill and 15 m in height with a storage of 3 M.m3 . or ' more. The
introduction to the register qualifies the data as follows.
s Japan reported only dams greater than 30 m high.
= Russia reported mainly hydropower dams.
'

.c_

""

_"

'' . '''., . .
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

in the 15 m to 30 m range. Therefore, it would seem reasonable to add a 20%


allowance for under-reporting. Postulating an average storage volume of
10 M.m3 per dam, this will increase the total storage by 50 km3
More difficult is the assessment of dams in the range 5 m to 15 m with a
storage of less than 3 M.m3 and all dams less than 5 m high. The 1995 National
Inventory of Dams maintained by the US Army Corps of Engineers lists 74053
dams over 2 m high with at least 60 000 m3 capacity, compared with the ICOLD
record of 6375 dams. The storage contributed by the 67 678 small dams not
included in the ICOLD register are estimated at some 12 km3 , some 5% of the
total. If the US data can be taken as typical for other countries, an allowance of
300 km3 should be made for storage provided by small dams. Thus the total
storage could be assessed at 6815 km3

2.2.2. Other sources


The estimated total capacity of the world's reservoirs is given in Water in crisis
(Gleich, 1993) as 7000 km3 ,
lJNESCO estimated in 1974 that the total storage of all reservoirs with
capacities of 5 km3 and above to be 4050 km3 This estimate was used by
Mahmood (1987), who assumed an allowance of 20% for the storage provided by
reservoirs less than 5 km3, to estimate total reservoir storage at 4880 km3 in 1987,
when the total nurnber of registered dams was approximately 20 000. Increasing
the storage pro-rata with the increase in number of dams gives a present day
storage of 6345 km3

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 ,

2.3.

WORLDWIDE DISTRIBUTION OF EXISTING STORAGE

2.3.1. Global water resources


The total world fresh-water resources are estimated at 35 million km? (Morris
and Fan, 1997). Of this, approximately 70% is locked up in the polar icecaps,
glaciers and permafrost, and approximately 30% is stored as groundwater. The
available water in lakes, rivers and swamps only accounts for 0,30% of the global
fresh-water resources, Natural lakes are estimated to contain 91 000 km3 , while
manmade lakes and reservoirs contribute 7000 km3 The water stored in natural
and manmade lakes and reservoirs is equivalent to 820/0 of the global annual
precipitation of 119 000 km3 and is twice the global annual run-off of
47000 km3 ,

18

SEDIMENTATION IN RESERVOIRS

2.3.2. Geographical distribution


The worldwide disuibution of existing storage and storage under construction, asdetermined from the ICOLD register is shown in Table 2.1. The Alnerica's
together with Northern Europe and mainland China account for 70% of the
existing world stock of reservoir storage.

2.4.

WORLD DEMAND FOR MORE STORAGE

2.4.1. Population
The world's population in 1990 is estimated to have been 5286 million, growing
at an annual rate of 15%. 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
Table 2.1.

Distribution of reservoir storage yolume

Region

North America

Number
of dams

1498

South America
Northern Europe

7205

2277

I
I
!

1845
1039
938

Average size of
reservoir:

Fraction of
world total

Gross
storage:
km3

M.m3

I
I

29%

256

16%

694
412

15%
I

3220

145

2%

45

Sub-Saharan Africa

966

575

9%

595

North Africa

280

188

3%

652

China

1851

649

10%

351

Southern Asia

4131

319

5%

77

44

148

2%

3364

277

117

2%

424

Pacific Rim

2778

277

4%

100

Middle East

895

224-

3%

250

25422

6464

100%

254

Southern Europe

Central Asia
South-East Asia-

World total

19

EVACUATION OF SEDIMENTS

Table 2.2.

Growth in world population

Year

Population:
millions

Annual growth rate in


following' decade: %

1990

5286

153

2000

6158

134

2010

7032

115

2020

7887

2030

8671

072

2040

9318

054

2050

9833

095

demand in much of the developing world is constrained by lack of availability


and is lower than that in the developed world: the total growth in demand is a
combination of population growth and per capita growth. Much of this demand
will be satisfied by increased surface and groundwater abstraction and water reuse. No direct linkage between overall demand and water storage can be
assumed.
An estimate of the growth in total water demands by Shiklamanaov is given in
Gleich (1993), which shows that the rate of growth in demand is consistently
higher than population growth rate and that the contribution of storage to the total
supply is greater still, as shown in Table 2.3.
Nearly 70% of the world demand for water is for irrigation. The bulk of the
world's storage is for irrigation and hydropower purposes, or a combination of
12000

18

(J)

8000

'E

c0

~
'3

0..

a..

6000

-... ---------- -Population

10000

--- -- --

......

------ .....

-...----- -- ---

..... ------

--

16
14

..c:

08

4000

06
04

2000
02

__~--~--------~----~O

OL-------~------~--------~
1990
2000
2010
2020

Year

Figure 2.1.

20

Growth in world population

*'

12 1Y

2030

2040

2050

0)

Cii
::I
c
c

<:

,I

SEDIMENTATION IN RESERVOIRS

Table 2.3.

Prediction of global demand (after Shiklamanov)


Rate of growth of demand: %

Year

Total

.,

Reservoirs

1950-1960

388

13-47

1960-1970

267

1112

1970-1980

251

616

1980-1990

221

354

1990-2000

231

261

,1 ~

~
!

.,

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
Table 2.4.

Growth in irrigation area

Year

Irrigated area:
million ha

Annual growth rate in


following 5 years: %

1975

190

222

1980

212

102

1985

223

123

1990

237

187

1995

260

21

EVACUATION OF SEDIMENTS

Table 2.5.

Growth in energy generated by hydropower

1990

2 112000

1995

2474000

321

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.

Table 2.6.

Comparison of actual and economical potential energy


Average annual
production:
GWh/yr

Region

Percentage
developed

Europe

525000

800000

66

North America

627000

1000000

63

South and Central America

491000

2325000

21

67 000

1000 000

750 000

3700000

20

2440000

8825 000

28

Africa
Asia and Oceania
Total

22

Economically feasible potential:


GWh/yr

..

SEDIMENTATION IN RESERVOIRS

5
-- - - -_ Water demand

- - - - -

Hydro demand

- ' - ' - Irrigation demand

--- ----- --- ----- ---- ..... ....

- - - Population
-

--_._-- .... __ .. _---------._------_ .. .


...

1
1970

1975

1980

Storage

---

.--.---.... ..

1985

1990

1995

2000

2005

2010

2015

2020

Year

Figure 2.2.

Comparison of growth rates

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.

DISTRiBUTION OF DEMAND FOR MORE STORAGE

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
Table 2.7.
Period

Annual growth rates and irlcrease in storage


Annual grovlth
rate per annum:
%

Increase in net storage:


km3

2000-2010

156

998

2010-2020

139 .

1032

?Cnn_"il~il

23

EVACUATION OF SEDIMENTS

Table 2.B.

European growth in irrigation and hydropower

Year

Irrigation
area:
million ha

1975

273

33

1980

321

1985

Energy:
GWh/yr

Annual growth in
following 5 years:
%

535 000

403

21

652 000

140

357

00

699 000

-032

1990

357

-68

688 000

139

1995

251

Annual growth in
following 5 years:
%

737 000

these population changes on water demand will be negligible, although changes


in the regional distribution of population will result in regional shortages and
surpluses which may generate demand for reservoir storage.

Irrigation and hydropower


The historic growth in irrigation and hydropower in Europe is shown in
Table 2.8.
The negative irrigation growth since 1990 reflects the abandonment of
uneconomic or marginal irrigation schemes in Russia since the break up of the
USSR. Although the connection between irrigation area and storage is far from
direct, these data suggest that there is no demand for new storage for irrigation.
The low growth in energy generation in the 1980s probably indicates water
shortages.
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 05% per annum.

2.5.2. North America


Population
The population of North America is predicted to grow from 278 million in 1990
to 306 million in the year 2000, rising to 358 million in 2020 and 389 million in
2050
with the growth rate declining from nearly 1% at the present time to
0-250/0 by 2050. As with Europe, the overall effects of these population changes
on water demand will be negligible, although changes in the regional disttibution

24

~.

.~

i
~<;

SEDIMENTATION IN RESERVOIRS

Table 2.9.

North American growth in irrigation and hydropower

Year

lITigation
area:
million ha

Annual growth in
following 5 years:

180

28

520 000

10

207

-19

547 000

22

1985

188

02

611 000

00

1990

190

13

614000

24

1995

203

1975
1980

Energy:
GWhlyr

Annual growth in
following 5 years:
%

693 000

of population will result In regional shortages and surpluses which may generate
demand for reservoir storage.

Irrigation and hydropower


The historic growth in irrigation and hydropower in North America is shown in
Table 2.9.
The overall growth rate of irrigation is 06% per annum. Although the
conIlection between irrigation area and storage is far from direct, these data
suggest that there is little demand for new storage for irrigation. Growth in
energy generation has been stronger, at an average rate of 14%.

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.

2.5.3. South and Central America


Population
The population of South and Central Amelica is predicted to grow from 440
million in 1990 to 523 million in the year 2000, rising to 676 million in 2020 and
839 million in 2050 - with the growth rate declining from 174% at the present
time to 071 % by 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.

Irrigation and hydropower


The historic growth in irrigation and hydropower in South America is shown in
Table 2.10.

25

EVACUATION OF SEDIMENTS

Table 2.10.

South American growth in irrigation and hydropower

Year

In'igation
area:
million ha

Annual growth in
following 5 years:
%

Energy:
GWh/yr

1975

117

08

117 000

Annual growth in
following 5 years:
%
112
I

1980

122

1985

14.. 8

1990
1995

167
198

39

199 000

72

24

282 000

53

35

365 000

55

477 000
I

The overall growth rate of irrigation is 267%. Although the connection


between irrigation area and storage is far from direct, these data suggest that
there is a steady demand for new storage for irrigation. The overall growth rate
of hydropower has been 72%
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 277% at the present time to 153% by
Table 2.11.

26

Annual growth rates, South and Central America

SEDIMENTATION IN RESERVOIRS

Table 2.12.
Year

African growth in irrigation and hydropower


In'igation
area:
million ha

Annual growth in
foUowing 5 years:

1975

94

12

1980

100

14

1985

107

13

1990

114

15

1995

123

Energy:
GWh/yr

Annual growth in
following 5 years:
%

37000

10-9

62 000

45000

I
I

-09
5-g

43 000
57 000

-62

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.
Irrigation and hydropower
The historic growth in irrigation and hydropower in Africa is shown in
Table 2.12.
The annual growth rate has been remarkably constant over this 20-year period
at an average rate of 135 % per annum - about half the population growth rate.;
The hydropower data reflect more upon the extremity of the 1980s drought than
upon the growth of new hydropower capacity.

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.

2.5.5. Asia and Oceania


Population
The population of Asia together with Oceania is predicted to grow from 3213
million in 1990 to 3785 million in the year 2000, rising to 4784 million in 2020
and 5787 million in 2050 - with the growth rate declining from 165% at the
present time to 06%- by 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.
Irrigation and hydiOpOWer
The historic growth in irrigation and hydropower In Asia is shown In
Table 2.13.

27

EVACUATION OF SEDIMENTS

Table 2.13.
Year

Asian and Oceanian growth in irrigation and hydropower


Inigation
area:
million ha

Annual growth in
following 5 years:

1975

1232

19

1980

1350

1985

I
I

Energy:
GWh/yr

I
I

Annual growth in
following 5 years:

235000

49

12

291000

37

1432

15

342000

52

1990

1555

34

437000

32

1995

1820

511 000

The annual growth rate has been reasonably constant over this 20 year period
at an average rate of 197% per annum - well above the population growth rate.
The average growth rate of hydropower over this 20-year pedod is 425%, 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.

RATE AND DISTRIBUTION OF CONSTRUCTION OF NEW RESERVOIRS

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 65%. 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

Table 2.14.

Annual growth rates, Asia and Oceania

Peliod

Annual growth rate:

2000-2010

20

2010-2020

15

2020-2030

10

Table 2.15.

Regional de117.andfor new storage


Demand for new storage (kIn3)

Reglon
.
2000-2010

2010-2020

I
I

2020-2030

Europe
South and Central America

Africa
Asia and Oceania

Year

1900

51

467

495

167

203

248

281

213

1032

939

315

998

Total

Table 2.16.

49

54

424

Rate of construction of new storage


Cumulative
storage:
]crn 3

I
I

11

Annual rate of increase


in following decade:

67

1910

21

1920

63

1930

121

1940

252

1950

116
6-7

I
I

76
51

414

I
I

112

1960

1196

.\

98

1970

3035

45

1980

4708

1990

5581

2000

5976

I
I

17
0-7

I
I

29

EVACUATION OF SEDIMENTS

6000

12

A
I
5000
CO)

E
.::<!.
CD

4000

E
"0
CD
Ol

I '"

::::l

>

............ ,
"'............"

3000

I
,

.e
til

ta

;e

2000

I
'J

"

\
\
\

\
\

- - - World gross storage


- - - - -

1.

10

Annual growth rate

1000

,,
,,
,,
.........

.......... .....

2
,

oL-~~~~==~~~~==~--~----~--~__-L__~O
1900

191.0

1920

1940

1930

1950

1960

1970

1980

1990

2000

Year

Figure 2.3.

Historic growth in reservoir storage

2.6.2. Distribution

1
1

of construction of storage

The distribution of the growth of new storage is set out in Table 2.17.

2.6.3. Comparison

of storage construction with demand

A comparison of the growth of storage constructed in the 1990s and planned


from 2000 onwards shows that in South America and in Africa actual
construction falls considerably short of demand. In the case of planned
construction in the decade 2000-2010 this may be explained by inadequacies in
the data.

I
1

Table 2.17.

Distribution of storage increase

,1

Annual growth rate in following decade: %


ear

Europe

North America

South America

Africa

Asia

1950

005

939

617

540

1084

1960

005

739

11,87

3307

1361

1970

0-05

369

824

585

1980

0-05

099

562

072

213

1990

0-05

0-03

159

049

116

2000

30

476

1
1

1
1

SEDIMENTATION IN RESERVOIRS

2.7.

RATE AND DISTRIBUTION OFLOSS OF STORAGE

2.7.1. Rate

of /OS5 of storage

The world estimate of land denudation is approximately 65 mm per 1000 years


(Walling, 1984). Of this, approximately 90% is transported as suspended
sediments and bed load, the remainder is in the form of dissolved matter.
The rate of loss of storage due to siltation varies around the globe. Extreme
examples include:
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 (01 % 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).
Table 2.18.

Regional sedimentation rates

R~gion

Estimated
annual loss of
storage due to
sedimentation:

Estimated
reservoir
balf-life:
yrs

North America

020

250

South America

010

500

Northern Europe

020

250

Southern Europe

017

294

Sub-Saharan Africa

023

217

Northern Africa

008

625

China

230

22

Southern Asia

052

96

Central Asia

100

50

South-East Asia

030

167

Pacific rim

027

185

Middle East

150

~"I

~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
lost to sedimentation, over 2000 average-sized dams would have to be
constructed around the world.

2.7.'2. Distribution of loss of storage


The distribution of the annual rate of loss of storage and the total volume lost to
sedimentation are shown in Table 2.20.
Table 2.19.. Extent of sediment data
Region

Total storage
used in
estimating
regional loss:
M.m3

Gross storage
used in
esth'llating
regionalloss:
%

109980

60

1038913

3832

04

Northern Europe

938 168

3067

02

Southern Europe

145 162

24030

165

Sub-Saharan Africa

575352

252 168

438

Northern Africa

188473

181 760

964

China

649322

42804

66

Southern Asia

318602

92712

291

Central Asia

148032

Nil

N/A

South-East Asia

117371

Nil

N/A

Pacific rim

277 124

20192

73

Middle East

223683

9006

40

Global total

6464730

738552

114

-.

Current gross
storage in region:
M.m3

North America

1 844 530

,
i

South America

32

SEDIMENTATION IN RESERVOIRS

7000

Projection assuming no new dams

6000
5000
C'l

.::

Q.j

01

4000

~
m

3000

;0

~ ::;:f'-_- -'-,~.= =~ ; .; ; J.'- ':-" "L. .-: " -;'. .l:. . -_-.l-: ~ _- L_- -JL. -. _J-. _. . l. -_- -L. ._- L_.- -J
Sedimentation

.-.=..-=
..- ""_1.

1900

1970 1980 1990 2000 20iO 2020 2030 2040 2050

1910 1920 1930 1940 1950 1960

Decade

Figure 2.4.

Storage lost to sedimentation

Table 2.20.

Distribution of sediment rate~ and storage loss

Region

North America
South America

I
)

Gross yolunie
in 2000:

Annual
sedimentation:

km3

1845
973

I
I

Ian3

Lost to
sediments:
%

Total storage
loss:
km3

369

79

112

104

25

17

Northern Europe

822

188

68

Southern Europe

135

025

56

574

132

78

32

Northern Aflica

188

015

2-4

China

526

1493

458

Southern Asia

233

166

131

Central Asia

132

1-48

269

035

80

Sub-Saharan Africa

South-East Asia

117

230
31

29

Pacific rim

232

0 75

76

IS

Middle East

199

3 36

277

38

Giobal total

5976

3085

11 8

567

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 05%, compared with the estimates for Europe by
this study of 02% and 017% per annum.

2.8.

TRENDS IN THE RATE OF LOSS OF STORAGE


The rate of loss of storage for.a given reservoir is dependent on the rate of erosion
of the catchment. In regions where the catchments have remained stable, e.g.
Northern Europe and North America, the rate of loss of storage is constant. In
regions where deforestation has occurred, the rate of catchment erosion and
consequently the rate of loss of storage increases. This phenomenon is clearly
visible in a number of locations, such as the Aberdare Forest in Kenya where the
natural vegetation is gradually being lost and replaced by small holding
subsistence farms. This has resulted in an increase in the rates of erosion.
As part of this study, data from 16 reservoirs in the US have been analysed to
detect any evidence of change of loss of storage over the life of the reservoirs.
The reservoirs ranged in size from 175 000 m3 to 23 km3 , with catchments
ranging from 9 km2 to over 100 000 km2 With the exception of Fort Peck
(23 000 M.m3) and Great Falls (68 M.m3), reservoirs that showed a definite trend
of reducing sediment accumulation rates with time, no discernible trend could be
established.
The data from Ringlet reservoir in Malaysia (ICOLD, 1997) shows clearly the
dramatic effects of deforestation. The 183 km2 catchment has been gradually
changed from forests to plantations and holiday facilities, which has resulted in
the specific sedimentation increasing by an order of magnitude from the midsixties to the present day (see Table 2.21).

Table 2.21.
Year

34

Ringlet reservoir, sedimentation


Specific
sedimentation:
m 3/km2/yr

Annual sedimentation:
. M.m3/yr

1965

165

003

1980

275

0-05

SEDIMENTATION iN RESERVOIRS

40
35

::::l

c:
c:

...

CiS

30

Based on data from


1105 US reservoirs

<V

c..
~

25

1ri

CI

e:!
0

t;

20

'0
Co"
tfl

10S

.Q
(l)

CI

e
ID

10

05

"i"":

0.0 1
0001

ttlt'

0-01

, "

It I

01

10

100

1000

10000

Aveiage size of reservoir: M.m3

Figure 2.5.

2.9.

Reservoir size and rate of loss of storage

RESERVOIR SIZEAND RATE OF LOSS OF STORAGE


In the USA it is estimaJed that 1235 M.. m 3 of sediments are deposited in the
resen-oirs annually (Glymph, 1973). The total storage of reservoirs in the USA
is estimated at 627300 M.m 3 (Morris and Fan, 1997). This represents a loss of
O 20% of the storage volume annually to sediments. However, the rate of loss of
storage varies substantially and shows a striking inverse relationship between the
rate of loss of storage and reservoir capacity, as shown in Figure 2.5.
The highest rates of loss of storage are found in the smallest reservoirs and the
lowest rates in the largest. Of the 1105 reservoirs studied, 730 have a storage
volume of less than 1233 m 3 and an average rate of loss of storage in excess of
Table 2.22.

Gross requirement for ne"Y\/ storage

Continent

demand

2000-2010
Europe

49

South America

467

Africa

167

35

202

Asia and Oceania

315

349

664

Total

998

567

1565

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.

2.10.

REQUIREMENTS FOR NEW STORAGE

New storage will be required in the future both to satisfy increasing demand
generated by the growing world population and to replace the storage lost due to
sedimentation. The estimate of the gross storage required in the next decade is
given in Table 2.22.

36

earc

In

ue

fa tor
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 hydrological conditions at the site
" details of the flushing system
0._ the way in which the flushing system is operated
the development and re-erosion of the sediment delta
details of the development of the incised channel during flushing
sediment inputs, throughputs and outputs.
(9

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:
Fortunately, there have been major advances in our understanding of sediment
transport by water and this has facilitated the development of numerical
simulation models. It has also facilitated a better understanding of the strong and
weak points in physical model simulation techniques. These models have also

39

EVACUATION OF SEDIMENTS

had the benefit of the field data in order to check their validity. Simulation
techniques can thus be used to:
extend the range of information beyond that covered by field data
systematically look at the importance or sensitivity of individual variables.
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.

3.2.

. THE MECHANISM OF FLUSHIN.G


Shen (1999) quot~s numerous researchers, Lai and Shen (1996), Albertson et al.
(1996), Shen and Lai (1996) and Morris and Fan (1997), who have studied
sediment flushing. Based on this work, Shen quotes three stages of flushing.
(a) When the water level in the reservoir is high, the water velocity in the

reservoir is too low to move much sediment. Only close to the flushing outlet
are the flow velocities high enough to erode sediment, and a flushing cone is
formed close to the flushing outlet.
(b) At intermediate water levels, water velocities at the upper end of the reservoir
increase and sediment is transported towards the flushing outlet. There
remains a flushing cone close to the outlet.
(c) When the water level falls to the top of the flushing outlet, scouring velocities
can be generated throughout the length of the reservoir. Retrogressive erosion
of previously deposited sediment occurs.
Shen (1999) concludes that stage three removes far more of the deposited
sediments, thereby regaining storage capacity. He also comments that stage three
'uses more water'. Indeed, stage three is the only realistic scenario for removing
significant quantities of previously deposited sediments from reservoirs.
When flushing is attempted without drawing down water levels, the high flow
velocities at the outlets are very localised and the impact is insignificant. The
water level in a reservoir must be drawn down close to the bed elevation at the
dam before flushing can be effective (Figure 3.la). Many authors have confirmed
this with observation, theory or modelling, including Mahmood (1987), White
and Bettess (1984) and Atkinson (1996), However, moderate lowering of water
levels during flushing will still increase flow velocities significantly at the
upstream end of the reservoir, where bed levels will be above the water level at
the dam (Figure 3.lb). Large sediment volumes will be scoured from these
upstream reaches and will re-deposit nearer the dam. Eventually, bed levels
upstream from the dam will rise to the water level during flushing and then
significant sediment quantities will be transported through the low level outlets
(Figure 3.1c).
Thus, flushing represents an extreme change in reservoir operation. It requires
draw down of the reservoir so that the velocity and volume of the flow are

40

FACTORSWHICH !NFLUENCE FLUSHING

Maximum water level

----------

Original bed level

'.

(a)

Maximum water level

(c)

Figure 3.1. Longitudinal profiles during flushing: (a) flushing with full drawdol,w1.;
(b) flushing yvith insu:fficient drmvdown; (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
undertaken if it may jeopardise future power or irrigation supplies.
..... The environmental consequences of passing sediments, that may have been in
, the .reservoir for .s.ome considerable time, to the downstream reach.

-.,

- \

-Ij

-I
!

3.3.

THE DEVELOPMENT OF CRITERIA FOR SUCCESSFUL FLUSHING

Sediment flushing is not universally applicable. If the conditions are right,


flushing represents an efficient and economical way of preserving reservoir
storage. If the conditions are not right, attempts to flush sediments from the
reservoir will be disappointing. This section considers those factors which
determine whether flushing is likely to succeed.

J
J

3.3.1. Sediment balance


The sediment balance considers the quantities of sediment entering the reservoir
compared with the quantities which might be flushed through the dam. The
analysis makes some simplifying assumptions but is useful in deriving one of the
indicators which determines whether flushing operations are, or will be,
successfuL This indicator is the sediment balance ratio, SBR.

Long-term equilibrium conditions


If flushing water levels are close to bed elevations at the dam (either as in Figure
3.la or as in Figure 3.1c) and long-term equilibrium conditions are to be reached,
then the sediment mass flushed must, in the long-term, balance the sediment
mass depositing between flushing operations. This balance can be expressed as:
(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)
The transporting capacity Qs' (tis) will be a function of discharge, channel
roughness, width and slope, and the properties of the deposited material. A low
sediment balance ratio indicates low flushing efficiency and the continued build
up of deposited s~~iments in the reservoir. A value of SBR> 1-0 indicates high
efficiency and the lpng -term stability of sediments within the reservoir.

The relevant parameters


1. The rate at which sediment is flushed, Qs'
The only method for predicting Qs during reservoir flushing which has been
widely tested, is an empirical equation derived by Tsinghua University and
reported in IRTCES (1985) and 110rris and Fan (1997):

(3)

where Qf is flushing discharge (m3/s), S is the longitudinal energy gradient


through the reservoir, W is channel vvidth (m) and " is a constant related to the
sediment type:

1600 (530, Atkinson (1996)) for loess sediments


650 (225, Atkinson (1996)) for other sediments 'with median size finer than
O-lmm
300 (lao, Atkinson (1996)) for sediments with median size larger than 0-1
mm,and
III
180 (60, Atkinson (1996 for conditions of flushing with a low discharge.
IJ

Discrepancies between predictions of sediment load derived from Equation (3)


and observations were relatively small for the flushing data from China, on which
the method is based_ Discrepancies \vere within a range of half to twice in 87%
of cases, which is very good for sediment transport predictions. Atkinson (1996)
compared the method with observations from reservoirs in India, USA and the
former USSR. Equation (3) was found to overestimate sediment loads by a factor
of t.'ree and even more \vhere flushing was not performed annually (as is
common practice in China) but leSS frequently where sediments were much
greater than the 01 mrn threshold. The threefold correction suggested by
Atkinson is recoITunended when conditions differ from those typical in China
(see above).

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
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.
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):
_.0

(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.
Vl't = 128 0.0.5

1000

Sanmenxia
Guanting

@ Guernsey

Baira

EB

.s=

:0
.~

1i5
c:
c:
ctS

..c:
()

100

"'0

<l.'l

(/)

.0

10~--~~~~~~--~--~~~~~--~~~~~~----------

10

Flushing discharge,

Figure 3.2.

44

1000

100

0.: m3/s

Channel widths formed in reservoir deposits during flushing

FACTORSWHICH INFLUENCE FLUSHING

4. The discharge used for fIushing~ Qf.


The discharge used for flushing reservoir sediments is, in the case of maximum
drawdown, the incoming river flow because there is little storage in the .reservoir
to cause attenuation.
Any flushing system must be capable of passing these river flows while
maintaining low water levels at the dam.
5. The duration of the flushing operation, Tf

This is under the control of the reservoir operating authority. Factors to be


considered include:
the current availability of water
the need to safeguard future supplies of water
the need to miniwise the loss of present ~'1d future po\ver supplies.

6. The number .of years between flushing operations, N.


Again, this is under the control of the reservoir operating authority. Factors to
be considered include:
/II

Ii

the need for the operation in the ~ght of the success of previous years'
operations
hydrological forecasts for the follovving months
likely demands for power and irrigation water supplies for the following
months, taking into account alternative sources of supply.

7. The trapping efficiency of the reservoir, TE.


The trapping efficiency of a reservoir depends on many factors. Brune's (1953)
curves give good guidelines in general terms. However trapping efficiency
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.
1

Evaluation of the sediment balance ratio, SBR


Equations (2) and (3), together with an estimate of trapping efficiency taken from
Brune's (1953) curves, can be used to derive the sediment balance ratio, SBR.
For long-term equilibrium to be obtained, values of SBR comfortably in excess
of unity are required owing to uncertainty in the prediction methods and input
parameters.
Discharge
flushing, Qf' and the duration flushing,
will depend on
the reservoir operation chosen, and the
can initially be set at the river slope
before impoundment. Predicted values for SBR can be used to guide the choice

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

. Initial
. capacity:
M_m3

Country

Sediment
balance ratio

Long-term capacity ratio


(LTCR)

(SBR)

Estimated
from
reservoir
surveys

Calculated

Calculated

Reservoirs flushed successfully


1

Baira

India

9-6

085

0-85

Gebidem

Switzerland

9-0

Approx_
1-00

0-99

Gmund

Austria

0-93

0-86

098

21

Hengshan

China

13-3

075

0-77

Palagnedra

Switzerland

5-5

100

1-00

33

Santo Domingo

Venezuela

30

0-97

1-00

11

I
i

Reservoirs flushed unsuccessfully


Guanting

China

2270

Low

0-20

0-2

Guernsey

USA

91

Low

0-26

10

Heisonglin

China

86

023-0-35

030

Approx_
0-70

Ichari

India

116

Approx.
035

0-36

Ouchi-Kurgan

Former
USSR

56

Low

Approx.

China

9640

Sanmexia

010
031

0-39

3-4

46

Sefid-Rud

Iran

Shu1caozi

China

1760

<026

013

96

Low

039

46

FACTORSVVHICH INFLUENCE FLUSHING

3.3.2. Sustainable reservoir capacity


The flushing of sediments depends on many factors. Prior to the construction of
the reservoir, the entire annual flows were used to transport sediment through the
reach and water depths were the natural river depths throughout the period. Once
the reservoir is impounded, depositional areas are created and only a proportion
of the annual flow is available for flushing sediments. Thus, some pennanent
long-term deposition is inevitable in reservoirs. The question to be asked is: what
proportion of the original storage volume can be retained by flushing?

Long-term capacity ratio~ LTCR


The long-term capacity ratio, LTCR, is defined as the sustainable storage capacity
divided by the initial storage capacity of the reservoir.
Flushing will cause a channel to be scoured into the reservoir deposits. In most
cases~ this channel will be narrower than the reservoir and so substantial deposits
will remain in the reservoir. In the long term these deposits will rise to an
elevation close to the maximum water level, leaving the volume created by the
incised flushing channel as the only storage volume remaining in the reservoir.
This storage volume is defined as the sustainable reservoir capacity. Figure 3.3
illustrates the process, it shows cross sections at two reservoirs ~vhere flushing
has maintained a relatively small sustainable reservoir capacity.
If a trapezoidal cross-sectional shape is assumed for the incised channel, the
sustainable reservoir capacity volume is't.lJ.en determined from:
minimum bed elevations at each point in the longitudinal profile, as shown in
Figure 3.1 - these can be determined from the water level during flushing and
the sediment balance calculations described above
e the ma."{imum water level,
.. the bed width of the incised channel, which can be calculated using Equation
(3) and t..he flushing discharge (bed width Inay be constrained by reservoir
width)
I
the side slope steepness of the incised channel (if the side slope is shallower
L~an the reservoir side slope, then the reservoir widths may constrain the width
of the sustainable section).
e

In the list above only the side slope steepness is not known. In wellconsolidated sediments, near vertical channel sides can occur, while slopes as
low as 25% 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

1750

E
C

~I

>
Q)
Q5

~l

"C
Q)

a:l

1745

I
1

-I

1740~----------------~-------------------L--------------__~

100

200

300

(a)

-\

(Before flushing)

Q5
Q)

-i

Sept. 1973 - (after flushing)

320

~
~

"C

325

E
C

-I

-I

315

310

r)

~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
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.

----~.
: ----------------------------------~----------------------------

48

FACTORSWHICH INFLUENCE FLUSHING

(a)

(b)

~ River channel

.-/-

Dam

------

Section
Section

"'C7

""=7'

(e)

Full supply level


Original river-bed levels
(d)

""':;:-------or---------,-----""""7'

~\ B_~
__

I I

lA'
),t

Full supply lever

_ _ _ Water level during flushing

,-- - - - - - - - -

--It- .
"I"

Bed width,

Reservoir bed elevation at dam

WbO!

~ Flushing channel width, Lt'~


(e)
Long-term capacity ratio, LTCR,
is approximated to:

Area B
Area A plus Area 8

Figure 3.4. Simplified reservoir geOl7'letry for application of capacity criterion:


(a) actual reservoir plan; (b) fitted reservoir plan; (c) simplified reservoir plan and
sections; (d) simplified reservoir elevation; (e) enlarged section immediately upstream,
of dam

49

EVACUATION OF SEDIMENTS

If the long-term capacity ratio, LTCR, is greater than 05 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.

3.3.3. Evaluation

of flushing criteria

The sediment balance ratio, SBR, is a measure of the propol1ion of the incoming
sediments which may be flushed from the reservoir. The long-term capacity ratio,
LTCR, is a measure of the proportion of the initial storage capacity which may
be retained by flushing from the reservoir.
Some of the factors which determine the values of the sediment balance ratio,
SBR, and the long-term capacity ratio, LTCR, are inherent characteristics of the
site. These include:
the shape and size of the reservoir
the imposed hydrological conditions
the imposed sediment inputs.
Some of the factors are controllable. These include:
the operation of the reservoir between flushing operations
the design of the flushing system, including elevation and capacity
/I the operation of the flushing system, including discharge and duration.
A review of the literature on reservoir flushing, produced information from 14
reservoirs where flushing had been attempted and where sufficient data were
available to test the criteria (Atkinson, 1996). The 14 reservoirs can be divided
into two categories: six where observations indicated that flushing would sustain
a long-term capacity in excess of half the original capacity, and eight where it
would fail to do so. Table 3.1 presents the results of the application of the two
assessment criteria. The criteria performed very well in distinguishing between
the six reservoirs where flushing was successful and the eight where it was not.
The predicted LTCR also proved to be a good indicator of the long-term
capacities that were estimated from the observations. The sediment balance ratio,
SBR, was, in most cases, not a constraint to successful flushing.

3.3.4. Practical criteria for successful flushing


There are several factors which impose constraints on sediment flushing. These
include:
the operation of the reservoir between flushing operations as dictated by its
usage
the design of the flushing system, including elevation and capacity, as dictated
by economics and hydrological factors

50

r'"

FACTORSWH1CH INFLUENCE FLUSHING

It

the operation of the flushing system, including discharge and duration, as


dictated by external demands for energy and water during the flushing period.

It is useful to review the factors which constrain successful flushing at a


reservoir, and so assess whether they can be overcome, for example by enlarging
outlets in the dam.
Four main constraints are identified and these have been assessed against field
data for reservoirs that have been successfully and unsuccessfully flushed. These
constraints are presented below.

Incomplete drawdown of water levels during flushing


It may not be possible to flush sediments from outlets close to the initial bedlevel upstream of the dam. Existing installations may have been built at a higher
level. There may be serious engineering/economic problems in building such
outlets at projected new installat~ons. Under these circumstances, the capability
of removing sediment from the reservoir by flushing is reduced.
By taking water height as elevation above the base of the dam, a drawdown
ratio is expressed as:
(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 07 indicates some degree of
constraint owing to insufficient drawdown.

Insufficient flushing flows to develop a long-term sediment balance


Flushing flows can be constrained for many reasons:
., 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 < 10 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 00006 and a flushing water volume of 864 x 109 m3 It
is assumed that the reservoir contains 01 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
discharge:
m 3/s

Flushing
duration:
days

Mean sediment
concentration:
ppm

Sediment
removed per
day:
Mt

Sediment removed
during flushing
period:
Mt

500

200

9600/5760*

0411025* .

82/50*

1000

100

14550

126

126

1500

67

18560

241

160

2000

50

22060

381

191

2500

40

25220

545

218

3000

33

28130

729

243

3500

285

30860

933

267

4000

25

33430

1155

289

4500

222

35880

1395

5000

20

38220

1651

330

5500

182

40470

1923

350

6000

167

42640

2210

369

I
I
I

* Using Atkinson (1996), see Section 3.3.1(1).

52

310

FACTORSWHICH INFLUENCE FLUSHING

hydraulic efficiency of removing sediments from the reservoir, in terms of loss of


potential generating capacity and of volumes of water for irrigation.
Experience suggests that a flushing discharge of twice the mean annual flow
is desirable and that the volume of water used for flushing should be not less than
10% of the mean annual run-off.

Reservoirs that are too narrow to develop an efficient flushing channel


The reservoir could be too narrow for the natural width of the flushing channel
to develop during flushing. This is particularly the case where flushing discharges
are high. To check this, the bottom width of the flushing channel should be
compared with the bottom width of the reservoir.
A flushing width ratio, FWR, is defined as:

FvVR

=VV/W

(7)

bed

where W is computed from Equation (3) and V/bed is a representative bottom


width in the reservoir. If the reservoir can be approximated to the shape shown
in Figure 3.4~ the representative bottom width should be taken as that which
occurs just upstream of the dam. FWR> 10 is required unless the side slopes are
shallow (see below - TVVR greater than about 2 would indicate that F'YVR is not
a constraint).

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:
TVVR

=[Wmin + (2 X Hrnax X tan a)J/~Vtop

..

(8)

where vVmin is the lesser of Wand }Vbed and V/top is a representative top width in'
the reservoir. If the reservoir can be approximated to the shape shown in Figure
3.4, the representative top width should be taken as that which occurs just
upstream of the dam. TVVR> 10 is required.

The effect of the size

of sediment in the reservoir deposits

The nature of the sediments entering and depositing in reservoirs influences


whether flushing is practicable.
The width of the incised channel formed during flushing is determined by the
flushing discharge and it is independent of sediment size for the raD.ge of
sediment sizes generally found in reservoirs, see Equation (4). However, the
sediment tral1.Sport rates are dependent upon sediment size as indicated by the
constant, 'Ii, in Equation (3). Coarser materials are more difficult to remove from
reservoirs than finer materials.

53

EVACUATION OF SEDIMENTS

Table 3.3. The influence of sediment size on the amount of sediment removed from,
reservoirs
Sediment removed during the flushing period: Mt

Flushing conditions

Sediments with a
mean size less
than 01 mm

Sediments with a
mean size greater
than 0-1 mm

Conditions
with low
flushing flow

N/A

N/A

N/A

50

1750

740

330

N/A

m3/s

Duration:
days

Loess
sediments

500

200
20

Flow:

5000

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.
The effect of widely graded sediments
In many rivers there is a mixture of sand and gravel in the bed material. In these
circumstances, there may be a fairly small proportion of the bed material
consisting of fine sand, but owing to its high transportability, a large proportion
of the material being transported by the river is fine sand. When an impoundment
is introduced to such a river, the material depositing is dominated by fine sand
together with finer cohesive material that is transported in the river as wash load.
These deposits are relatively mobile compared with the general river-bed
sediments and hence are amenable to flushing. A parameter that can identify
these favourable circumstances for achieving a sediment balance by flushing is
the sediment size ratio, SSR:
(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 riverbed 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.
Table 3.4.

Effect of the sediment size ratio on sediment balance


Flushing period required: days

c. -ment Size ratIO


Sill
(SSR)

Tarbela,
Pakistan

Tungabhadra,
India

Rooiport.
South Africa

(> 100)

(> 100)

(> 100)

04

62

94

(> 100)

0-2

38

58

84

0-1

23

30

58

0-06

15

Ii

21

46

0-05

13

19

43

0-04

11

17

40

15

36

0-03

9 .

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:
there is a single flushing period of 30 days annually
sediment inflow to the reservoir for all material (wash load and bed material
load) can be described by the simple relationship:
Concentration = Constant x Discharge1.2

Table 3.5. Effect of the proportion of sand and coarser material on extending the life
of reservoirs
PropOition
of coarse
sediment

Factor by which reservoir life is extended

Qllushin/ Qmean
I

02
04

Tarbela

Tungabhadra

19

14

25

18

Rooiport
1-4
18
!

I
I

06

38

27

27

08

76

5-4

55

09

152

108

110

02

27

16

16

04

36

21

21

06

54

32

31

109

63

62

217

127

125

08

09

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)
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 02 to 09 in each
case.
These results indicate that where a large proportion of the material deposits in
the delta, say P sand >O8, 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 .
., The sediment sizes in transport in the river can be of paramount importance to
the success of flushing in a reservoir.
From the point of view of achieving a sediment balance, a large factor is
required benveen the sediment sizes being transported in the river and the sizes
found in the river-bed material. Such conditions are typical for gravel rivers
with a widely-varying bed material composition.
If a sediment balance can be assured, then a predominance of fine sand, and
other material that deposits in the delta at the head of a reservoir, ensures that
flushing greatly extends reservoir life.
$
Therefore, from the point of view of sediment size alone, delta deposits of fine
sand and coarse silt are the most likely to produce success in flushing a
reservoir. Coarser material may inhibit a sediment balance arid finer material
will deposit in the body of the reservoir outside any incised channel and so will
not be available for reworking during flushing.

Evaluation

of criteria at existing reservoirs

Application of these criteria to the 14 reservoirs presented above is given in


Table 3.6. Unfortunately, there is insufficient data readily available to include the
sediment size ratio, SSR, and the proportion of sand and coarser sizes, Psand ' in
this table.
When compared with field data, the criteria are able to distinguish reasonably
well between the reservoirs where flushing is successful and those where it is not:
almost all the criteria were met for the six successfully flushed reservoirs (figures
in bold) and at least one criterion was not met for each of the eight other
reservoirs.

57

EVACUATION OF SEDIMENTS

Table 3.6.

Application of constraint criteria to existing reservoirs


DDR value

SBRd value

FWR value

TWR value

India

068

24

3-4

16

Gebidem

Switzerland

093

20

67

15

Gmtind

Austria

089

58

5-2

13

China

077

Approx.4

0,1

71

Palagnedra

Switzerland

100

33

14

10

Santo Domingo

Venezuela

1-00

11

14

18

004

05

Reservoir

Country

Reservoirs flushed successfully


Baira

Hengshan

Reservoirs flushed unsuccessfully

Guanting

China

081

03

Guernsey

USA

044

3-2

14

026

Helsonglin

China

0'77

Approx.1

006

08

Ichari

India

031

33

99

14

Ouchi-Kurgan

Former USSR

014

110

Approx.2

Approx.03

Sanmenxia

China

075

48

026

09

Sefid-Rud

Iran

0'96

43

03

0-1

Shulcaozi

China

037

15

10

2-1

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.

3.4.

SUMMARY OFTHE REQUIREMENTS FOR EFFECTIVE FLUSHING

For effective flushing the following factors need to be considered/satisfied.

3.4. I. Hydraulic conditions required for efficient flushing


Riverine conditions must be created in the reservoir for a significant length of
time. Flushing is most effective when the reservoir is fully drawn down to a level
approaching the conditions which applied prior to impoundment. The reservoir

58

FACTORSWHICH INFLUENCE FLUSHING

level must be held constant at as Iowa level as possible throughout the flushing
period. To achieve this:
ED

the hydraulic capacity of the bypass must be sufficient to mairitain the


reservoir at a low level during the flushing period.

3.4.2. Quantity of water available for flushing


There must be enough water available to transport the required volulue of
sediment. For a given quantity of water used for flushing, it is more efficient
hydraulically to use a high discharge for a short period than to use a low
discharge for an extended period. This has the following implications.
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.

3.4.3. Mobility of reservoir sedjments


The nature and quantity of river sediments are important factors in determining
vvhether the quantity of water available for flushing is adequate to remove the
desired quantity of sediment from the reservoir.
(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, 0001 and 0002. In
smaller rivers the eauivalent
range
may
between 0002 to 0005.
1.
...

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
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.

3.4.4. Site-specific factors


The most suitable conditions for flushing are to be found in reservoirs which
approximate in shape to the incised channel which develops during flushing. If
the reservoir is too narrow, the incised flushing channel cannot develop its full
equilibrium width. If the reservoir is too wide, large areas of sediments will
remain on the flanks of the incised channel.

J
J

L()ng, relatively nan"ow, reservoirs are better suited to flushing than short,
wide, shallow reservoirs.

3.4.5. Constraints on the ultimate capacity achievable by sediment flushing


In cases where the amount of water available for flushing, combined with
considerations of the shape of the reservoir, and the nature of the sediments
within the reservoir, lead to a restriction on the amount of sediment which can be
removed, the long-term capacity ratio should be assessed in order to check what
percentage of the original reservoir capacity can be retained long term by
sediment flushing.

The greater the sustainable live storage the more attractive flushing systems
become. Subject to the economic circumstances, flushing systems will
normally be worthwhile if a sustainable live storage of more than, say, 35% of
the original live storage can be achieved by flushing.

3.4.6. Economic assessment

A full economic analysis covering the whole life costs and benefits of the
flushing system should be undertaken.
.

3.4.7. Summary
Reservoirs in the 'upper and middle reaches' of rivers are likely to be best suited
to sediment flushing for the following reasons.

60

To be worth doing, the benefits of sediment flushing measured over the


anticipated lifespan of the works, must exceed the penalties of loss of power
during the draw down period and possible loss of stored water for irrigation and
other uses.

FACTORSWHICH INFLUENCE FLUSHING

3.5.

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.
The longitudinal slope available for the flushing channel is relatively small,
thus limiting the amount of sediment transport.
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.

NUMERICAL MODELS
The previous section described the factors which influence the efficiency of
. sediment fiuslling and gave guidance on some of the hydrological and design
parameters which need to be satisfied.
Detailed analysis of specific sites requires the use of numerical models that can
provide much firmer estimates of flushing performance. Numerical models can
take into account many details that are precluded from the simpler desk
calculation techniques. These include:
e
4D

details of the reservoir topography


'details of the long-tenn development .of sedimentation using representative
flow sequences
details of the annual/monthly/daily operational n11es for the reservoir, in terms
of required releases, rule curves for water levels and maximum rates of change
of water levels, etc.
details of the sediments in motion, including graded sediments where these are
a factor of importance.

The calculation methods described in Section 3.4 rely on several assumptions


and can only provide an approximate estimate for the design and operation of
reservoir flushing systems. A more accurate method is numerical modelling,
albeit with a requirement for much more input data.
A one-dimensional model is often suited to the simulation of reservoir
sedimentation. More complex two-dimensional or th..ree-dimensional models
will, in general, require too much data and computational time because
simulations are usually required to cover periods of 50 t() 100 years into the
future and have a time step of aday or an evep shorter ;p~rlod.
For each time step in a 'cme-dimensional'1inodel, th---\vater levels and flow :. ",.,:'
conditions are predicted from discharges and/or changes in storage, 8.J.1.d hence
sediment concentrations within each of, typically, 10 size fractions are routed
through the reservoir. Bed level changes are determined, using the concept of
sediment continuity, from the cbanges in concentrations through the reservoir.
These changes in bed level are used to update the bed elevations stored in the
model. Usuallybed elevations are stored as full cross sections rather than single

61

EVACUATION OF SEDIMENTS

values, so additional rules are required to determine how deposition or erosion is


distributed across the sections.
White and Bettess (1984) and Basson and Olesen (1997), as well as other
authors, present one-dimensional numerical model applications. The White and
Bettess (1984) model has now been combined with reservoir survey analysis
software, which uses the accurate Stage-Width Modification Method (SWIMM),
to form the PC software RESSASS (REServoir Survey Analysis and Sedimentation Simulation).
Two- or three-dimensional models can be used to assess the localised impact
of flushing near low-level outlets. Atkinson (1996) briefly presents threedimensional modelling in an idealised reservoir to investigate the extent of
influence achieved by flushing. Such modelling would usually provide little
useful information on the feasibility of flushing, but may prove invaluable as a
component of the design process.

J
1
1

1
1
1

1
1

1
1
J~

I
62

1
1
I

I WI
e Imen
e

fit

len e
G

..

flu hing

4. Worldwide experience
4. J.

of sediment flushing

INTRODUCTION
The main purpose of this chapter is to answer the following questions:
how many reservoirs are being flushed?
where are they?
are they used for water supply (potable and irrigation) hydropower or flood
detention?
., what methods are employed for flushing (the facilities and the operational
regime)?
Q
what is the nature of the sediment?
how successful is the flushing?
iii what constraints (operational, economic and environmental) affect the flushing
sy'stems?
/I what downstream effects occur?
e

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
-

.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.

4.2.

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
Chapter 3. A few of the main points are given below, as these help to explain
some of the experiences of flushing performance.

FLUSHING

Flushing is a technique which, by using a suitable combination of the drawdown


(water level lowering) and increased flow in the reservoir, allows previously
deposited sediments to . be discharged from the reservoir basin into the
..downstream ri~er or irrigation system. Flushing is undertaken over a relatively
short period - usually a few days or weeks and would typically be annual,
although there are some cases where it is undertaken once every few years.
Flushing may be undertaken with the reservoir effectively empty ('empty
flushing'), so that riverine conditions are established, or with the reservoir
paItially drawn down (,pressure flushing'). It may be undertaken either during
the flood season, as is most common, or outside it.
Flushing can be distinguished from sediment 'routing' techniques, which aim
to pass the bulk of the sediment load without deposition in the reservoir.
Examples of these techniques are:

J
r

J
J

'sluicing' by drawdown through the flood season


'sluicing' by drawdown during the main annual floods
density current venting.
Routing - particularly sluicing - results in the seasonal pattern of sediment
outflows largely following the pattern of sediment inflows, whereas flushing
typically compresses the annual sediment load, which may occur over two or
three months, into a few days or weeks.
Inevitably, there is a potential overlap between the techniques, such as in cases
where a significant propoltion of the annual sediment load is passed without
deposition, but where flushing is relied upon to erode those sediments deposited
during the sluicing operations or during floods outside the sluicing period.

66

-{
J.

WORLDWIDE SEDlMENT FLUSHING

4.3.

WORLDWIDE EXPERIENCE OF 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
or parallel for flushing, only a single case is included in this
operated in
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:
available to those countries for
t.1.e resources - fin~ncial and technical
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.
G>

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.

4.3.2. Flushing techniques


Of the 54 cases included in Table 4.1, 50 were
to involve significant
while three were predominantly routing (sluicing) and one flood storage
r,:;;>cpr,J'f"'I1r was considered to be essentially uncontrolled, so that the mode would
Of the 50 fi ushing
a closer resemblance to sluicing, rather than
t-!l1C'hl1"\ ......

67

EVACUATION OF SEDIMENTS

Table 4.1.

Summary of reservoirs flushed


ii0

Reservoir/dam

Baira

India

Barenburg

Switzerland

Bajiazui

China

Cach!

' Former USSR

Chiyu
Dalingkou

I China

""0

::lE

3 .~
'" ,t:tl

J.6

F
F

M&F(l997)

China

Fergoug

Algelia

Ferrera

Switzerland

-.

F
F

Gebidem --; ...

~wi cierI and

1968

2)

1945

07

F :

Guanting
Guernsey
Heisonglin
Hengshan

~
China

China

Honglingjin

China
China

FWH

1953

182

HI

1927

IF
IF
I

1959

43

61

FRD

1966

84

1960

38

F
R

1975

02
76

1938

30?

=1

Sudan

HIW

1964

Kunda Pal am

India

Liujixia

China

Jensanpei

Taiwan
!China

Jiaojiazhuang
Khashm El Girba

F
D
FR

La

M&F (1997)

Yes Bhargava et at. (1987)


UNESCO (1985)

I
141-435

China

IF

1974

84

Naodehai

China

IF

1942

63

Nebeur

Thnisia

Ouchi-Kurgan

1961

04

PaJagnedra

Former USSR
Switzerland

PD
FR

1952

28

Prieto

Puerto Rico

Rioni

Former USSR

Sakura

Japan

60

1956

1960

USA

Sanmenxia

China

Sanshenggong
Santa Maria
Santo Domingo

China
Guatemala
Venezuela

1974

0-7

Sefid-Rud

Iran

HI

1962

35

Shiaodaokuo

China

Shimalin

China

Shuicaozi

China

Warsak
Yanouxia

Pakistan
China

F
F
F
F

9
M

Yes SNCOLD (1982)

Lo

M&F(1997)

M&F(l997)

ht

Yoon, 1992

i M&F(l997)

Yes M&F(1997)
M&F(l997)

I
F
F
F

L
N

FM
LoP I

8-1]
15-45

La

HI

1960

08

F
F

68

Zemo-Afchar

Former USSR

! 1927

Zhenziliang

China

1958

Drawdown
Full
Partial

M&F(1997)
Yes K&C (1979)
Yes M&F(l997)
M&F(l997)
M&F(1997)

F
1958 I 19

Qian (1982)
Yes UNESCO (1985)

F
H

Intake forebay

M&F(l997)

Yes C&Z (1992)


Yes UNESCO (1985)

F
FRD II F

22

2 in series

Yes UNESCO (1985)


P&D(1988)

M&F(1997)

San Gabriel

Yes . Hwang (1985)


M&F(1997)

Yes Jowett (1984)

DF

Yes M&F(1997)

F
F

HI
H

Yes J&M (1963)

LI
F
P

Nanqin

Mode
Flushing
Density current
Routing/sluicing
Uncontrolled

5880

FIFI

USA

1924

(1982)

Yes B&P(1986)
I

Morris

1954

i SNCOLD

Yes M&F(l997)

Puerto Rico

M&F(l997)
M&F(1997)

La

F
F
FR

M&F(l997)
Yes M&P(l997)
Yes i R&S (1982)

11

Yes UNESCO (1985)


Yes Zbang et al. (1976)

New Zealand

Qian (1982)
38
17-21

M&F(l997)

La

Mangahao

Hydropower
Irrigation
Flood control
Water supply
Multipurpose

F iF
F
F

Loiza (Carraizo)

Purpose

FD

1953

i Algeria

12-18
I

Parallel resrs

M&F(1997)

India

Iril Emda

EI

F
H

M&F(1997)
M&F(1997)

China
, Austria

Hongqi
Icbmi

F
1985

M&F(l997)
Yes M&F(I997)
, M&F(l997)

25

M&F(1997)

China

China

Comment

Principal
reference

Yes J&K(1984)

Donfanghong

Grimsel

:l~

~.g
E

Dashikau

Guanshan

,I

~~

17

Dashidaira

Groiind

jL

E-

'00:

c c

!:

;::

0
01

1966

China
' China

Genshanpei

68

Costa Rica

Chirurt

1981

""0

.~

~0

..,

~~

.;:.'"

c~
<.)0-

Sediment removal

~
~
I:l..

Country

P
P

23-83
592

10-67

I
Season
Flood
Early flood
Late flood
Non-flood

Yes IWHR(1983)
Yes Mahmood (1987)
M&F(1997)
I
Yes UNESCO (1985)
Yes Zhang et (d. (1976)
Enhancements
Mechanical
Lateral channels
Longitudinal channels
Piping to induce lateral erosion
Fluctuating pool level

3 in series

~.

WORLDWIDE SEDIMENT FLUSH!NG

Table 4.2.

Summmy of experience infiushing


Numbers of reservoirs flushed

Country

Hydropower

Irrigation or
water supply

Flood
control

Algeria
Austria

China

Costa Rica

Former USSR
Guatemala
India
I

II

Iran

Total

21

15

2
1

l
Japan

Unknown

New Zealand

Pakistan

Puerto Rico

Sudan
Switzerland

I
1

1
1

1
1

2
1

Taiwan
Tunisia
US A

Venezuela
Totals

19

I
,

11

25

50

cases, five were considered to involve a degree of sluicing and in five densitycurrent 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

Table 4.3.

Distribution offlushing experience by purpose

Hydropower, irrigation and water supply

Flood control

Multipurpose

25

Total

Table 4.4.

Summary offlushing techniques

Drawdown for flushing

Flushing season

Enhancements

Full

21

Early flood

Mechanical

Partial

Flood (not specific)

Lateral channels

Late flood

3 . Longitudinal channels

Non-flood

Not stated

36

Not stated

22

Lateral piping

Fluctuating pool

Total sample

50

Total sample

I 50 I Total number*

12

* Reservoirs where one or more enhancements are recorded

or partial. In six of the cases where an enhancement is recorded, little or no


additional data on the flushing operation is available.
Enhancements to flushing were apparently attempted at a total of 11 reservoirs
out of 50 cases. Excluding the case where fluctuations in the pool level was the
only such measure, this leaves ten cases where the enhancement required
physical activity by labour and earthmoving plant within the reservoir basin. This
represents 20% of the cases. Deliberate fluctuations in reservoir level during the
flushing operation are probably undertaken more widely than indicated in the
table.

4.3.3. Sediments flushed


Most of the references provide little or no information on the sediments being
flushed, although it is clearly the case that most of the sediments readily removed

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.

CASE STUDIES OF RESERVOIR FLUSHING

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.
JiJangahoo reservoir (New Zealand, 1924)
This example suggests that sedimentation was a consideration in the design, but
that no specific planning for sediment flushing was included, because it was
expected that more dams were to be built upstream. Fortunately, a generous lowlevel outlet could be made available by recommissioning gates on the diversion
tunnel, which had remained out of use after problems 25 years earlier. This was
successful in removing a large proportion of the accumulated sediment, after
which annual flushing has been undertaken.

Guernsey reservoir (USA, 1927)


This is a small reservoir in relation to its annual inflow (43%), but it does not
have low-level flushing facilities and apparently there was no attempt at flushing
for the first 30 years of its operation. Partial draw downs for annual flushing
between 1959 and 1962 scoured sediment from the upstream part of the basin
and redeposited it in the lower basin, having a very small effect on the total
volume of sediment deposited in the basin. Fortunately, the construction of

71

Table 4.5.

Detailed list of reservoirs subject to flushing

Reservation/dam

Country

Purpose

Year built or
sll\rlcd
impounding

Year
modified

I Calchmenl

Max depth: m

Surface
area: km'

i area: km'

Original
capacity:
M.m'

Basin
1enlllb:km

Annual inft ow:


M.m'

1990

Designed for flushing?

AM",'
sediment
inHow:Mt

India

Baita

Cachi

Costa Rica

Gebidem

Switzcrland

Hydropower

1981

Hydropower

1966

Hydropower

1968

5] (dam height)
35 (diversion
lunnel)

1996

785

69

200

lB

4-1

I
324

14

24

.0.3
(Atkinson,
1996)

(Atkinson,
1996)3500
(J984ref)
54

1500 (appro~.)

9'()

420

Yes, via diversion tunnel

()'81

Single bottom oUllel suitably


localed

05

Yes, two flushing lunnels below


bydropower 'nlake

I
Gmund

Austria

Hydropower

1945

Gaunting

China

Flood control

1953

1967

1563

311

0J24

094

093

43400

43

229

30

2270

135

02107

1250

73 (19505)
7 (19805)

Yes

Generously-sized bOllom outle~


but flushing not practicable
because of downstream impacts

13 (B&P

I
USA

Guernsey

Heisonglin

Chin.

lITigation &
hydropower

1927

Irrigation & flood


control

1959

42000

29

96 (1963)

24

91

value)

J.7

2100

No

(1927-57)

370

30

29

86

0-70

142

Probably, as lhe disch.tEe capacity


of]O mS/s is much grelltctlban the
mean inflow
I

Hengshan

China

Irrigalion & flood


control

1966

163

65

I'()

133

158

118

IYes, although main flushing outlet


is 15 m above original river-bed

Honglingjin

China

Irrigalion (probnbly)

Icbari

India

Hydropower

Jensanpei

Taiwan

Waler supply (sugar


cane)

1938

Khashm EI Gima

Sudan

lITigation, hydropower

1964

I
!

1960

1364

<42

16-6

432

11-6

5300

077

I
1975

37

1955

II

57 (1976-84 Apparently not for basin, but


lIushing provided for sediment
mean) 22
(median)
exclusion at inlBke
Flushing gallery added 1955

70

106

950

84

& water supply

Loiza (Carraizo)

Pucno Rico

Water supply

1954

Mangahao

New Zealand

Hydropower

1924

534

23

27

Probably, as lbe discharge capacity


of 28 m'/s is much greater Ihan Ihe
mean inllow

27

449

Seven 7 m )( 73 m bottom outlets


controlled by radial gales

038

No, upstream reservoir buill as


sediment trap

I
!

Chillll

Nan'lin

Naodehai

\ China

Irrigation & flood


control

1974

Originally flood
control. lanerly also

1942

453

1970

4-5

29

4501

121

069 or
0531

168

265

16

102

Yes

I irrigation

Yes, lowcring by 5 m during Hood


season. through 8 bonom oUllets

35

17

564

IS 000

13

138

55

26

55

3041 (199 in
SINCOLD
1982)

008

NO! known, but bollom outlel


available

1966-71 &
1990

688400

55

120

9640

43000

1600

Yes, bu! facilities bad to be


upgraded in Stages

None

427

65

()'2

Yes, apparently

56200

50

Probably, us thare are five lowlevel outlets

Ouchi-Kurgan

FonncrUSSR

Hydropower &
irrigation

1961

Plllagnedru

Swil.7.crland

Hydropower

1952

1974

Sanmenxia

China

Muhipurpose .

1960

(m340)

Venezuela

Sanlo Domingo

Hydropower

1974

Irrigalion (primary) &


hydropower

1962

Hydropower

1958

! Irrig.tion&

1960

l'()

3'()

450

82

25

1760

5000

28

"

96

514

()'63

No (no bottom outlet)

42

170

21 100
(1961-70)

153

Apparently nol

6600

Approx.5

53-8

342

Approx.02 -[

Sefid-Rud

lImn

Shuicaozi

China

Pakistan

Warsak

67340

hydropower

lemo-Akhar

Fonner USSR

Hydropower

1927

Zilcnziliang

China

lrri~li"n

1958

(prohably)

18

* Key: A6.1, e\c. Alkinson (1996)


Chl9, elc. Morri. & Fall (1997)
DETR
This "'I,on
LTCR
Long-tcrlll capacily ratio (Atkin.on. 1996)

<14

1740

I<

36-6

Protr.bly. as the discharge capacity


of 57 m'/s is much greater th..... the
mean inflow

S~"\Iillh"11t:nit1R

Mudd studics'!

e.\pcricm:e

Rcfcrcm:cs

C'lS&!

LTCR:

Slud i l!~"

Mt::m mit:

~ri~inall\'

t!$limalt:d a.__ 0'0'12 M.m.l/vr (.lbmn 4%

interrupt ion lO
thcrt!aftc:"

mCf':lhs

E;ilim;JlI.:d lhat 18%

nUW! thtl~uyh

wiltKJut deposition.

pas~c:; by dCtUiit!l current "cnlin~ and 18% dcposill!L.!

Yes. :t((c:rrm.'hlern~
futlnd

Used Llivcniiun lUllltcl. ck::uin!! OJ8 M .m' in 40 huurs


go:nt:rJliun~ ann"'utli nashing P(\'Poscd

o(ori~inal 'l';r"~e): bUI ()45 M.nr' ao:urnulalcd in firsl I~

54~

A6.1
DETR

JM''''''' (19921: R:unfre7. & RUdri;u.:7.\I99Z): M".,.is &


Fan 11~9il

Commenced 1973 :lnd undcruIkcn I~ [imc~ in 13 \lars.


reduced lr.lppinl:: i'rt1m 8~~ [('I ~7C;r. (fi~,t1tt:.'i on Id:t nc~d
- ufter nushln~ SI:utr.:L.!'.')

~5

Cil 19
DETR

da.rifyin~

Virtually nu sccJimcnl .3CCUI11UI:ltiun.

h..:C:lU~C

of gOl".;e:lyre

1;l!omt.try J.nd annual tlu~h.in:;

02 Mllyr initially. ".'ducin~ to 007 Mllyr ufter upsl=m


","ervoir built 1967
350 M.m' dCpt"ilCd 1953-60: sUbsequenrly m~ny
up~ueam l"C'ScrYoirs constructet.!. 5ubs(:lntially n:ducing
sedimcm inl1ow:,. B&P value for annual scdim1!nl inllow
recommended for plill1ning pu"",ses in 1986

R~:;.crJl1ir cmpltcJ flu!-J. day~ per yc:v ;,md about J ~t.nl"


of WOlter used; unt!crta~cn rrum ('Iutsct; hypa.'is tunnel
considered. but ~jcch",-d un C(tst grounds

Phy!\il:al

Flusbing undeOUlken intermillCnlly 194:1-60: annuully


Ih."",rLOr

Yc.,

DUWUl1s ",,,I. (1982): SNCOLD (1982), UNESCO 1I9H5):


Alkin.on 1.1996); Mom, & F:m ( 1997)

162 M.m' deposilion in fir~t :; yt!Ol'S of op.!r.ltion (6%


slur"lI" los., p<r year): capadty reduced to j87 M.m' by
1973

Ao.3
DETR

Only one nu..t1ing operalion (1954) reported, removing


10% of annu:::al inUow, parlly hy venting clen:oity currt!nt!'

IRTCES (1985); UNESCO (1935); Atkinson (1996);


Moni, & F:ln 11997); Binnic & Partners (1986) (some dot.
incon5istent betw~n source.c;)

A7.1
DETR

20

AlIen1fued in Four y~ 19'9-62: nc;u ccmsidc::n.=d


ccnntlmic:li ur en'ectiv~. 0lS rt!covcreti <Q2% of origin.II

J.r~"Cki

A7.2
DETR

26

&: Murphy (1963): UNESCO (1985): Mahmood


(1987): AllUn",n 11996); Moms 8< Fan (1997)

Zhan;., 01. (1976): Xia i1~SO): UNESCO (198;1:


Atkinson (1996); I-Io.,.is ," Fan (1997)

, From 1962. llc::nsity cum:nt venting and tklod Se3.C;lln


sluicing r<:duced lrap efficiency 10 about 15%: lalenll
erosion IAO<:hniquc successfully implemented from 1980.
ra:ovt!ring some losl storage: longu:nn c:::1pxity c~pccled
!Q be 30-35% of oriinal

319 M.m dopa,ilcd 196C~73. rcuching deplh of 27 m :II


dom

A7.J
CIt25

DETR

AM

IRTCES (19S5): UNESCO 119S'): Atkin.'iOn (1996):

Emplied & nu..hod ror 37 days in 1974. removing OM


M.mJ of depo.sits: 52 days in 1979 removcd 103 M.m"

Moms & F:tn (1997} (su,,"..: d~b inc,,"si~lent bt:"tween

I
I

77

DETR

SOLH't.~)

057 M.m"' deposited per year 1960-63 in impound!ng


mode. representing 35% s[omgc loss per ye:u-

\Varer 1evel lowered in nood lensan, resuttin!! in suh"t~\ntial


reduction in r:ue of "crag. In.'. 10 0",,' M.m; p<r ye:"
1~6.J.-7J: technique is e.~5entially routing/!iiiluicing

ZI,.ng er al. (1976): IRTCES (1985); UNESCO (198;):


Alldnson (1996)

S.:c!imcnlO1tion reached spillway crest after one year; 85%


tr:::apping much gre-Jler (h:L"l indicated by Brune's curvcs

Flushed annually by fully-op.ning spillway gaI.s:


problems with :abr.1sion dumage to spilhv:::ay ~d roller

Bharg.Y:1., al. (lq87): Mohan ., al. (1982); Alkinson


(19961

A7A
DETR

Hwong (1985); Paul.l:: Dhillon (19881: Monis & Fon

DETR

( '.~,!):

99

Rieno..1 &: S<hnollc (1932): Alkin50n (1996)

rut

~9 3% ofstorn2C lust 1927-'7 when ~ime"I<onlriburin1:!:


C'atchm.en( red~t!d from 14000 to 1800 km!
-

A6.!
Ch~1

DETR

36

.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 15 month." aa1l1uillly,


virtuaity mesting ~ub~qucnl sedimcntarion. hut not
re.... toring c3f':Jcity. minor mising of im['lOunding level :1bout
1942 and 1958

year
I
Cap~ity

seriously de~ie1l!d

L..:.lS{ 53t':~ ot' ..:a9:.u.:i(y


aw:le.s blocKctl

(I'1~7)

Flushing opo<"Jtions in 1971 .nd 1973 c.ch ",moved 85 Mt

lCJ5J-~~~

:.hrcc {lOU mm

lowlcv~J

Mechanic:Ji mt:thods auemptcd unsuccessfully in 199J.;


pass-through planned and .,petted to <educe
dredging con.;idered in (995; \cchniq!J~ :s
I cssctuially r,outing/sluicing

HEC6 (Morris & Hu.


1992)

~...'!!im"nt

UNESCO 1198S): EI H.gToyeb (1980): El FailhSiI:ld


(1980); CFGB (J 973 '" 1982)

DETR

Webb &: Soi~rL.orc7. (1997): Morri .l; Fun (1997)

Ch2Q

JOW"lr (1984); Alkinson (\ 996)

OETR

.~cdimc:H.:ltk'll':;

59% of Slara!e 1o", by 1958: prublem inc, in~I) ",riOu..


by mid-I 96C.'

Slor.gc loss 53;~ by 1983 (~ppa ...:ntly bas.a ~n WL of


t rS In. whereas nVL = ! 24 enl: life span then l!:tpl!::wd :c
be 2000 if RUlOhing nca in~tigatt:d

CilPllCity reduced

LO iiOOUL

50% hy ! 950.

bUl

:ccuvcr:t.! to

about B<X~ by ""dy 19iO..

Flu.hod in 1969 through lowl".,.cl dive"ion tunnel anLi


75% of accumulated .!<diment removed in one monlh;
subsequentJy :1I1nl.L:llJy emptied and l1ushr.:d during ) .. wt:ck
clo.sure of po\verhoust

Dun~ity current venting commenced

p~ing :lbout

Chen", Zhao (1992): Mmris 3< Fon (1997)

DETR

64% of annual scdimonlload in 1977-84: ~xp"ritnen,"1


ftushing from 198J. with gaud rcsuIL'Y: (.'Uncludcd Rushing
,hould O~ under!okcn for 4 clays overy 3-4 years

I
I

IRTCES (19851; L'NE5CO (19851; A.kinson (1996)

ilo!totn oulict> ung.l<:ci prior to 1970. 0 fl\1Sllinll appears

Bod level, nJ>" up to 23 m by 1969, ,edimen. volume


oppe-.Jrs to have stabilised at 30 M.m.' sir:<:c t953

19i7;

to have been j);&IUr.:tI

ISluicrd for ~ month., annually sine. 1963


I

1973 Hood c;tuscd lS M.:rr' dC(l<~.liit:on (33% of origin:J.1


bmtom \lur!el

IRTCES (1985): UNESCO (19851: Arkinson (1996)

.
Lie::hi & Hacoerli (I 97O'J: SNCOLD
(1996)

Flushing (aioe'!! by m""banic;ll pl.nt) II/iS-03179


rcmovc:u 2,.", ~I.mJ: 1760 m 10n1: scdiment bv!)a.5.S tunnel

.slor~f!} and .liuhnK:~c.:d

119~2):

AtKinscn

A7.5
DETR

10

A6.3
OETR

100

11.7.6
Ch14
DETR

39

A6.6

100

:~d;:i:lQ~~~;\~nl~~i~:~:dl~: villUmly full cupaelty can

low .. It:'IIct Out[CL;; flushed for 4 months annually; six


devl!!0r>ment r.rage~ de~ri;,cd in li1ernture

I
osa M.m depo;'.e!! in two years 1975-78: on "'I.m' in

Only on. 1Iu.<hing opcmtian reported. nner 4 ye:u; iLieu


by bulldozers; 1Iu.<b.d 50-60% of ~eposilion in j doy.:
conelucled Ill.t ilushing .huuhJ be annuul

four years 197-1-78

K;umdick & Ch.:!nlot

Yes

~eu !~,:c!s ::.t

rmp!:::,'!.:n:et.!

::''tpcrimt:~t::~!y from 195; b:.1t !imi!:!c by

(1~79);

Atkinson (1996)

DETR

A7.7

Toio"i r al. (1991): Mahmood (IY'7); Aikin,'on (1996);

Flus hing (:::about 4 month"/yrJ commcncdi in 198C: after 7


years 25% of lost storage h:r.tJ been reco\'cred: from t 99~
fluodplain erosion ~nn:lllced using diversion ciumnel!::
~."(pe.;:~~d mat longumn staragt: CJposcity could be up to
90% uf original

Sc',ere. causing less or 2 t % of !.he :.tor.lge c:lp:!clty ~r


year up 10 19S0 (Imp efficiency 73%); mas I nf sedim"",
relc:JSC' occurred in d..:n!ittj' currents

8,: 3 ~'l.ml {~3% d' 5tarJg~) tos~ :~1:::1-81 ;


dam oni:- "] m ~dow impoundin~ tc.:\'ct

Zhang &: long (1980): UNESCO (19851: Atkinson (19961:


Mom, &: F,n (1997)

Rehanili!.lLion from 1966 includod con;[ruttion of Inrger

Severe. with 18110 ,'vlt dep<>.<ilcd in first 18 month.,

DETR

nigh elcv:1tion uf !ipillway and short dl.!rltiun :;u1nual1y lo


abOUt one third of inflow

(WeHR i

!9~~):

IRTCES

(19~51: liNESCO (IQi 5r

A'kin,on ( 1996)

I
I

A7.Pt

DETR

DETR
right b.u:k. lc:lding to puw'.:r

int~~.;s

76% <li Cllp:tCtLj' lo~t in Orsl 10 yt'tlrs

43 M.m.' dcp0:iitl"C.I
r.1vUl!. rcp(!!..:cntin~

p~r

yt::tr 195lJ-6 1 in impoundln~


12% ${Qi.!g..! )o!t; per YC:lr

I
:

hnpiemcnlcd from 193'J. with. full dr:l'.\'do','In :md ~pp~Jr!:d


to k..:ep SitU:lLion :;tubll! up 'D J953 . n:rnoving :tbout l M.rn'
f':r yt:lr
I
\\:'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'
Itjrjl-1J: tC'chnique:s .!!i5t!'nt1i!lIy rm.!li;"1gJ~bicir.g

13

ChD

'fUC~ {19S5): Atkinson i IY96): l.iN ESCO


Morris &: Fan (1997)

Zh:;.ng

t:(

tTl.

!!~76j

DETR

{Ii.j~j,;

II

39

~~

()

~
Table 4.6.
Reservoir

Year

Capacity:
M.m3

;;:.,'

-~

Mean
annual flow:
M.m 3

Mean or
median
sediment
inflow: Mt

Mangahao
Guernsey

1927

Zemo-Afchar

1927

Jensanpei

1938

81

Naodehai

1942

168

265

16

Gmund

1945

093

135

007

Palagnedra

1952

55

199

Guanting

1953

2270

Shuicaozi

1958

Heisonglin

91

Flushing
discharge:

m 3/s

Flushing
volume:
M.m3

Ratios: %
CIl

SC:
tlm3

VII

Q/I

2100

17

125

54

6600

450

30

43

19

188

26

215

05

Atkinson
(1996)

95

07

75

008

28

15

1250

13

182

06

96

514

063

50

19

66

307

1959

86

142

071

10

61

83

2221

Sanmenxia

1960

9640

43000

1600

22

Warsak

1960

170

21 100

153

Ouchi-Kurgan

1961

564

15000

13

-~:

-1. - --'

25

045

432

2800

-l - j -

-- - L.~

583

Current
or
recent
estimate

26

03

Comment

~rn

3:

a.>

()

()

::l

tI)

Insufficient data

Note I

70

Y?

98

85

100

100

Assisted by bulldozers

Note 2

Insufficient data

Y?
Y

Note 3

17

50

Note 4

08

90

10

0-4

23

L-J

'-

-' -~

Vl

30

---L-J'

-t

45

30

589

""T1
Vl

Y? Original capacity not given

39

08

C"-.

20

"

63

J~"-=",,,

LTCR%

1924

-1~",,";,,,j

o
o

Summary of key flushing parameters

10

40

L_-J

L-......:

Y? Note 4

--L-J

,_.

I --.

L.- . ..J_

.I

Sdid-Rud

1962

1760

Khashm EI Girba

1964

950

5008

50

100

R4

7100

1067

35

28

63

2]

1966

1].3

15g

Cach(

1966

54

1500

Y? Note 6

84

17
08

36

71

3393

429

20

2}

56

Note 7

Y
-

05

75

15

90

Y Note 5

--

I-Jt>ngshan

1968

75

88

Gebidclll

13

147

--.

..

07

99

99

100

96

75

36

34

N?

Note 9

85

85

Note 10

~-

SanlO Domingo

1974

30

450

02

10

07

67

70

Nunq'in

1974

102

121

05

14

8-4

52

365

1975

116

5300

22

02

19

1981

24

2700

03

0]

13

Note 8

---.. lchari
- -.

naira

100

117

03

NOles:
I. Flushing diseharge given as 120-1L10 ml/s at 12-13 III drawdown; Q and V from Atkinson (1996)
2. Value or sediment inflow ('or planning, as recommended in 13 innie & Partners (1986) report; flushing not acceptable due to downstream constraints
3. Sediment management includes sluicing during 1100d scason and lise of lateral channels
I
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)
7 . Much greater Ilushing disc\larges possible via a higher outlet (abollt 25% of water deplh above base of dalll)
8. Mllch t~realer Illlshing discharges possible if needed
9. No hollom ourlet; llushing via gated spill.way only
10. Mean anllual !low taken as mean of two values reported in li rerallll'e

~
0

;0

0
n1
(/I

rn

0
3:
fl1

-I
11

r
C

VI

Z
(.)

"

. :..

.:0,

EVACUATION OF SEDIMENTS

upstream reservoirs controlling most of the catchment appears to have reduced


the need for flushing.

Zemo-Afchar reservoir (Former USSR, (927)


After over a decade with limited drawdown, apparently having little effect and
allowing about 75% of the storage capacity to be lost, active flushing was
undertaken between one and four times per annum. The -data are ambiguous, but
suggest that the long-term accretion has been arrested and some reversal
achieved.
Jensanpei reservoir (Taiwan, 1938)
The capacity of this reservoir is believed to be about 30% of the mean annual
inflow, which may be considered as the boundary between hydrologically small
and large reservoirs. In the first 18 years of its operation, over 60% of the storage
was lost, but this was then arrested and an equiliblium maintained through annual
flushing. This is undertaken during the latter part of the non-flood season, which
coincides with a period of no water demand from the industrial consumer.
Naodehai reservoir (China, 1942)
This was originally an uncontrolled flood detention reservoir, with a capacity of
63 % of mean annual inflow, but gates were later added to the bottom outlet to
allow some impounding of clear water for irrigation. The mode of operation
probably bears more resemblance to sluicing than flushing, with riverine flow
est~blished for much of the year when flows are low. During floods, which
contain most of the annual sediment load, the water levels rise and there is some
deposition over the flood plains, which will dry and consolidate between floods
and which will not be amenable to subsequent erosion.
During the first 30 years of operation, the available storage volume has ranged
between 58% and 80% of the original capacity, with the lowest value having
occurred in 1950. It appears to be dominated by massive deposition in the largest
floods, followed by a period of progressive erosion, but the data are too limited
to judge the degree to which any active flushing may be practised through the
operation of the gates.
Gmiind reservoir (Austria, 1945)
This reservoir is hydrologically very small, at under 1% of the mean annual
inflow. There. is a coarse bed load, which is bypassed around the reservoir in a
tunneL Periodic flushing was undertaken initially and the storage loss reached
20% by the early 1960s. Annual flushing commenced in 1960, an additional
bottom outlet was constructed in 1963 and -sediment inflows were substantially
reduced after 1967 by the construction of an upstream dam. As a result, the
storage loss stabilised and a modest recovery occurred. Flushing is greatly aided
by the ability to control inflows by releases from the upstream reservoir.
Palagnedra reservoir (SWitzerland, 1952)
The reservoir is hydrologically small, at less than 3% of the mean annual inflow.
Sediment inflows appear to have varied substantially from year to year, with

76

VVORLDVv'IDE SEDIMENT FLUSHING

major floods contributing the majority of sediment. It suffered sediment


deposition amounting to 27% of the original capacity by 1968 and 69% by 1978
(the latter probably largely the result of the August 1978 flood). A sediment
bypass tunnel was constructed in 1974 and a major flushing operation lasting
four and a half months was undertaken in 1978-79, aided by bulldozers, clearing
most of the deposited sediment. It is not known what flushing operations have
been undertaken since, but it appears likely that annual flushing would enable
most of the capacity to be retained.

Guanting reservoir (China, 1953)


This is a hydrologically large reservoir, at about 18 times the mean annual runoff. The rate of sedimentation has reduced progressively over the years, from an
initial value of about 3% per annum to 03% in the 1970s, apparently due to the
construction of a vast number of reservoirs within its catchment and the diversion
of :bighly turbid flows for warping the agricultural land. Only one case of partial
drawdown is reported, in 1954, which removed only 10% of the annual sediment
inflow. Downstream impacts, including a water supply intake, two hydropower
schemes and the effects of accretion on flood risks, prevent substantive flushing
being undertaken as the primary means of sediment management.

Shuicaozi reservoir (China, f 958)


Shuicaozi reservoir is hydrologically small, at
than 2 % of the annual inflow.
With no botton1 outlet available for flushing, only partial drawdown was
undertaken by way of the spillway, which was found to be of limited efficacy. In
spite of six flushing operations between 1965 and 1981, sedimentation was very
severe, consuming 85% of the oliginal storage capacity by 1981. An improved
fiuslIing procedure was adopted
1984 and was found to be successful, but no
further data are available.
I-Jeisongfin reServoir (China, 1959)

This is a small reservoir, but it is hydrologically large with a capacity/inflow ratio


of 61 %. Following sedimentation problems at many earlier Chinese reservoirs,
this reservoir was used as a test-bed for sediment management techniques.
Initially, it was operated only with impounding, resulting in average siltation of
60/0 per annum over the first three years. Subsequently, the mode of operation was
changed to emptying the reservoir during the flood season, when the water is
most turbid, and impounding only during the non-flood season. This technique
which is essentially sediment routing or sluicing - reduced the trap
efficiency to 15%, but further improvement was required in the longer term.
Starting from 1980, experiments were carried out, resulting in the development
and routine use of lateral erosion techniques. Between 1980 and 1985 deposits
amounting to about 6% of the storage capacity were recovered by this method,
at a very high sediment/water ratio of 23 %.
It is expected that a long-term balance bet\veen sediment inflovvs and outno\vs
can be maintained at Heisonglin reservoir, with a residual
capacity of

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
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
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.
The reduction and control of sediment deposition and the development of a
sustainable sediment. management . regime became a high priority for the
Sanmenxia reservoir. The development occurred over six stages between 1962
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
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
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
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
unsuccessful. The Warsak reservoir has apparently reached a broad equilibrium,
with virtually no residual live storage.
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
annually, but apparently with only a limited drawdown. The volume of deposited
sediment has stabilised at 50-55% of the original capacity since 1968.
Sefid-Rud reservoir (Iran, 1962)
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,
reducing the storage capacity at a mean annual rate of 21 % 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
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.

1
1
r

1
]

I
I

1
I

I
I
I

78

WORLDWIDE SEDIMENT FLUSHING

Recovery and maintenance of storage capacity has been aided by lateral


erosion, by piping and by the use of a longitudinal diversion channel. It is
anticipated that, by creating a new diversion channel each year, it would be
possible to reach a Iong-telTIl storage capacity of 900/0, compared with 75% by
flushing alone.

Khashm EI Girba reservoir (Sudan, 1964)


No information is available on the ratio of the reservoir capacity to the annual
inflow. The limited data available on flushing operations suggests that it has been
successful.
Hengshan reservoir (China, 1966)
This is a small reservoir on a steep stream and is hydrologically large, having a
capacity/inflow ratio of 84%. In the first -eight years of operation, 24% of tl]e
original capacity was occupied by sediment deposits. Flushing, wpich was
undertaken with the reservoir empty in 1974, 1979, 1982 and in 1986, was
effective in recovering lost storage and indicated that flushing every few years
would be sufficient in this case. Although the Hengshan reservoir is hydro- .
logicaUy large, flushing was probably effective because of the steep valley
gradient and side-slopes.
Cach{ reservoir (Costa Rica, 1966)
The Cacri reservoir is hydrologically small, at 36% of the mean annual inflow
and the al.1-Dual sediment load would have a deposited volume rather over 1% of
the original storage volume. For the first seven years it was operated without
flushing and trapped 82% of the incoming sediment load. Between 1973 and
1990 flushing was undertaken 14 times. It was effective in the lower part of the
basin, but less so in the upper part, which is progressively filling with sand and
coarser material. Overall, however, flushing has been considered successful in
maintaining the storage capacity at the Cacm reservoir.

Gebidem reservoir (Switzerland, 1968)


Gebidem is a hydrologically small reservoir, with a capacity/inflow ratio of 21 %.
Sediment inflows are high due to glacial activity, with stone sizes up to 100 rnm,
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
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.
Santo Domingo reservoir (Venezuela, 1974)
This is hydrologically small, at less than 1% of the merul annual inflow, and
contains two brfu'1ches, being built at the confluence of two rivers. Sediment loads
\vere expected to be high in relation to the original storage capacity, at about 8%
per annum, and model studies had been undertaken which suggested that the
installed flushing facilities would be sufficient. For the first four years the

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.
Nanqin reservoir (China, 1974)

Nanqin is a hydrologically small reservoir, with a capacity/inflow ratio of 8%.


Initially it was used for flood detention only, then from 1976 flows were
impounded to a middle level for irrigation. By 1983,53% of the storage capacity
had been lost, following whIch, from 1984, an improved regime of sediment
management was put into place. The steep longitudinal slope of the reservoir was
suitable for density current venting, which was practised from 1977, achieving a
trap efficiency of 36%. The first empty flushing was undertaken at the end of the
1984 flood season and was highly effective, removing all of the sediment
deposited that season, together with an additional amount deposited earlier,
equivalent to 7% of the original storage.
J'he operating rules that were subsequently derived, maintain the high pool
level during the flood season to trap bed load deposits near the upstream end and
prevent them armouring the more erodible deposits further downstream. Empty
flushing is to be undertaken every three or four years, at the end of the flood
season, and density current venting is promoted. It was estimated that a longterm storage capacity of the order of 75% of the original capacity could be
maintained by following these rules.
Ichari reservoir (India, 1975)
This is hydrologically very small, with a capacity of only 02% of the mean
annual inflow. The annual sediment load is highly variable, with a median value
having the potential to consume about 20% of the original storage per annum.
This was borne out in the first year of operation, when the storage capacity was
reduced by 23%. Over the first six years the total loss of storage was 60%. The
sediment ranges up to cobbles and has severely damaged the spillway roller
bucket.
Although the dam includes facilities for excluding coarse sediment at the
hydropower intake, there is apparently no low-level outlet for flushing sediment
from the dead storage of the impoundment. Flushing by way of the gated
spillway was undertaken annually from 1976 and it appears likely that an
approximate equilibrium, with the dead storage entirely filled with sediment,
existed from about 1980. The long-term storage is likely to be about 35% of the
original capacity.

80

WORLDWIDE SEDIMENT FLUSHING

Baira reservoir (India, 1981)

The Baira reservoir is hydrologically very small with a capacity 1.1 the order of
01 % of the mean annual run-off from the catchment. In the first 18 months of
operation, almost 20% of the original capacity had been consumed, representing
at least double the average annual sediment load assunled during the design. The
construction period diversion tunnel had been fitted with gates to facilitate
flushing and model studies had suggested that this wouldbe capable of removing
virtually all the deposited sediment. The first flushing operation was successful,
removing over 80% of the deposition in 40 hours, and it appears that annual
flushing should be effective in maintaining a large proportion of the original
storage capacity.

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:
, the hydrology of the catchment
the sedimentology of the catchment
" the storage capacity of the reservoir
o the sediment deposition'-potential
8
the shape of the reservoir basin
e the low-level outlet facilities provided
operational limitations
e whether full or partial drawdown is to be deployed
e the scope for enhancements to flushing
o downstream impacts
o the critelia for judging the success of flushing.

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
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
in number and all hydro.lo.gically small)
en the availability and co.ntro.l o.f flushing flews when they are needed .
. 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
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
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
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
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
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
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

!
,.--L

1
_1

1
1
1

1
1

'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 01 % 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.
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 carryover 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.
Reservoirs which are initially hydrologically large and impnlctlcable to flush,
may become practicable to flush as sedimentation reduces the storage capacity,
allowing an acceptable residual capacity to be sustained in the long term. The
Heisonglin reservoir probably falls into that category.
o

Sediment deposition potential


As well as the relationship between reservoir storage capacity and mean annual
inflow, it is relevant to consider the potential accretion which would result if there
were 100% trapping of the sediment load. A potential accretion rate of 1-2% per
annum probably represents a reasonable boundary between reservoirs where
flushing should be started early and those where it might be delayed for perhaps
20 years. Potential rates of 5% or higher certainly spell danger, requiring flushing
to be fully planned in the design and implemented from the outset. Of the
reservoirs listed in Table 4.6, for which this information is available, 50% have

83

EVACUATION OF SEDIMENTS

potential rates (in tonnes of sedimentlm3 of water) above 6-7 % (corresponding


to potential accretion of about 5% or more), the highest being over 20%. Very
few reservoirs in Table 4.6 have potential accretion rates of less than 1-2%, but
this is not really surprising, as there is clearly less need for, and interest in,
flushing where the potential rate of accretion is very low.
Hydrologically large reservoirs, except where sited in the areas of the highest
sediment yield, tend to have low rates of potential annual accretion, so may not
be expected to need active flushing during their economic life. The Guanting
reservoir would appear, on the basis of the raw figures, to now fall into this
category, the sedimentation problems having arisen because of higher historic
sediment loads, which were nearly all concentrated into the reservoir arm that
had a lesser storage capacity. Nevertheless, in the case of large dams, proper
consideration needs to be given to the long-term conditions, including the
practicability of decommissioning and the problems associated with any eventual
release of the reservoir deposits.

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

example of a reservoir with limited inflows available in a significant tributary


valley is the Guanting reservoir.

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 noncomplial"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:
recreation, such as boating or swimming
water supply intakes
III hydropower intakes
reservoirs.
@

Ill)

Criteria for the success of flushing


Past attempts to define objective criteria in order to judge the success or
otherwise of flushing have been:
G

L~at

there should eventually be a balance, over time, between the sediment


inflow to and outflow from the reservoir
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.

Criteria such as these, however, are probably unnecessarily restrictive.


Whether an absolute sediment balance is required depends on the potential rate
of loss of storage. In the case of a large reservoir which might lose 2% of its
capacity per annum, flushing which lowers that rate to 1% might be considered
successfui, as it would lengthen substantially the useful life of the reservoir.
Similarly, an ultimate equilibrium long-term capacity of only 200/0 might be
considered perfectly acceptable at a reservoir vvhere this was expected at the
outset and where alternative solutions for water supply, power generation or flood
control are less attractive.

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
or has been unsuccessful are probably limited to the following:
Guernsey
Guanting
Warsak
Ichari

Effective flushing is not possible, due to limitations of low-level


outlets, but apparently it is not needed, because of vastly reduced
sediment inflows.
Major flushing is not feasible because of downstream constraints; the
reservoir is required to contain sediment and prevent significant
.
downstream sedilnentation.
Flushing is not possible because of a lack of low-level outlets; a lack
of live storage limits hydropower generation to the 'run of the
river' .
No low-level outlet, so cobbles and gravels must pass down the
spillway, damaging the roller bucket.

J.
J
J

4.4.3. Summary of findings


The findings from the review of the worldwide experience of flushing can be
summarised as follows.
The hydrology and sedimentology of the catchment need to be fully
understood in the planning of flushing facilities for new or existing reservoirs
and need to provide the background for analyses of past sedimentation and
flushing performance.
Successful hydraulic flushing is more likely to be practicable in reservoirs that
are hydrologically small, with a storage capacity less than 30% of the mean
annual inflow. The smaller the reservoir, the greater the chance of it being
successfully flushed and the greater the likely residual storage capacity.
Flushing is vital for the preservation of long-term storage in reservoirs where
the sediment deposition potential is greater than 1-2% of the original capacity.
Even in large reservoirs with a potentially long life, consideration should be
given to possible eventual decommissioning problems when deciding whether
or not to flush.
The shape of the reservoir basin can have a large impact on the practicability
of effective flushing and on the residual storage capacity. Narrow steep-sided
the easiest to flush.
reservoirs in valleys with a steep longitudinal slope

are

88

...l

J
1

WORLDWIDE SEDiMENT FLUSHING

"

"

Wide valleys, where the impoundment covers former floodplains, can be


flushed less effectively, because the deposits tend to consolidate and 'are
remote from the flushing channel.
For effective empty flushing with full drawdown, the low-level outlets must be
both low enough and of sufficient capacity to allow the draw down to be
controlled during the time of year when flushing is undertaken. Proportionately larger outlets are required for flood-season flushing than for flushing
outside the flood season.
Operational considerations, such as water and power demands can inhibit the
ability to flush successfully, but they must not be allowed to prejudice the longterm preservation of an important resource.
Full drawdown and empty flushing have been found to be much more effective
tJ."'1an partial drawdown.
Fluctuations water level and discharge during flushing are beneficial to the
promotion of bank slumping, increasing the rate of sediment discharge.
The deployment of lateral and longitudinal diversion channels has been
successful in promoting flushing in reservoirs that are hydrologically large or
u1.at contain significant proportions of deposition in areas remote from the
main flushing channel .
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, may be a
practicable alternative.
The degree of success in flushL.'1g should be judged by whether it makes a
worthwhile difference to the beneficial uses of the reservoir, rather than simply
by whether it meets numerate objectives, such as a long-term balance between
inflows and outflows, or the retention of a certain percentage of the original
storage volume.

89

ro hi
it d to

5. Geographical areas suited to flushing


5.1.

WORLDWIDEVARIATION IN EROSION RATES

5.1. I. Factors that affect erosion


The erosion rate depends on a cOlnplex interaction of the following factors:

. (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

(e) Land use and human impact


These factors are discussed in Appendix A4.1. It is not easy to generalise
between areas of high and low erosion rates as it depends on their geographical
location.

5. 1.2. Estimates

of global sediment yield

The estiInates derived from more than a dozen studies of global average rates of
denudation have ranged from 006 to 016 rnmlyr (Morris and Fan, 1997).
Estimates for the aggregate worldwide sedilnent yield of between 15 and
20 x 109 t/yr have been given (\Valling and \Vebb, 1996). Areas with sediment
yield over 1000 tlkIn2/yr are 88% of the total land area and account for 690/0 of

93

EVACUATION OF SEDIMENTS

the total sediment load. Regions with less than 50 tJkm2/yr account for about half
of the land area and 21 % of the sediment yield. Case studies of erosion rates are
presented in Appendix A4.2.

Continental variations in erosion rates


A number of estimates have been made of sediment yield on a continental basis.
Results from two such studies are presented in Tables 5.1 and 5.2.
Sediment yield in Asia is four times larger than South America even though
South America experiences the highest rates of run-off in the world. The
sediment yield for basins in Asia is over twice the world average and contributes

Table 5.1.

~I

Continental variations in sediment yield (Mahmood, 1987)

Area
I

mm

km3

North America

756

158

154

Asia

740

257

250

Africa

740

197

1600
790

South America
Europe

Sediment:
Mtlyr

Run-off:
km3

Precipitation

Yield:
tlkm2Jyr

66

171

1460

109

84

108

280

6350

474

380

192

42

109

530

39

35

270

262

118

305

1790

133

97.

75

73

27

230

17

50

60

0-4

28

3000

224

1000

13420

1000

165

70
I

Australia
791

71

25

69

65

Oceania
I
i

Total

1028

1000

386

1000

Table 5.2. Continental variations in sediment yield (Jolly, 1982, taken from Gregory
and Walling, 1973)
Suspended sediment:
tllcrn 2Jyr

Continent

Suspended sediment:
10 Mtlyr

27

550

600

16160

Australia

45

230

Europe

35

330

96

1990

63

1220

Mrica
Asia

North America

~j

J
J

South America

94

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.

Vaiiations in sediment yield by drainage basin


The areas where extreme erosion rates occur are, of course, hidden in the
statistics that define areas according to continent. Various attempts have been
made to identify the drainage basins worldwide which have the highest sediment
yields. Table 5.3 presents the results of one such global study.
Table 5.3 shows high erosion rates mainly in Asia, Oceania, the USA and
in Eastern Europe. The two highest values of sediment yield, both around
8000 tf1<I112/yr are for Chinese rivers.
A more recent study has revealed even higher sediment yields in some areas
where sufficient data exists for analysis. These are show later in Table 5.5.
A detailed study by Mahmood (1987), the results of which are shown in Table
5.4, gives the water and sediment yields from the world's major rivers at ocean
level.
Table 5.4 covers a wide variety of the worlds' drainage basins and shows rates
of sediment yield varying from 4 to 2581 tf1<lIi2/yr. A large number of the major
rivers with high sediment yields occur in:
lit
G

Asia, especially in China and India


Oceania, especially in New Guinea
South America, especially in Peru and Colombia.

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.

Areas of high erosion rates


Further COITh.'TIents on areas of particularly heavy erosion: by various researchers,
are given below.
o

Areas of high erosion include mountainoLls areas, such as the Andes,


Himalayas anYd Karakorams, parts of the Rocky mountains and the African rift

95

J
EVACUATION OF SEDIMENTS

1
Table 5.3. World maximum recorded suspended-sediment yields greater than
2000 tlkm 2/yr (Jolly, 1982, from Gregory and Walling, 1973)
River

Location

Average annual yield:


tlkm2/yr

r--'-

Ching

Changchiashan, China

8040

Lo

Chuantou, China

7922

Waipaoa

Kanakanaia, New Zealand

6982

Tjatabon

Java, Indonesia

6250

Lo-Lo

Luyang, China

6068

Pietracuta, Italy

4570

Semani

Urage, Kucit, Albania

4150

Soldier

Pisgash, Iowa, USA

4072

Shkum Bini

Paper, Albania

3590

Kosi

Chatra, India

3130

Yellow

Shenhsien, China

2957

Indus

Kalabagh, Pakistan

2498

Santa Anita

Arcadia, California, USA

2374

Eel

Scotia, California, USA

2292

Marecchia
...,;. .

1
1
I

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
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, .
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
country in the world. Streams draining the central range produce suspended
sediment yields of 13 760 tlkm2jyr. One small basin exports 31 700 tJkm2jyr
(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).
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-

96

.---L

1
1
1
f

1
j

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

River

Country

Drainage area:
million km 2

Run-off:
cm/yr

Oceania

New Guinea

Purari

0031

248

South AmeLica

Peru

Chira

002

25

Sediment:
tlkm 2/yr

Yield:

ppm

2581

1039

2000

8000

Asia

China

Daling

Asia

China

Haiho

Asia

China

Yellow

India

Damodar

Asia
I

North America I USA


Asia :

I
II

I Copper

1800

36 000

005

1620

40500

077

1403

22041

01.02

50

1400

2800

65

1167

1795

1128

1720

1083

1057

917

928

01.06

66

148

Gange sIB rahm

Bangladesh

002

Asia

Vietnam

Hungho

012

103

South America

Colombia

Magdelena

9.24

99

043

100

616

619

80

500

625

0061

126

492

390

097

25

454

1849

31

27

310

1143

Asia

I Burma

Irrawaddy

USA

Susitna

Oceania

New Guinea

Fly

Asia

Pakistan

Indus

Asia

I India
I China

Orinoco

99

111

212

191

Colorado

9.64

211

6750

Europe

Mekong

Brazil

017

I
I
I

I Vietnam
I China

340

59

Pearl

9.44

69

157

228

Amazon

~'15

102

146

143

al1

145

2286

II

C1

30

130

433

(09

S4

111

204

Negro
Rhone
i

203

79

Brazos

531

326

Mexico

South America

214

North America

I USA
II Ar2:en tina
I France

66

Venezuela

North America

246

9.07

South America

South America

46

6833

Po

I
I

241

Italy

Asia

r
194

Europe

Asia

Yangtze

I
I Liaohc .

USSR

~; ~

005

Godavari

4l,,:)Ld.

1\

..

North America

Asia

.'

97

1
EVACUATION OF SEDIMENTS

Table 5.4.

continued

Continent

Country

River

Drainage area:
million krn 2

Run-off:
crnlyr

Sediment:
t/km 2/yr

Yield:
ppm

North AmeIica

USA

Mississippi

327

18

107

602

Africa

Tanzania

Rufiji

018

94

1889

North America

Canada

Fraser

022

51

91

179

Europe

Romania

Danube

081

25

83

325

Africa

Mozambique

Limpopo

041

80

6600

North America

USA

Yukon

084

23

71

308

North America

Canada,",.

181

17

55

327

North America

USA

002

60

50

83

' ''''M ackenzie


Hudson

"

"

"

1
1
i

1
I

Asia

Iraq

Tigris-Eupha

105

50

1152

Europoean AItic

USSR

Indigirka

036

15

39

255

Africa

Egypt

Nile

296

38

3700

South Amelica

Argentina

La Plata

283

17

33

196

Africa

Nigeria

Niger

121

16

33

208

Asia

USSR

Amur

185

18

28

160

Oceania

Australia

Murray

106

28

1364

Africa

South Africa

Orange

102

17

1545

Africa

Mozambique

Zambesi

12

19

17

90

Asia

' India

Mehandi

013

52

15

30

European Artic

USSR

Yana

022

13

14

103

European Artic

USSR

Sev. Dvina

035

30

13

42

North America

USA

Columbia

067

37

12

32

Africa

Zaire

Zaire

382

33

11

34

South Amelica

Brazil

Sao Francisco

064

15

62

European Artic

USSR

Kolyma

064

11

85

European Artic

USSR

Ob

25

15

42

1
1
j

-1

" I'

98

European AItic

USSR

Yenisei

258

22

23

European Artic

USSR

Lena

25

21

23

North America

Canada

St Lawrence

103

43

1
I

..J..

1
T

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).

Areas of low erosion rates


Global minima below 2 t/k.rn?/yr have been documented. Douglas (1973) cites a
yield of 13 tlkm2/yr for the Brindabella catchment (261 km2) and 17 t/k.m 2/yr
for the Queanbeyan River (172 km2) in the southern Tablelands and Highlands of
Table 5.5.
1983)

Values of sediment yield in excess of 10 000 tlkm2/yr (l;Valling and vVebblec

Country

River

China

Dali

China
China

Kenya

II

Mean annual
sediment yield:

Source

tfkrn'1jyr

961

25600

MOli and Meng (1980)

Dali

187

21 700

Mou and Meng (1980)

Dali

3893

16300

j Mou and Meng (1980)

Perkerra

1310

(Unknown)

Taiwan

Drainage
area: km 2

I!

(Unknown)

19520
31 700

Dunne (1979)
Li (1976)

Java

Cilutung

600

12000

I
I Hardjowitjitro (1981)

Java

Cikeruh

250

11 200

Hardjowitjitro (1981)

New Guinea

Aure

4360

11126

Pickup et aI. (1981)

Waiapu

1378

19970

Griffit.~s

Waingaromia

175

17340

Griffiths (1982)

Hikuwai

307

13 890

Griffiths (1982)

South Island

Hokitika

352

17070

Griffiths (1982)

New Zealand

Cleddau

155

13 300

Griffiths (1981)

North Island

New Zealand

New Zealand

J
I

(1982)

99

EVACUATION OF SEDIMENTS

New South Wales, Australia. Values of less than 1 t/km 2/yr have been reported in
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.

" 5.1.3. Maps

of global variation in sediment yields

A number of maps have been produced to illustrate global variations in sediment


yields. The maps of Strakhov (1967) and Fournier (1960) are based on 96 and 60
observations respectively and are shown in Figures 5.1 a and 5.1 b. There are large
discrepancies between the maps with values on the Fournier map frequently of
an order of magnitude greater than on the Strak...lJ.ov map.
A later study by Walling and Webb (1983) was based on 1500 stations with
basin sizes from 1000 to 10 000 km? The highest values are associated with the
loess areas of China and the Cenozoic mountain areas around the Pacific
Margins. High values occur in mountainous areas, mediterranean, semi-arid and
seasonally humid climates (Walling and Webb, 1983). Low values occur in the
desert regions, areas of low relief and glaciated regions (Walling and Webb,
1996). A better likeness can be seen between the maps produced by Lvovich et
aI. (1991) and by Walling and Webb (1996), both based on drainage basins of
between 1000 and 10 000 km2 (Figures 5.2a and 5.2b)
In this study we have used the map produced by Walling and Webb to generate
rates of sediment yield in every country of the world. These are tabulated in
Appendix A4.4. Despite the use of a relatively small unit, such as the country,
highly variable rates were found within many countries. A number of large
countries would clearly be better sub-divided into smaller homogenous regions
but this is a task beyond the scope of this study.
Jansson (1988) makes a number of observations on the 1983 map (Figure
5.2b).
Small rivers in Taiwan, Java and Borneo and in the mountain areas of Pakistan
and Soviet Central Asia have high yields. Thailand and Cambodia have low
values while the Philippines are intermediate. Low values occur in Africa
except for the mountains in northern Africa and pal1s of South Africa and
Lesotho.
In South America the values for large drainage basins are low. In the north, .
smaller basins which drain the Andes have high values. In the Andes in
Venezuela and Colombia, the small rivers have . extremely high sediment
yields. Western Central South America has "fairly high levels while the
mountains of northern Argentina and Bolivia have high values. The lowest
rates of .s~diment yield occur in northern Chile, in the mountains east of the
desert.

100

",

- ,.

AREAS SUITEDTO FLUSHING

5.2.

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).

CLIMATIC ZONES OFTHEWORLD

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.

5.2.2. Precipitatjon regimes and their seasonal variation


High annual precipitation
Figure 5.3 shows the global mean annual precipitation for 1998 in rnm per
month. This shows that annual precipitation values are greater than 150 mm per
month along the eastern edge of Asia in Vietnam, China, Bangladesh and NepaJ
and into the Pacific Islands of Japan, Malaysia, Taiwan and Papua New Guinea;
High rates in Mrica of over 100 rom per month occur in the western centraf
region around Cameroon, Gabon and Zaire. In South America values over
100 mIn per month are found predominantly in Brazil, extending northwards into
Guyana and Surinam. In Central America values over 100 Dl1n per month occur
in Guatemala, Honduras and Nicaragua and in North America they are limited to
a region north of the Gulf of Mexico and to a narrow strip along the northwestern coast stretching into Canada.

Low annual precipitation


.Precipitation rates under 10 IT.L..TD. per month occur in the Saharan region of Africa,
Israel and countries of the Middle East, NIongolia and nearby regions of central
Asia and in the northern regions of Greenland and Canada.

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, Madaascar, Tanzania, Zambia and Angola.
In Asia, high p;ecipitation values- occur in the Pacific Islands and in the northern .
part of Australia.

101

()

.......

. ..
.. .. . . . .. .
.. ... .. . .. .. ...... ...... .. ...... .
...-...............
...
. .. .. ., .. ..............
. . . . .. . . . .. . .
.............................
. .. . ....................
. . . . . . . .. . .. . . . . . . .. . . . . .
1 --'

o
z
o-n

..

..

"!j

..

..

..

.. . .

It

.,

..

~~:::~; ~:~ ~~ ~mwm lj;m\ij:i{:":"""

(/)

3:
m

(/)

t km-2 yr- 1

240

100

50
10

,0

Arid regions

(a)

Figure 5.1.

't"

'

'

Global patterns of sediment yield: (a) after Strakhov (1967); (b) after Fournier (1960)

"

,-

,"

,., '

L--- '

' . L ,,--

'" L.--,- ,

'

AREAS SUITEDTO FLUSHING

103

<

()

~
0

z
0

"m

Vl

-0

3:

-I
Vl

Suspended sediment yield: t km-2 yr- 1


5

20

..

200 1000 5000

(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)

1--

l- -

l--.

~ .

1--

"I

"

~,

~ .~

xv

Sediment yield: t km-2 y r -1

1000

;0
m

750
500

Vl
Vl

250
100
50

VI

Deserts and
permanent ice
(b)

),

I@
-;

0
I~
C

Vl

Vl

Figure 5.2_

continued

Z
G)

-Nean ,monthl}!. precipitation JGPCpmonitoring) _


forthe.- yeor lJan -Dec) -1~98 -.nmm/month
90N..---------r------,-.-----,------....,..------r-o------.

()

o
z
o
"m
(/)

3:

-I

30N - t - - - - -- - t - " I

(/)

EQ~--------~----~

3~~---------+----~--~

6~~--------~----------~----------r----------+--------~~-------~~

9~~--------~----------~----------~---------+----------~--------~
1BO
12ml
60W
(I
GOE
1eo
1
Figure 5.3.

10

25

50

75

tOO

150

Annual precipitation/or 1998 in mm per month

1 :
~

OD

300

0400

800

800 1000

. .

.I

_..__ ..__ _.. _#_ . . __ _

~_. _

...._... . ..
~_"'

:..a ._~

__

~ ; ~ !a.'

. . ..;:J~ 5~ ~~

Ea~-----------~-----~~~

:;;0

VI
VI

C
-l

9~~----------~----------~----------~~------------~----------~~--------~
1fiD
12ml
!)OW
0
tiOE
120
1SO

1Q

25

50

15

100

1M

200

300

400

-l

o
r"C
VI

I:

Figure 5.4.

Precipitation distribution during winter 1998 (December to February)

Z
C)

1
EVACUATION OF SEDIMENTS

Precipitation distribution during spring (March to May). Figure 5.5 shows the
global precipitation distribution during the nOl1hern spring of 1998. In spring,
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
values in Asia are higher in eastern China and Japan with the islands of the
Pacific still experiencing high values.
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
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,
China, the other eastern Asian countries and to the Pacific Islands.
Precipitation distribution during autumn (September to November). Figure 5.7
shows the global precipitation distribution during the nOl1hem autumn of 1998,
with regions of high precipitation now more isolated in America and confined to
a narrow strip along the western coast, Central America and the northern region
of South America. In Africa, high values occur in the western region around
Cameroon and Gabon. In Asia, the highest values are again in the islands of the
Pacific stretching into Japan, India, Vietnam and Cambodia.

The information shown in the four seasonal maps is available on the Internet.
It would be of interest in the present context to derive a world map showing the
degree of seasonal polarity of precipitation, as this may correlate with erosion
rates and sediment yields. However, this exercise was not possible within this
study.

5.2.3. Koppen classifIcation


There have been many climatic classifications produced but one of the most
frequently used is the Koppen classification, with eight climatic regions based on
four temperature zones and one moisture zone and the seasonal domination of air
masses. Details of a version of the Koppen classification by Pidwimy (1999) are
given below and a discussion of alternative classifications is included in
Appendix A4.3.

Tropical wet
e
e

Koppen classification Af.


Maritime tropical air masses all year.

The climate has consistent high daily temperatures ranging from 20-30C.
Monthly temp~rature averages range from 24-30C . . The annual range of
monthly temperatures is about 3C. Precipitation is uniform with a total over

108

1
1
1
1
1

1
I

J..

1
1
1
1
1
I
J-

, 1

-I

Mean ~onthlyprecipnutron (GP9C monitoring)


for :Dlpnng (Mor.AprM(]y) 191)6 Inmm/rnonth

[O-~-----------+---~'-

3(}S - i l - - - - - - - - - I - - - --

-+.i

rn

;0

Ul
Ul

C
-l

D~~----------~-----------F~---------+~--------~----------~~----------,
1
12[}11(
150'1/
0
1 ZOE
1BO

-I

o
."

10

25

50

15

, 00

t flO

200

300

.4QO

600

800

100'D

Ul

:c
Figllre 5_5_

Precipitation distribution during spring 1998 (March to JV/ay)

zC)

~
()

Nean. ,monthl.r' Pfacipitation(GPC.C m


,' on,itoring,)
for summer (Jun',Jul,Aug) 1998 In mm/montft

9{JN------...----~--.....--...------.......------..,._----..-;...,.....----_.,

z
o
-n
Vl

OON

3:
m

-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
Figure 5.6.

Precipitation distribution during summer 1998 (June to August)

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

Vl
Vl

..

90S -\-------+------""i-~----_Ii__------_I--1BO

12mI'

M't.'

tiOE

---__I------~

120E

1BO

C
-l
m
Q
-l

."

10

50

75

100

ffl<J

200

300

400

600

800

1000

r
C

Vl

Figure 5.7.

Precipitation distribution during autumn 1998 (September to November)

Z
C)

EVACUATION OF SEDIMENTS

Table 5.6.

Temp: C

Colombia, 5 oN, elevation 65 Tn


Jan.

Feb.

27

27

Mar. , Apr.

28

28

May

June

July

Aug.

Sept.

Oct.

Nov.

Dec.

Yr

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
rainfall is given in Table 5.6.

....

.-J

Tropical wet and dry


Koppen classification Aw, Am and BS.
Maritime tropical air masses during high sun season and continental tropical air
masses during low sun season.

- I

-!

.._ \

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
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.
During the rainy season the climate is similar to the tropical wet climate. During
the dry season, semi-desert conditions prevail. Some regions experience intensification of rainfall due to monsoons and orographic uplift. A typical monthly
distribution of temperature and rainfall is given in Table 5.7.

J
i

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:
low relative humidity (10-30%) and cloud cover
low frequency and amount of precipitation
high mean annual temperature

J
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

Pptn: mm

40

233

982

1059

577

267

206

71

18

3467

J
I

112

J:

AREAS SUITEOTO FLUSHING

Table 5.8.

}Vadi Haifa, Sudan, 221V, elevation 160117.


Jan.

I Feb.

Mar.

Apr.

May

June

July
32

33

Temp: C

15

17

21

26

31

32

Pptn: mm

, 1

Aug.

I Sept. I Oct.

Nov.

30

28

22

I Dec.
I
I

Yr

17

35

, 3

high monthly temperatures ... .


e high diurnal temperature rahg~s
e high wind velocities.
. .
"'-.=-
.J

~,r': ~

.'

_" ......-

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 29-35C
and the average diurnal rangejs :}Jetween 14-25C. A typical monthly distribution
of temperature and rainfall is tsiven in Table 5.8.

Mid-latitude wet
"Koppen classification Cf a#d~f.~: .
.;, :."}~.7:' :'
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 21-26C. 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; .. . . . .
Precipitation is fairly evenly distributedthr~ughog"t~~~!JJ~ year with .variable
annual totals depending on the latitude and the coiltitIentalposition of the regions.
A typical montpjy distributi6n. 'o(~~mperature am(r9:lrifall is given in Table 5.9.
IY1jd-latitude winter dry .
o

Koppen classification C\vand Dw.


Maritime tropical air iflasse's in summer and continental polar air masses in
winter.

This region is charactelised by a strong seasonal pattern in temperature and


precipitation. The region is located in the interior of the continents in the midEngland, 51 5N, elevation 5 m

Table 5.9.

I Jan .
Temp : cC
Pptn: rnm

I
[

Feb.

54

40

I Mar. I APr.'
,
,

' 37

38

I
I

May

I June

12

16

46

46

I July
i

18

56

I
I
I
I

Aug.

17

59

I Sept. I oct./ Nov. I Dec. I Yr


I
,

15
50

I
I

11

10

48

, 5"95

57

64

113

EVACUATION OF SEDIMENTS

Table 5.10.
1

Calgary, Canada, 5JON, elevation


Jan.

Mar.

Feb.

Apr.

Temp: C

-10

-9

-4

17 I 20

, 26

35

i
I

Pptn: mm

i
i
:

10

52

July i Aug.

June

May

32J m
Sept.

Oct.

Nov.

Dec.

Yr

-7

15

444

13

17

15

11

-2

88

58

! 59

35

23

16

latitudes. The continental location causes a large annual temperature range.


Summers are hot and humid with intense summer convectional storms. Continental
polar air masses are associated with cold, dry! weather in the winter. A typical
monthly distribution of temperature and rainfall lis given in Table 5.10.

Mid-latitude summer dry (Mediterranean climate) .


Koppen classification Cs and Ds.
i
Sunlmer dominated by continental tropical alr, winter dominated by maritime
polar air masses.
Found on the western sides of continents between 30 and 40 o N. Precipitation falls
mainly in the winter due to the mid-latitude cyclone. During the summer these
areas are influenced by stable subtropical highsi, producing dry, warm weather. A
typical monthly distribution of temperature and rinfall is given in Table 5.11.

Polar wet and dry

Koppen classification ET.


I
Maritime polar in summer and continental polar or Arctic in winter.
i

Cold winters, cool summers with a summer rainfall regime. The areas experiencing
this climate are the North American Arctic coast, Iceland, coastal Greenland, the
Arctic coast of Europe and Asia and the Sou~hern Hemisphere islands. Annual
precipitation is less than 250 mm with precipit4tion during the summer. A typical
monthly distribution of temperature andrainfal~ is given in Table 5.12.

Polar desert
Koppen classification EE
I
Continental Arctic and continental polar air masses.
These regions occur in continental areas of thd high-latitudes, such as Greenland
and Antarctica. No solar radiation is received for about half the year while during
I

Table 5.11.

Italy, 42N, elevation 131 m

I Jan.
I
I

Temp: C

I Julyl

Feb.

Mar.

Apr.

May

10

13

17

22 I 24

88 I 77

72

63

48

June

Aug.

I Sept. Ii Oct.
i

Nov.

Dec.

Yr

24 , 21

16

12

15

22

128

116

106

881

Pptn: mm

76

114

14

70

AREAS SUlTEDTO FLUSHING

Table 5.12.

Greenland, 815N, elevation 35 m

Temp: C
Pptn: mm

I Oct.

Nov., Dec.

Yr

-8

1-

19

-24

I -26

-16

21

16

35

204

I Feb. I Mar.

Apr.

May

June

July

Aug.

Sept.

-30 /-30

-33

-23

-11

12

19

Jan.

23

20

37

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 OC. A typical monthly distribution of temperature and rainfall is given in
Table 5.13.

. 5.2.4. Relationship between climate zone and erosion rates


A map of climatic regions based on Koppen is presented in Figure 5.8.
Jansson (1988) took sediment yield data from 1358 drainage basins and
cOlTelated sediment yields with the Koppen climate classifications.
For each Koppen classification the values of sediment yield have high standard
.
deviations due to the strong influence of a few extremely high yields.

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

I.

Table 5.13.

Antarctica, 66,5 oS, elevation 30 m

I Jan.
Temp:

Pptn: mm

[ Feb., Mar.

I APr.

, May

I June i July i Aug.

I -2 I -5 1-10 1-14 I -16 1-16 1-17


I. 13 I 19 I 51 ! 44 ! 92 i 67 I 77

Sept.

I Oct. I Nov. I Dec.,

Yr

-17

-17

-14

-7

-3

1-12

I 95

52

43

46

26

I 625

115

EVACUATION OF SEDIMENTS

~----------~~~--------------------------------~~~~

Figure 5.B.

A Tropical rain climates


f no dry period
w dry winter
s dry summer
m monsoon rains, dry period

C Warm temperate rain climates


a warmest month >22
b at least 4 months >10
c 1-4 months> 10
d coldest month below -38

B Dry climates
s steppe
w desert
h hot, mean temp. >18
k cold, mean temp. <18

D Boreal climates

"

E Snow climates
T tundra
F always frost

Climates of the world according to the Koppen classification

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 35 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

I
Af,Cf

Climatic
group

Reasons for grouping

Af

Cf is found in tropical mountains. Similar precipitation


conditions as Af

AW,Am, Cw,
outside Argentina

Aw

Cw is found in tropical mountains. Cw (Argentina excluded)


has a median yield of 246 rll<m2 and a mean of 446 tlkm"

BSh,BSk

BS

Similar rain conditions. Not many low sediment yield values in


BSh. Many low and many high values in BSk

BW

Similar rain and sediment yield conditions

I
I
I
BWh, BWk
Cfa

-.

Few low values

Cfa

cefb in contrast has many low values)

Ctb, Cfc

Ctb

Df

Df

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

Ds

Dsb and Ds are found in mountainous areas in Csa or Csb but


have more snowmelt run-off than Csa or Csb

Cfc similar climate as Cfu


Mountains in Cfa or CfD but Df climates have more snowmelt
erosion than Cfa or etb

.-

Cwa, Cw in
Argentina

Dfa, Dfo, Dfe, Dfd

Cwa

Cw in Argentina has similar rain conditions as Cwa in


Argentina. Cwb not included because rain in warm month may
be of importance and few data from only one country

Dfa-d

All have snowmelt run-off in spring when vegetation cover is


sparse. Temperature similar to snowmelt erosion

Dwe

Similar snO\vmelt and rainfall conditions. Dwa and Dwb hU"'e


different rainfall and snowmelt conditions from Dwc and Dwd

D we. D\vd

ET, ET (Mt)

__::"._0

I
I

ET

Similar climate

has about 450/0 within the two lowest classes. Low values are found in Central
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

Sediment yield classes: Vkm 2

c:::J

0-10

CJ

11-50
51-100

h',~; :J,hl

101-500
Snow

501-1000
_

>1000

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)
Climate zone

Country (nLlmber of rivers/zone)

Albania (312)
Cfb/Csa
Algeria (27/1)
CsafBSk
Argentina (22/13/5/5/2)
BSkiCwafCw/ET (Mt)IBSh
Australia (4/3/1/1/1/1)
CsbfBSh/Csa/CfbIBWhiAm
Austria (4/3)
Df/Cfb
Bolivia (3)
Cwa
Brazil (12)
Aw
Bulgaria (24/6/6)
Cfb/CfafCsa
Cameroon (311)
AwlBSh
Canada (49/18/2)
DfblDfclBSk
Chad (4/1)
-- ~
AwlBSh
Chile (2112)
ET (Mt)/Cs
China (1015/4/3/1/1)
BSkJCfaIDwa/Dwb/Cwa/BWk
Colombia (1512)
Cf/Aw
Costa Rica (10/6/1)
Awl Af/Cf
Cuba (1)
Aw
Czechoslovakia (7/5)
Cfb/Dfb
Denmark (2)
Cfb
Ecuador (1012)
Cw/Aw
El Salvador (2)
Aw
Finland (16/3)
DfclDfo
France (4)
Cfb
Germany (55)
Cfo
Great Britain (9)
Ctb
Greece (414/1)
Csa/Cfb/Cfa
Haiti (1)
Aw
Honduras (1)
Cw
Hungary (6/2)
Cfb/Cfa
Iceland (8/1)
ET/Cfc
India (5/5/2/1/111)
CwalAwfBSfET (Mt)/CwfBSh
Iran (6/2/1)
Csa/BSk./BWh
Iraq (1)
BWh
Israel (2/1)
CsaIBSh
Italy (26114/811)
CfalCs;?}CfbIDf
Java (6)
Af
Kenya (611)
CwaIBSh
Lesotho (9/5)
Df/Cfn
Madagascar (1)
BSh
Malaysia (3)
Af
Morocco (14/2)
CsaIBSk
Nepal (1)
Cwa
New Zealand (42)
Cfb
BWh
Niger (1)
Nigeria (11n)
BShlAw
Pakistan (1)
Ds
Panama (2)
Af
Papua New Guinea (3/2)
Af/Cf
Peru (2)
Cw
Philippines (5)
Am
Poland (53/28)
Dfb/Cfb
Romania (50/412)
Dfb/Ctb/Cfa
South Africa (17/8/7/4/2/1/111)
I' BSklCwb/CtbfBWklBWhlCsb/CfaJBSh
Soviet .(55/351221 IS!) 41 121 10/9/5/4/3/2/111) , DfblDfclD wcfDflDs/BSklETICfaJDwblDwdIDfaJCslDfdlCtb

119

EVACUATION OF SEDIMENTS

Table 5.15.

continued

Country (number of riverslzone)

Climate zone

Spain (18/3)
Sweden (14/3/2)
Switzerland (9/5)
Taiwan (16111)
Tanzania (1)
Thailand (25/13111)
Trinidad (1)
Tunisia (6)
Turkey (1)
USA (83116/14/12/916/4/212)
USA-Alaska (7/5)
Venezuela (4/312)
Yugoslavia (28/311)
Zimbabwe (111)

CsaJCfb

- 1

Dfc/CfblDfb
Df/Cfb

CwaJCfa
Bsh
Aw/CwaJAm

BS
Csa
Ds
CfalBSk/CsblDfblDfaJDflDsb/CfblB sh

DfclET
Cf/AwIBsh
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
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
country provides a manageable unit for some purposes.

1
1
,

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.

5.3.2. Factors affecting erosion rates


Factors causing high erosion rates have been outlined earlier in this book and the
main aspects are summarised below.

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.

Drainage basin area


There is generally an inverse relationsrLip between sediment yield per unit area
and catchment area. Higher rates occur in small drainage basins due to a higher
overall slope, higher percentage of erodible material a.lld less opportunity for
eroded material to be deposited further down the catchment.

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

5.3.3. Sediment delivery ratio


Only a proportion of the sediment that is eroded will be transported down the
catchment to be deposited in a reservoir. The efficiency of the transport process
is expressed by the sediment delivery ratio, SDR, which is the proportion of
sediment eroded from the land that is discharged into rivers (Morgan and
Davidson, 1986).
This measure is required to convert the estimated soil erosion within a basin
into a value of sediment yield. Values vary from about 3 to 90%, decreasing with
greater basin area and lower average slope (Morgan and Davidson, 1986). There
are a number of factors that affect the sediment delivery ratio from a basin. The
size of the drainage basin has an influence on the sediment delivery ratio with
more opportunity for deposition and lower overall slopes in larger drainage
basins leading to lower ratios. The following factors also influence the sediment
delivery ratio.

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.
Distance from basin outlet
Channel networks with a high drainage density are more efficient for transporting
sediment than basins that have a low channel density, meandering low gradient
channels, or those clogged with debris.
Soil and vegetation
Finer particles are transported more easily than coarse particles, therefore higher
delivery ratios occur for soils with fine grained erosion particles. However, silts
tend to be more erosive and produce higher delivery ratios than clays.
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.

Catchment size and slope


Large, gently sloping catchments will have lower delivery ratios than smaller and
steeper catchments.
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

5.3.4. Hydrological characteristics


In addition to the factors causing a high rate of sediment inflow, discussed above,
there are some specific hydrological characteristics of the catchment above a
reservoir site that are required for successful flushing.
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 referred to in the Koppen classification as
'tropical wet and dry'.
River discharges must also be sufficient to transport sediment loads through
the reservoir. Regions of low precipitation, like the Sahara and other desert
environments, will therefore 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. As stated in Chapter 3, 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.
Geographical regions suitable for successful flushing must, therefore, provide
a large enough annual run-off compared with the volume of a reservoir to allow
use of a sufficient proportion of the water for 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:
parts of Central America extending into South America
o areas in North and South America where the rivers are fed by the Rockies and
the Andes
Q
parts of Central Africa from the Ivory Coast in the west to Sudan in the east
19 areas in Central Asia where the rivers are fed by the Himalayas, including
Pakistan, India and Nepal
" parts of Asia including Cambodia, V-ietnam and Thailand.
o

123

Si
e

In

.m:i

pecific

sti

d sIgn

an
era ons

: .' ,

Site-specific investigations and design


considerations
Chapters 2 to 5 inclusive are concerned with general issues relating to the
flushing of sediment from reservoirs. Chapter 2 assesses 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 volume of storage
that is likely to be required to meet continuing demand. Chapter 3 provides a
review of the current state of knowledge of reservoir flushing, and Chapter 4
considers the worldwide experience of flushing. Chapter 5 identifies areas of the
world where flushing is likely to be most useful.
Thus far it is possible to identify locations where sediment flushing is likely to
be useful. However, there are many detailed factors which need to be evaluated
on a site-specific basis before the technical viability and economic soundness of
sediment flushing C8..s.'1 be confirmed. This final chapter provides details of the
nature of these site-specific" investigations, including design considerations for
the sediment bypass itself.
There are numerous stages for such investigations as follows.

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

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
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
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
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

128

~-

-/

:4
J

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:
measures to reduce the amount of sediments entering reservoirs such as
catchment conservation or upstream storage
measures to manage the sedimentation process within reservoirs such as
variations in the operating rule curves for the reservoir
measures to evacuate sediment from the reservoir including dredging and
sediment flushing.

Q)

System simulation modelling


System simulation modelling is required to evaluate the conflicting demands of
hydropower production, irrigation and other requirements, and must be able to
assess the impacts of the various reservoir operating strategies. The simulation
model must be able to replicate the outputs of water and power under a range of
operating strategies so that an optimal economic and technical solution may be
identified. In addition, it must be possible to take account of the effects of other
reservoirs upstream and downstream of the one under consideration.
Economic and financial analysis
The main aim of economic and financial analyses is to assist in the identification
and selection of the most favourable sediment management option. For each
option the most important factor~ from the economic viewpoint, is to define the
'with' and 'without' project cases. These will illustrate the net economic impact
of t1.e availability of water resources over time, including any seasonal variations.
Evaluation of the impact of alternative investment phasing is also important.
The greatest challenge in the evaluation of projects \vhich promote
sustainability of reservoirs is to assign realistic values to the benefits of extending
reservoir life. This is beyond the scope of this study. Work, however, is
progressing in this direction (Palrnieri, 1998).

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.
efficiency of reservoirs. Transactions of the A,m.erican Geophysical
Bnme, G. 11. (1953).
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

Chen, J. L. and Zhao, K. Y. (1992). Sediment management in Nanqin reservoir. International


Journal of Sediment Research, 7, No 3.
Comite Francrais des Grands BalTages (CFGB) (1973 and 1982). EDF-CNEH, Savoie Technolac,
73373 Le Bourget-Du-Lac, CEDEX
Chang, M., Roth, F. A. and Hunt, E. V. (1982). Sediment production under various forest-site
conditions. In: Recent developments in the explanation and prediction of erosion and sediment
yield, D. E. Walling (ed.). International Association of Hydrological Sciences, Publication No.
137, pp. 13-22.
Chaudhry, M. R. (1982). Flushing operations of Warsak reservoir sediment. Proceedings of the
Pakistan Engineering Congress, 58, Paper No. 454,Lahore.
Church, M. A. and Slaymaker, H. O. (1989). Disequiliblium of Holocene sediment yield in
glaciated British Columbia. Nature, 337, 452---454.
Cyberski. (1973). Accumulation of debris in water storage reservoirs of central Europe.
Dawans, P. H., Charpie, 1., Giezendanner, W. and Rufwenacht, H. P. (1982). Le Degravement de
la Retenue de Gebidem. 14th Congress on Large Dams, ICOLD, Rio de Janeiro, 1982.
Dedkov, A. P. and Moszherin, V. 1. (1992). Erosion and sediment yield in mountain regions of
the world. In: Erosion, debris flows and environment in mountain regions, Proceedings of the
Chengdu Symposium, July 1992, IAHS Publication No. 209, pp. 29-36.
Dendy, F. E. and Bolton, G. C. (1976). Sediment yield runoff drainage area relationships in the
US. Journal of Soil and Water Conservation, 31, No.6, 264-266.
Douglas,!. (1973). Rates of denudation in selected small catchments in Eastern Australia.
University of Hull, Occasional Papers in Geography, No. 21, 127 pp.
Dunne, T. (1979). Sediment yield and land use in tropical catchments. Journal of Hydrology, No.
42,281-300.
Dunne, Dietrich and Brunegngo. (1979). Rapid evaluation of soil erosion and soil lifespan in
grazing lands of Kenya, hydrology of areas of low precipitation. IAHS publication 128,
proceedings. of conference, Canberra, Dec.
El Faith Saad, A. (1980). Sedimentation and flushing operations of Roserires, Sennar and
Khashm El Girba reservoirs. International Seminar of Experts on Reservoir Desiltation,
Communication No.2. Tunis, 1980.
El Hag Tayeb (1980). The limited experience of desilting in Sudan. International Seminar of
Experts on Reservoir Desiltation. Com. 14. Tunis.
Fournier, E (1960). Climate and erosion. P. U. E, Paris, 201 pp.
Fredriksen, R. L. (1970). Erosion and sedimentation following road construction and timber
harvest on unstable soils in three small Western Oregon watersheds. US Forest Service
Research Paper. PNW 104, 15 pp.
Gleich, P. H. (ed.). (1993). Water in crisis. Oxford University Press, New York.
Glymph, M. (1973). Summary: sedimentation of reservoirs.
Gogus, M. and Yener. (1997). Estimation of sediment yield rates of reservoirs in Turkey. ICOLD
19th Congress, Florence, Vol. 3, Question 74, Reply 78.
Goldman, S. J., Jackson, K. and Bursztynsky, T. A. (1986). Erosion and sediment control
handbook. McGraw-Hill, New York.
Goldsmith, E. and Hildyard, N. (eds). (1984). Environmental and social effects of large dams.
Vols. 1-3, Wadebridge Ecological Centre, UK.

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.
International, 10, 7-13.

~Vater

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.
172-176,

~Vater

International, 9,

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.
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
l!te~. States Engineering District, St Paul, Minnesota, US.

I'
~.

Langbein, W. B. and Schumm, S. A. (1958). Yield of sediment in relation to mean annual


. precipitation. Transactions of the American Geophysical Union, 39, No.6, 1076-1084.

,
!

Li, Y. H. (1976). Denudation of Taiwan Island since the Pliocene Epoch. Geology, 4, 105-107.

.~

Liechi, W. and Haeberli, W. (1970). Les Sedimentations dans Ie Bassin de Compensation de


Palagnedra et les Dispositions plises pour Ie Deblaiement des Alluvions. 10th Congress on
Large Dams, ICOLD, Montreal.
Lvovich, M. I., Karasik, G. Ya, Bratseva, N. L., Medvedeva, G. P. and Maleshko A. V. (1991).
Contemporary intensity of the world land intracontinental erosion. USSR Academy of
Sciences, Moscow.
Mahmood, K. (1987). Reservoir sedimentation: impact, extent and mitigation. World Bank
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.

J
J

I
!

,-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
r

Mou, J. and Meng, Q. (1980). Sedilnent delivery ratio as used in the ' computation of the
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

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 palaeohydrology, 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 Benwen (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
I

8. Bibliography

1
!

!
i

Abdelhadi, :M. L. (1995). Environment and socio-economic impacts of erosion and sedimentation 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 perspective. Water Intemational, 6, No.4, 219-224

a global

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 Ministry of Development Cooperation, The Netherlands.

a literature review.

Dendy, F. E. (1967). Sediment deposition in reservoirs in the United States. ARS #41-137, US Department of Agriculture.

Paper

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

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 SouthEast 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
systems. Journal of the Chartered Institute for Water and Environmental Management, Feb.

146

-f

".t.,","
;,

.1

BIBLIOGRAPHY

,,~

I1
f

Palmieri, A., Shah, F. and Dinar, A. (1998). Reservoir sedimentation and the sustainable
management of dams. World Congress of Enviromnerztal and Resource Economists, Venice',
June, pp. 23-27.

Partheniades, E. (1965). Erosion and deposition of cohesive soils. Journal of Hydraulics


Division, ASCE, 91, No. HYl, 105-l38.

Patzkowski, L. G. (1992). Rock creek - cresta sediment management project: an overview of


geotechnical and sediment characterization studies. 9th. Annual Dredging and Placer Mining
Conference, Reno, Nevada.

i
i

1
j

I
I

I
~

~
~

Pemberton. E. L. (1978). Reservoir sedimentation. Proceedings of US-Japan Binational Seminar


on Sedimentation and Erosion, Honolulu, Hawaii.
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.
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.

Roehl, J. "V. (1962). Sediment source areas, delivery ratios and influencing morphological
factors. IAHS, Publication No. 59.

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.

!
~

Roosebooffi! A. (1975). Sediment transport in rivers and river


Pretoria, South Africa ..

basins.~

Thesis. University-"of

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

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

Table AI.I.
Region

Country

Notes

Africa
Africa
Africa
Africa
Africa
Aldca
Al'rica
Africa
Africa
Africa
Africa
Africa
Africa
Africa
Africa
Africa
Afiica
Africa
Africa
ArriclI
AfriclI
Afril:a
Africa
Africa
Africa
Africa
Africa
Africa
Africa
Africa
Africa
Africa

Angola
Benin
Botswana
Cameroon
Congo
Congo (ORC)
Ethiopia
Gabon
Ghana
Guinca
Ivory Coast
Kenya
Lesotho
Liberia
Madagascar
Mal awi
Mali
Mlluritius
Mozambiljue
Namibia
Nigeria
Scnegal
Seychelles
Sierra Leone
South AI'rica
Sudan
Swaziland
Tanzania
Togo
Uganda
Zambia
Zimbabwe

Africa

Total

C. Asia

Kazakhstan
Kirghizstan
Tadjikis(an
Uzhekistan

Volume

3
213
967

575352

595

12
11
7
14

89786
23303
28182
6761

7482
2118
4026
483

2
I

c
c

1910--19

1900--9

1920--29

1930--39

1940-49

i950-59

Av. vol No. Volume No. Volume No. Volume No. Volume No. Volume No. Volume

630
9446
1734
867
77
231
14325
1592
44
22
5319
443
2624
328
220
220
150279 37570
237
119
37917
1724
162
2428
1970
281
0
0
431
43
I
4
13 440
6720
61
7
57 102 9517
51
662
40414
898
11520 5760
I
1
22
22
57
30583
5587
1117
42
250
1135
1135
I
I
200
200
47
16
187117
878

IS
2
9
2
12
8
I
4
2
22
15
7
I
10
4
2
9
6
13
45
2
2
I
539
5
6

<1900

Totals
No.

C. Asia
C. Asia
C. Asia

[COLD data - world total storage volumes (M.m 3 )

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

' 0
0

0
1267

0
0

0
I
0

125

1960-69

No.

Volunic

No.

142

44

I
5

2
0
I
I

0
8
2417
7

146
0
36
35
1925

0
0
I
0
0

Volume

7
2

293

<>

ISO 000
237
9
16
3
0
0

2
I
I
0

8
1536
4

3
I
I
6

36
33
295
15079

I
0

1100
16
0

I
2
2

I
2
I

1980-89

1970-79

No.

Volume

Volume

5
I

5124
1710

10915

()

I
0

2
0

14

()

5
0

12
0

11
0

164
0

26
I

679
931

22
I

3534
0

15
0

173
0

413(>
24
85
3410

3
4

0
692

1600

192

220
0

87

14
2
0

28444
170
0

3
9
I

8339
2155
4

0
I
I

0
70
1950

13

125
0
3

16053
0
23

124
0
2

2194
0
177

()

n
0

(,

1134

()

I 1270

I
0
1

I 1270

44
0

5099
0

0
0

41
()

0
183266

2
36

6
1431

51

0
1016

0
82

0
1201.1

0
5

66

6440

164

358850

235

124018

219

46580

132

30900

17

1326

630

56888
723
180
2928

29878
20060

4160

4
4
3
6

6
3

3
2
4

2380
10542
1564

2390
140
13 300

15830

0
0

I
2

41
39

0
2

0
39

0
5

0
56

16

11

164

31

1696

27

4967

26

251

96

144

l'

3
0
I

I
0
0

8
I

2170
4
1273
98
9331
250

I
I

25

0
21

0
4

54260
222
14856

200
0
0

104
2
I
I

0
0

3
6
23

2269

()

C. Asia

Total

44

148032

3364

4790

17

60719

13

52207

14486

China

China

1851

649322

351

15

24057

177

26113

479

256549

583

103536

220

58090

120

57278

260 123667

1851

649322

351

15

24057

177

26113

479

256549

260 123667

2
1

1579
6500
70
48
340
126

3158

37
0
0
4

2101
0
0
267

39
0
0,
2

826
0
0
157

82
0

77

41

2368

41

983

China

Total

S. Asia
S.Asia
S.Asia
S. Asia
S. Asia
S. Asia

Afgh anistan
Danghtdesh
India
Nepal
Pakistan
Sri Lanka

4010
3
69
46

3158
6500
279548
144
23436
5816

S. Asia

Total

4131

318602

- )

45
0
0
I

3287
0
0
101

40
0
0
0

6658
0
0
0

35
0

2446
0
27
87

85

2560

46

3388

40

6658

213
0
0
9

25601
0

550
0
7
0

1329

414
0
19
12

36

557

224

30088

446

I)

o
z

3
I
I
4

707
0

61

I
0
8

0
0

Volumt!

144
0

No.

0
22
1824
4656
50
1135
I

1
2

No.

Volume

No.

I
I

()

ute

1990-1998

583

103 536

220

58090

120

57278

()
0
6500
55938 1128
0
0
7961
16
621
8

57076 1178
3
0
15209
II
527
8

38901
144
184
2727

116
0
21
0

1045
0
48
0

683

85119
0
0
0

71020 1152

72 812 1200

41956

137

1093

683

85119

()

o
"
m

V1

3:
m
Z
-f
V1

.Mid. E"sl
Mid. Easl
Mid. Easl
Mid . 1;",1
Mid.E'lsl
(vlid.E:ISI
Mid. Easl
Mill. Hasl
Ivlid. ['ast
Mid. Fasl
Mid. E'lSI

Aruh.'nill

Iii

Azerbaidj'lIl
CYI'I1IS

17

l.d>:IIIOII
Saudi Ambia
Syri'l
Turl,ey

5
311
41
1i25

I7H9
183K5
2B6
2S!!1
32935
':>7183
203
161
2114
342
109233

'lblal

895

22:161B

5'2
14
66

(iL."Ofgia

11';111
Inil(

JII.-.I:1I1

---'-'- - - - - - - - ---Mid. Easl

- - --

-----.-

N. "f,-il' a
N. Arri.:a
N. Africa
N. A!'rica
N. Afrim

1\lgcri;1

EUypl

Lihya
t\'tllruccCl

"ruilisia

- --.- - - - N. Afri':3
'lllial
_- _--0_._.N. AIII('ri~a Call:llla
N.A llwric'l USA

--

---

N. AlIler;":1 'lhlal
N. ElIr<lp"
N. Ellrllp"
N. Ellmp~
N. EllfllpC
N.Ell rll!,c
N. Ellrupe
N. ElIwpe
N. EUI"llI'"
N.l'lIl"1lp"
N. Europe
N. Eul'Opc
N. Eurup~
N. Eurllpc
N. Eurnpl!
N. Europe
N. EUIll(1c
N. EllrClfX'
N. Europe
N. Europe

Pacilic l~illl
j'al'ifiL: \(jill
P:tdlkRil\l
Pacific Ilim
Parili.: Rim
1"lcilkRill1
P;ll'ilic Rim
Pacilie Rim
Pa.:ilic Rilll
Pa~ilic Rim
P;wilic Rim
Pacilic Ililll

Ln
W

- --

PHeili. Rim 'lillal

---

()

II

()

(I

I)

I)

II

1/

()

175

()

()

2511

3
0

5
0
0

-- --

I)

0
n
II
I0
0

()

II

(I

t)

II

I)

'l

tl

IJ

I)

(I

I)

()

10
132

48
1638

652

5
6115

6T!4

517 670

9
665

47.'i~

- -I
3

()

1929
19 Inn
160SY

()

0
141
15

0
17
14
0
0

0
0

Il

0
0
0

9
14
0

403
0

(I

t)

II

II

II

6
I

1414
331

0
5
6

'2

6104

\0

1911

381

36~63

414

II
0
2
158
92
0
6

I
0
2
35
27
0
3

281

22

9')

19

2142

71

1S1.S

90

5202 153

484
H7
2
9ti
1077
7(-,5
70
86

89855
10 775
153
74348
21 OKS
161161
34312
16515
333
3301
6714
2808

186
124
17
774

33
2

240
10

14
I

627
95

21
6

144
30

21

5102
24

21
20

2050
2
I

1
7
0

17
10
0

3
43
15
50

._-

:ms

- - -- .
277 124

I
26
14
()

I)

1221
15

\I

543
83

14
8
I
2
3

/I
612

0
0
1532
54

30110
5
177

9
15
I
9

II

168

43

68429

\06

6
I
II

259
1620nO
290

2238

31

6423
0

41i
563
1137

0
29
61

473
0
0
5143
199

tl

40
0
4
37
12
93

6')696 349

84923

925 336912 1672

592755 1393

I
0
6
38
9
I
'2
I

25
0
5726

-II
0
21
826

60S
II

654

~ - --

()

906
2\J
31
j

I
'2
14
75
34
Il
4

51:2()
162
201
1859

0
14
2

.5

3
18

10929
3320
74J

II
II

954
276
10281l
27

1006

17

166

24457 117

19631

--

3
5
7639
2755
850
0
307

2
4
16
90
49

-~--.--

t-506874

r--

0
485
35
2994
1811
225700
47
5329
18216
0

3
3

5
7
2953
3998
SKI
\3
0
5\0
462
79

3
101
9
19
4
53
7
58

204
11 765
883
173475
41
7927
19 7~('
38!li!

Y
0
II
93
76
4
2
2
12
0
I
2
75
10
IJ
4
30
6
32

315 264 553

425

226517

382

195444

R6
II

45850
64

16

0
')

0
:2
37
3
20

"

0\7
3
68

t----.

16
5

1187
71

51
7

S'S\()
1608

93
12

121'13
5901

,II
22
9
9

106
255
86
2332
322

4
37
62
7

3(1'l3
832
1361
2112
94

2
137
102
12
\0

128
3504
191
16350
nll5

4
3
8

322
424
198

I
208
140
18
12
I
8
I
6

2556
6444
1138
4316
2907
0
875
1!50
367

336

341)20

5(11)

37,197

52

II

.6

51

II

12

31

57

160

332

I
4

31
12
II

100

53

493

7J

5655 135

~449

142

6(i74

44H

4371

0
2
14

885
12 632

3
I)

II

60

0
39
0
\0 995
248SS
t--

()

[I

II

"I -0
0

122

5815

15
195

1i32
)-125

757 172 306

2111

4057

0
I
3
54
6
4
I

0
1900
72
3133
181
420
5

0
0
0

0
0
0
672
100
0
0

3
2
96
48
7
0

277

3984

132

340
65
0
735
1I2U
9154
724
162265
20
11028
2510
541

116 155495
641 16 MIl

2539
2054
639
1575

(--

264

335

168 972

40

43811

455767
51 \07

2777

221

23

35
151
55
114
161

23

41 178

432092 142
160663 1251

I
2
I

--

122

/0

(- -

3526
459

64 050

9
69

1-17 211691 128


Tl8 125221 1544

0
J2

8
I
0
15
7
151

52

133

0
6

18
13
3

0
76
14
3490

8551
26000

32428
52495

20

3
I
21
4

593
13
231
6
1206
1756
28
21330
80
U
0
165
U
6
18437 295

0
I

--

787
2182
17

8176 55
61520 294

:!I)

22
490
192
III
77

15H

I)

75.)
0
0
595
197

II

II

II
III

165

II

22061

0
482

86

106
0
27
0

7
8
4
3
8
I
I
0
12
6
72

0
0

6
0
0
I
0

(I

131
190
29
350
5568
3053
II
160
I
327
12241

3
I
13
29

'/31
5000
0
373
0

0
'2

- - - - - --- I - : - - - - -

0
2

2
2
0
6

0
0
0
2

56

2298'2

98

(I

21

52Y

(I

()

6800

314

7(1

5
0

0
.5
I
0

0
' :\

5'/

188

3
0

120

18

248

J
I

CiK

7
16
'2
5
3

18

412

()

14

56
117
0
tl

'2
I
I

I
160011
I

-- ---

9JS 168

I)

(I

'1277

35
15
0
3

0
II

0
312

0
1

9
321l
69
91
28
19[1
20
517

.--- - -

2679
994
H66
6"1
79
1220
9259
53496
3598
731 637
206
3:l216
46 B8'1
6864

<In!)

(J

II

()

19:1
3'16
28
14
149
62
173
31
26
6\0
1029
163
52
R040
7
201
234 '1
13

12 12247 85 17679 82
246 III 747 296 19284 332

256 674
5364 142 36R6 28K 30994
._r---'t - - 1-- - --. - . 252
Iii
0
0

0
I
0

II
0

--

-.

t)

1597
81

--._0

()

0
0
0

I 3~6 860

16
5
20
3

UK

'Ihiw;ln

1111

IS

Ukr:linc

I<llr~:llN)

47:1

55
569
311

Sw'-'lIcn

N. Z"'II:md
I'. N.Guillca
l'elll
Phjlipllincs

18~

111

Rllssia

J{orl~il

289

16

Slnwnia

Auslralia
Chik
Fiji
Indonesia
Japan

' -'

4732
48
Ifl71100 27833
3]6
28
I~ 54\)
15~
l11fi5
II

I
I

7205 IM'14530

lkigilll\l
[klllll;1rk
I'illiand
Fr:mc.:
Gl'I"IIHIIIY
kdand
Ireland
Lllvi"
Lillluania
Luscmhllrg
Mllld'lI'ia
Nl'lherlalld,;
Nlww:IY
Polalld

---- [ - . _ - N. Europe TOlal


---.--- - -- --- . -

99
6
12
92
8tl

112
1081
6
20ti
49<)
4399
25
32
7

68
17
77

1322
463
640
()

4
0

4K

1
51
8
5
6
12
2
II

2780
2U944
2RO
10493
52
631
92
4011

0
4044
653
327liU

6
,9

1.
I
3

<)

95

II

I
5
0
0

---II

I
0

46

0
175
236
111805

"

\I

()

II

0
JOUO
0

69

38307

93

43341

92
14
2
27
165
'187
6
19

9808
1222
153
9277
1579
63911
3139
614

36
7

271H
291

0
2

0
10011

27
198
68
3
0

14515
3004
217K
6726
0

\I
56
0

41997
16119

II

690
2425
505

157

52

299

3581U

3115

29888

75

4'1738

._257
-

10.
202
170
12
14
2
5
4

574
3586
5453
1053
8978
33)
1069
3003

I)

1)95

4
6
9

525

70958

531

II

28 116034

II

-0
-U

~
n

Table Ai.i.
Region

c
~

continued

Country

Volume

No.

1920-29

1910-19

1900-9

<1900

TOlals

NOles

1930-39

1940-49

1950-59

Av. vol No. Volume No. Volume No. Volume No. Volume No. Volume No. Volume

No.

1960--69

Volume No.

Volume
------

S. America
S. America
S. America
S. America
S. America
S.America
S. America
S.Americ:n
S.America
S. America
S. America
S.America
S. America
S.America
S. America
S. America
S. America
S.America
S.America
S.Amcrica
S.America
S.America
S.America

Antigua
Argentina
Bolivia
Brazil
Colombia
Costa Rica
Cuba
Dominican R.
Ecuador
EI Salvador
Guatemala
Guyana
Haiti
Honduras
Jamaica
Mexico
Nicaragua
Panama
Paraguay
Surinam
Trinidad
Uruguay
Vene..:uela

5
130380
372
540386
10913
2288
3755
2373
70()3
2430
460
44
I
9035
220
121253
1252
5184
33690
20
48
12335
155467

5
1304
53
922
223
254
77
216
226
486
115
22

1498 I 038913

694

a, d

a, b
d

S. America Total

100
7
586
49
9
49
II
31
5
4
2

I
8
2
536
4
5
4
I
4
5
74

0
0
77
0
0

0
0
1874
0
0

0
0
3
0
0

0
0
179
0
0

0
0
16
0
0

0
0
540
0
0

3
0
35
0
0

32
0
2029
0
0

Albania
Austria
Bosnia
Bulgaria
Croatia
Czech Rep.
Macedonia
Greece
Hungary
Italy
Ponugal
Romania
Slovakia
Spain
Switzerland
Yugoslavia

S. Europe

Total

306
148
25
180
29
118
18
46
15
524
103
246
50
1\87
156
69

117

10

4144

544
0

I
15

92

3220

14
0
100
9

1057
161
1665
0
0
4

2
0

14
3
16
0
0

I
I

0
220
39

20

2658

1
0

624
0

I
0
0

5
0
0

0
I
44

0
0
9678

447
0
11825
250
0
5

17
2
103
10
2
15

0
6
620
0

I
I

0
0

0
0

0
0

0
0

0
0

0
0

0
I

0
3

92

1966

II

296

27

5228

47

2323

0
0
I
0
0

0
0

17706
1841
56477
4094
4360

0
0
53
0
0
I
10
0
0
3
113
1
0

145162

45

85

182

----

62

----

0
7
0
0
0
4
0
2
0
12
5
0

0
0
0
0
0
7
0
0
0
17
0
0
0
57
3
0

0
3
0
0
0
7
0
0
0
16
0
I
0
44
5
0

30

84

76

0
0
0
0

0
35
0
0
0
28
0
0
0
38
0

I
0
656
94

0
4
0
0
0
2
0
0
0
79
I
0

0
41
1I
0

852 138

0
5

0
0
0
3
0
0
0
1375
22
0
0
636
303
0

0
4
0
0
0

10
I
1
0
57
3
2
1
45
10
2

2344 136

I
5

3
4
I

1057
6330
380
460

Volume

No.
~ ~

----

21
2
3

34593
0
85063
85
13

2
3

175
465

()

Volume

----

I
29
3
0

197
54
38298
1999
0
0

0
44

(I

20357

I
2

0
&04

21000

(I

532

0
4

0
3981

58 141283

44

1940

10

7085

133
I

16573
23
4000
12690

65

15293

21

I
42
2
2946
16 142726

0
3

357

8800
345

I
I
3

I
29
0

I
I
16

202

31466

322

147118

0
26

0
3126

2631111 191 350268

o
3:
m
Z

--i

66377
-------

1
30
3
19
5

3922

424

21629

JO

577
712
2289
2145
151
1567
1211
5875
19
2159
1627
2062
611
19353
1703
703

125
26
4
54
4
23
3
3
6
43
15
74
9
195
18
II

3857
700
814
2865
24
729
76
2721
30
634
899
6097
428
7444
343
1502

629

42764

613

29163

73
25
9
78
4
23
8
4
5
77
13
29
13
211
47

98
29
7
23
14

11
2
7
3
33
21
81
7
186
9
28

596
389
343
1034
231
187
19
1246
12
1070
1103
2674
131
7723
17
1620

9
9
I
4
2
5
0
26
0
31
21
20
7
139
2
2

29
94
5
43
0
253
0
2924
0
1273
644
728
258
5037

559

18395

278

0
2
0
0
0
0
2

..

4
0
0
0
0
0

42

0
28
0
34
0
38
0
3

1100
0
1355
0
6120
0
1865
0
143

11330

109

10587

()

o
..."

(.I')

(.I')

2
3

3910 143

II

31
6
22
I
0

9530
18
I
6
60 161247
14
3162
2
4

145
I
2
0

27
0
91
10

13
2
3
0
116
22
2
7
159
32
10

0
65

No.

18282

698
123
2221
608
404
359
447
0
3174
3011
10
408
9093
893
178

0
2
0
4
0
0
0
34
7

Volume

68

0
183
0
16
0
31
0
0
0
882
291
0
0
2073
278
168

22
0
0
0
192
4
41
0
715
31
14
2
2427
459
4

No.

5
79705
0
135220
3821
2220
2957
1141
190
1430
0

34020
425
16
0
20
0
560
4754

~217tO
o

4204
151
98590
1593
51
782
0
12
0
0

Volume

utC

1990-1998

I
0

13
0
0
1
6
0
0
6
52
6
0

~~~~

615
0
3856
3
0
7
0

17
19
143
46
35
29
93
312
4
24
74
72
37
48
26
63

5061
2841
3576
8324
1014
3454
1669
14354
62

3
0
35

-~-

S. Europe
S. Europe
S. Europe
S. Europe
S. Europe
S. Europe
S. Europe
S. Europe
S. Europe
S. Europe
S. Europe
S. Europe
S. Europe
S. Europe
S. Europe
S. Europe

No.

1980--89

-~~~~~~

1129
110
226
313
1037
8423
20
12
2467
2101

1970-79

_"

,.

S.E. A~ia
S.B. A$iu
S.E.Asia

Brunei
C:unhmlia

::i.E. Asia

Millays; ..

S.E.Asi:1

S.E.l\siil

M,YIIIIHlur
Sil1gllllure
Thailand

S.n.Asia

VkWiltll

S.r:. Asia

'2
'2

Lnos

45

2325

22
119
7(nn
491
465

204

75
78534

15
385

16.5

165

277

117 311

424

I
59
5
:I

23'1

7030
28 ')(,0

1blal
Worltllllllli

25432 6464730

I
14

18

2442

1
II
I

28
18536
165

33

21 18(i

I
4

-1-

S.I1.Asia

I
I

'11

254 1208 11308 1355

0
II

()

(J

Il

.,'--

9986

0
0

96

96

()

75

0
I

75

0
'2

44

44

4
I
I
20

27

7030
6657
B27
)0
30976

44

19
3

19720
1423
17
27 ROI

I
135

~~

49005

223

3ri

1150

37

1J73

0
0

---0

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 ~.,c1udesYmyr~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

o
X

EVACUATION OF SEDIMENTS

Table Al.2.
CountTY

Reservoir sedimentation rate data


Reservoir

Catchment:
km 2

Capacity:
M.m'

Survey dales
Stan

India
India
India
India
India
India
India
India
India
India
India
India
India
India
India
India
India
India
India
India
India
India
India
India
India
India
India
India
India
India
India
India
India
India
India
India
India
India
India
India
India
India
India
India
India
India
India
India

Uri
Maharashta - Koyna
Maharashta - Khadakwasala
Maharashta - Yeldari
Maharashta - Ramtek
Maharashta - Dama
Maharashta - Gima
Maharashla - Ekruk
Maharashta - Mhaswad
Maharashta - Visapllr
Maharashl3 - Mangi
Maharashta - Asolmendha
Maharashta - Nalganga
A.P. - Sriramsagar
A.P. - Nizamsagar
A.P. - Himatyasagar
A.P. - Cembum Tank
A.P. - Kaddam
A.P. - Ramappa Ll!ke
A.P. - Lakhamvaram
A.P. - Dindi
A.P. - Palair
A.P. - Manjira
Kanataka - Tungbhadra
Kanataka - Bhadar
Tamilnadu - Lower Bhawani
Tamilnadu - Vaigai
Tamilnadu - Mettur
Tamilnadu - Upper Bhawani
Tamilnadu - Setnur
Tamilnadu - Aliyar
H.P. -Pong
Punjab - Bhakra
Uttar Pradesh - Mamtila
U.P. - Ramganga
U.P. - Dhukwan
Uttar Pradesh - Tehri
Bilhar - Panchet Hill
Bilhar - Maithon
M.P. - Gandhisagar
M .P.-Tawa
GUjrat - Ukai
Gujrat - l(adana
W.B. - Mayurakshi
W.B. - Kangsabati
Orissa - Hirakud
Kerala - Malampuzha
Kerala - Peec.hi

India
China
China
China
China
China
China
China
China
China
China
China
China
China
China
China
China
China
China
China
China
China
China
China
China
China
China
China
China

156

12750
892
507
7362
212
404
4729
412
1243
412
304
246
316
91751
21694
1308
993
2656
184
268
3919
1687
16770
28 180
2435
4200
2253
42200
34
10826
195
12562
56980
20720
3134
21340
5188
1087.8
6294
23025
5980
62224
25520
1860
3790
83395
148
107

2797
110
934
117
237
609
94
87
43
34
93
76
3172
841
108
106
124
82
60
74
57
51
3751
239
933
195
2709
101
235
109
8579
9869
1133
2450
106
3540
1581
1349
7740
3645
8510
1543
608
1135
8105
228
113

Total

591737

78412

Sanmenxia - Yel10w River


Sanmenxia II
Hongshan
Guating - Yongding River
Fenhe
Liujiaxia - Yellow River
Yanqouxia - Yellow River
Qingtongxia - Yellow River
Bapanxia - Yel10w River
Danjian kou
Celian
ZhenziJiang
Naodehai
Gondzui
Bikou
Shimen
Hongsiba
Wangyao
Fengjiashan
Miugong
Dongxia
Shixiakoll
Chang shan tao
WenYllhe
Bajiazui
Yang maowan
Zhaikou
Luhun

688000

16200
9640
2560
2270
721
5720
216
606
49
16000
200
36
168
357
521
105
34
203
389
183

43402
181 800
182700
285000
215900

77
175
348
105
496
120
185
1320

End

TOlal
sedimentation:
M.m'

Vol. lost:
2

1961
1870
1963
1914
1910
1965
1871
1888
1902
1957
1918
1963
1970
1930
1927
1956
1958
1919
1909
1943
1928
1966
1953
1963
1953
1958
1934
1965
1957
1962
1974
1958
1956
1974
1907

1986
1940
1983
1987
1941
1979
1991
1990
1988
1989
1987
1985
1984
1975
1976
1978
1977
1975
1975
1976
1977
1977
1985
1974
1983
1983
1984
1985
1982
1981
1986
1987
1990
1986
1980

175
23-9
841
145
21
495
264
45 3
296
33
:27-0
42
7949
5336
286
21
459
2g
20
22
11
187
5882
310
378
22-4
5283
36
27 6
31
422J
9158
2735
970
472

06%
21 7%
90%
12-4%
09%
81%
281%
520%
688%
96%
290%
55%
251%
635%
265%
20%
371%
34%
33%
30%
20%
367%
157%
13 0%
41%
115%
195%
36%
118%
29%
49%
93%
241%
40%
445%

1956
1955
1960
1974
1972
1977
1955
1965
1957
1955
1957

1985
1979
1976
1980
1984
1984
1975
1972
1984
1977
1982

1852
1554
3591
96
5412
702
620
57
13349
80
254

117%
115%
46%
03%
6-4%
46%
102%
05%
165%
35%
225%

7514

9-6%

1960
1960
1960
1953
1960
1968
1958
1967
1975
1968
1960
1959
1963
1967
1976
1973
1960
1972
1971
1960
1959
1959
1960
1959
1958
1970
1970
1960

1978
1989
1987
1994
1989
1989
1968
1980

5450
5690
670
630
330
1410
J61
566
15
1130
205
29
2
206
218
28
7
77
63
97
41
35
47
20
249
17

336%
590%
262%
278%
458%
247%
745%
93-4%
306%
71%
1025%
806%.
12%
577%
418%
267%
206%
379%
162%
530%
532%
200%
135%
190%
502%
142%
43%
47%

N/A

1986
1983
1973
1986
1987
1986
1988
1986
1990
1990
1989
1983
1988
1986
1988
1990
1990
1990
1983

8
62

Annual sedimentation

Notes

%
M.m'/ann

550
784
673
571
938
166
747
534
357
835
336
1588
601
619
547
446
96
910
267
112
14
102
652
1158
300
398
250
5347
102
838
2800
554
388
2580
30
1400
587
1029
975
268
725
393
1667
216
593
2465
9505

701
070
034
4 20
020
007
353
022
044
034
010
039
019
5678
11 86
058
010
242
005
003
007
002
170
1838
282
126
090
1057
018
110
016
3517
31 58
804
809
065
726
639
648
2245
160
45 10
1003
310
082
4944
036
\02

003%
031%
045%
017%
003%
058%
023%
051%
080%
030%
042%
025%
179%
141%
054%
009%
195%
006%
005%
009%
004%
334%
049%
118%
014%
046%
039%
018%
047%
015%
041%
032%
071%
033%
061%
021%
040%
048%
029%
004%
053%
065%
051%
007%
061%
016%
090%

604

35729

0-46%

4401

30278
19621
2481
1537
1138
6714
1610
4354

187%
204%
097%
068%
158%
117%
7-45%
718%

6278
891
207
009
1030
2180
)87
027
428
332
334
171
121
IS1
069
778
085
040
270

039%
446%
575%
005%
289%
418%
178%
079%
211%
085%
183%
222%
069%
052%
066%
157%
071%
022%
020'11.

17

3540
3693
881
1528

Mt/ann

% vol.

m'/km /yr

N/A

j
i
J
J

APPENDIX I

Table Al.2.

continued

Country

Re~ervoir

c~,_,
Ic",,""
I
km z
t.Lm}'

Survey dates
Start

I 42804

China

Total

I
I

End

ChilUl
China

Total-on3!) reservoirs
1981
80986
83 357 reservoirs with 460000 M.m' storage willI an aver.lge annual loss af 23%

Netherlands

HlIringvUet

Romania
Romania
Romania
Romania
Romania
Romania
Romania
Romania
Romania
Romania
Romania
Romani:!
Romania
Romania
Romania
Romania
Romania
Romania

Pangamli
V:lduri
Balea

I Tumu

21
11
21
62

Total
K...pelilny
Hrieov
Nosice

Slovakia

Total

Japan

29 dams
13.5 d:lms
169 dams
15S dams
152 dams

J.:lI:l!ln

207

S
8
36

Spain
Spain
Saain
Spain
Spain
Spain
Spain
Spain
SP:1!tl

202
26
149
123
570
308
138
86

23

, T:lb'lls - Buendia
; ucar - Alcon
Ebro - La Tranqu:::ra
Sur - R~(Ie"'o.do
CaL:1.lonia -'"Riudecanll5
S<!gura - Camarillas
Norte - Rioseco
I

049

094%

075
360
450
1104
1431
132
096
108

017%
016%
015%
0-22%
025%
022%
038%
051%

3755

22-4%

6-3%

Tot;u-Gains

Spain
Spain
Spain
Spain
Sp;z.in

284%

\.,5

247%

69%

2737

Spain

14969

1870
579

1520
2
84
2
3
35
4
7
4

Tagus
Ta!<u$ - Palmaces
Tagus - El Vado
Jlicar - Amadorio
Ebro - Arguis

275
379

Jucar - .~eIlQS
Iuco.r - Bus(!o
Ebro G:tl!iu~uen
luco.r - A:ouillo de Dan Bias
NOlie -Alfilorios
Gu:!d:llqtlivi~ - G'.lado.lmena

220

32
57
17
3
1311
8
4-

1955
1942
1965
1973
1969
1984
1956
1962
1954
1909

557-9
647
569

-92742
-10835
--4962
-403
-3724
-3478
-2014
-2
-In':::
-0504
-t:!6

1957
1958
1960
1974
19i5
1960
1973
1959
1948
1954
1972

1960
19:.8
1979
1912

1927

330
690

H,)

1973

2:2

1350

3J..7

1960
1990
1969

1984
1995
1993
199[
1977
1990
1985
1%9
1990
1983

324
855
3201

198J
1976
1994

1992
1981
1993
1994
1979
1968
1984
1979
1991
1980
1994
19S0
1979
[981
1988
1994
1 1989

0
0()75

0121
0154

oun
0189
0285
0326
0561
0629

0691
0723
07:58

014
008
028

166% 1
091%
077%

4500

034%

-22%

-047
-O5S
-007
-019
-005
-001

-469'0

-526 j

-0\9%

-320
-020
-018

~7*

-Q~:!2

-83%
-5-4%
-30%
-06%
-07%
-14%
-2~1%

000
000
000
001
000
0-01
(}O2
002
003
002
Q.1O
002
()O2
005
001
002
0-1 I

00%
38%
OI%8-6%
56%
0-5%
6-7o/c
47%

12-8%
20%
12%
44%
25~3~c

0. 793

06%

0-807
0-S4t
e9!
0965
0994

\0-1%
210%
9-1%
44%
106%

H),

03%

0-220/0

-029%
-0[0%
-067%
-015%
-038%
-010%
-002%
-021%
-006%
-003%

-IS8%

I 321%

6,65

32-4%
581%

825

\994
1994
1994
19941994

4825
2089
8055

t3 200
1112

Spain

1994
1994
1994

6707

425 reservoirs studied in 1979


Jucar - Alarcon
Guadatquillir - EI Pint:ldo
S~ra - Sllntomera
Sur - Guadalteba
Guadalquivir - Ameen!!
Guadalquivir - NegrJ.tin
Douro - Barrio. de Luna
Guudaiquivir - Los Hurones
l'ugus - i'lorbollon
Guadian:l - Gasset

Spain

1991193
i98!/90
1971180
1961170
1951160
1941/50
1931/40
tol930

1-3.5
052
096
043
063
[.63

Total 729 clams

Spain
Souin
Spain
Spain
Soain
Spain
SOllin

Spain

1992
1989
1992

39
691
1305

256%
41-2%
440%
535%
358%
323%
210%

683

Jupan

Spain
Spain

1957
1962
1963

12
0132
H
54
:5-15
1147

011
OCO
021
0C3
005
007
007
007
005
016

[ISH

Japan

Spain

214%
069%
/-17%
031%
241%
)50%
064%
120%'
3-93%
4<12%
4<88%
197%
1030%
440%
446%
398%
294%
262%

i7 322

23i7
2942
4911
5672
599
251
[88

27 dams

Spain

024%

03%
14%
2-9%
65%
98%
108%
227%
365%

35 dams
24 dams

Spain

53
443

JaPan
Jacun
Japan
Japan
lapan
Japan

Sp:!in

241
014
003

10%

Ramnieu
Raureni
Govan
' Babeni

Romania

099%

493%
110%
280%
75%
505%
255%
lQ.8%
241%
746%
74-l%
781%

Slovnkiil
Slovakia
Slovakia

800CO

70

Dnesti

1/9%

142%

33
055
28
003
437
13
08
13
132

% vol.

! 1500

1987
1981
19&7
1987
1986

1982
1983
1986
1986
1986
1986
j
1986
19S~ 1 1986
1976 1986
1974 1986
1977 1986
1975 1986
\978 1986

Notes

Mtlann

51071

1999

I~I

~edimcntJtion

M.mJ/a,.'l11

I 281%

\970

1966
1966
\967
1968
1970
1973

Annuul
rn"lkm!/yr

12013

1964
1965
1963
1963
1965

10
0
9
.5
7
5
2
2
I
\3 j
IJ tI

Garleni
Lilteci
Bucau
Oiesti
Cerbureni

I",

.5

Vanarori
RaCOV!

1 Vol.

1000
I

T~,'

sedimcntation:
M.m"

000%
O':H%
0-00%

0480/"
009%
002%
042%
023%

064%
007%
017%
014%
0-60%
0-04%
015%
0-40%

0-03

1-14%
0:6%

025
005

002%

2~%

157

J
EVACUATION OF SEDIMENTS

]
Table Al.2.
Country

Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain
Spain

continued
Reservoir

Catalonia - Foix
Guadalquivir - Gergal
Guadalquivir - Bennejales
Jucar - Forata
Segura - Argos
Ebro - Santa Maria de Belsue
Segura - Anchuricas
Sur - La Vinuela
bro - Santolea
Ebro - Moneva
Guadiana - Valuengo
Jucar - La Toba
Guadalquivir - CubiJIas
Segura - La Cierva
Catalonia - San Pons
Douro - Burgomillodo
Jucar - Sichar
Guadalquivir - La Bolera
Douro - Linares del Arroyo
Jucar - Guadalesl
Guadiana - Torre de Abraham
Ebro - Las Torcas
Guadalquivir - Cala
Ebro- Pena
Guadalquivir - La Minilla
Guadalquivir - El Tranco de Beas
Jucar - Maria Cristina
Norte - Penarubia
Jucar - Beniarres
Sur - Guadalhorce
Tagus - CazaJegas
Guadalquivir - Bernbezar
Guadalquivir - Puente Nuevo
Tagus - Guajarez
Tagus - Entrepenas
Guadalquivir - Torre del Aguilla
Segura - Cenajo
DOUTO - Agueda
Ebro - Cueva Foradada
Jucar - Benageber
Guadiana - Zujar
Ebro - La Estanca de AJcaniz
Ebro - Sotonera
Jucar - Ernbarcaderos
Catalonia - Sau
Guadalquivir - Guadelen
Tagus - EI Burguillo
Sur - Conde del Guadalhorce
Guadalquivir - Guadalcacin
Ebro - Ribaroja
Segura - Talave
Segura - Valdeinlierno
Tagus - Gru;el y Galan
Douro - Santa Teresa
Tagus - Riosequillo
Guadalquivir- Bornos
Ebro-Oliena
Guadalquivir - La Brena
Guadalquivir - Guadalmellato
Guadalquivir - Pedro Marin
. Segura - Alfonso XII
Segura - Puentes
Jucar - Contreras
Ebro- Yesa
Guadalquivir - Dona Aldonza
Tagus - San Juan
Ebro - Barasona
Segura - La Fuensanta
Ebro - Talarn - Tremp
Ebro - Mequinenza
Guadiana - Cijara

Spain

Total-Loss

Turkey
Turkey
Turkey
Turkey
Turkey
Turkey

Caygoren
Demirkopru
Buldan
Kerner
Yalvac
Kararnanli

158

Catchment:
km 2

300
1058
190
117
1221

737
167
800
187
760
60
761

64
965
1334
469
965
1955
1665
559
375
439
1060
644
3495

323
16952
1281
268

763
31 I
1846
1858
1361
2694
465
1195
420
852
1042
2344
2181
3766
863
1250
1201

7456

Capacity:
M.m;

6
36
104
39
12
13
8
170
49
10
20
II
21
8
25
15
52
56
58
16
60
9
59
22
60
500
23
12
31
134
11
347
289
25
8911
70
472
22
29
228
309
14
189
9
177
173
209
78
77
219
47
25
924
496
49
215
101
116
163
19
42
32
872
471

23162
71
235
258
1530
1670

Start

End

Total
sedimentation:
M.mJ

1928
1979
1958
1969
1970
1931
1957
1986
1932
1939
1959
1935
1956
1929
1957
1953
1960
1967
1951
1965
1974
1946
1927
1930
1956
1945
1920
1961
1971
1972
1949
1963
1972
1971
1956
1947
1960
1931
1926
1955
1989
1944
1963
1952
1963
1954
1931
1921
1917
1969
1918
1897
1961
1960
1956
1961
1959
1935
1965
1954
1916
1884
1975
1960
1955
1955
1932
1933
1916
1966
1956

1983
1985
1978
1983
1991
1980
1979
1994
1993
1984
1985
1980
1990
1987
1968
1989
1976
1979
1980
1989
1988
1979
1984
1989
1984
1990
1991
1994
1991
1991
1990
1994
1994
1982
1979
1992
1992
1980
1992
1992
1994
1971
1986
1983
1979
1977
1991
1991
1969
1982
1993
1984
1990
1989
1970
1990
1985
1991
1992
1977
1985
1985
1994
1986
1977
1992
1993
1991
1990
1982
1983

1158
1305
1395
1542
1666
1742
1759
1799
1851
1991
2131
2174
2299
2429
2591
2603
2729
2'828
2954
3008
315
3195
3603
3619
364
3675
375
3788
3-831
397
4052
4899
4 97
5391
5611
5643
6403
6582
6617
6663
7122
7133
7288
7539
8-495
9718
10936
11051
11972
12224
12344
12473
1284
13387
14024
14815
1518
15869
16323
17-893
18184
18726
19595
2078
220439
24258
24764
25 273
69592
92-822
138 111

193%
36%
13%
40%
142%
134%
220%
11%
38%
199%
107%
198%
109%
32-4%
104%
174%
52%
51%
51%
188%
53%
355%
61%
168%
61%
07%
161%
316%
124%
30%
368%
14%
].7%
21 6%
01%
81%
1-4%
299%
231%
29%
2-3%
510%
39%
838%
48%
56%
52%
142%
155%
56%
262%
499%
14%
27%
289%
69%
150%
137%
100%
942%
43-3%
593%
22%
44%
97' 6%
150%
349%
108%
270%
6 1%
83%

002
022
007
011
008
004
008
022
003
004
008
005
007
004
024
007
017
024
010
0]3
023
010
006
006
013
008
005
011
019
021
010
0 16
023
049
024
013
020
013
010
018
142
026
032
024
053
042
018
016
023
094
016
014
044
046
100
051
058
028
060
078
026
019
103
080
102
066
041
044
094
580
5 12

035%
060%
007%
028%
068%
027%
100%
0 13%
006%
044%
041%
044%
032%
056%
094%
048%
033%
042%
018%
078%
038%
108%
011%
029%
022%
002%
023%
096%
062%
016%
090%
005%
008%
196%
000%
018%
004%
061%
035%
008%
046%
189%
017%
270%
030%
024%
009%
020%
030%
043%
035%
057%
005%
009%
207%
024%
058%
024%
037%
4 09%
063%
059%
0 12%
017%
4 43%
040%
057%
0 19%
036%
038%
031%

859

37%

3240

0 14%

246
5636
315
2098
59
99

190%
692%
685%
563%
684%
397%

088
1445
098
466
023
038

068%
177%
214%
125%
263%
153%

Survey dates

23323
1510
6590
180
2500
133
164

130
814
46
373
9
25

\1971

I ::~

1954
1973
)973

1999
1999
1999
1999
1999
1999

Vol. lost:
%

Annual sedimentation
m"/km 2/yr

583
2193
5467
1865
1697
2311

M.mJ/ann

Mtlann

Notes
% vol.

J
- -J

'-1
-1
1

1
j

1
J
1

J
J
4

J
1

APPENDlX I

II

Table Al.2.

continued

C(luncry

Turkq

Turkey
Turkey
Turkey
Turkey
Turkey

Sdc:vir
Cubuk- I
Bayi ndir
Hilfanli
Kcssikkopru
Alcinapa
Sevhan
Kmalkaya
Cip
Surgu

Turkey

TOlal

Pakistan

Tarbela

Germany
Germeny

Saxonian Reservoirs
Bal c!eney
Ba\'aria - Forggensee
Bavaria - Saalachsee (flushed)
B :lv:ui~ - Sylven5rei nsee

Turk~y

Turkey
Turkey
Turk~y

G~r:nany

Germolny
Germany
Germany
Malaysia
Malaysia
Malaysia

C,,",mom, 1c ".o; ,>


km!
M_ m~ _

R<!sc:rvair

~~

721
660
70
261 70
360
589
19 :!54
11 30
236
2i5

7'

14;;

940
I US

R i ng! ~l

Tunisia
Tunisia
Tunisia
Tunisia
Tunisia

Kasseb
EI Kebir
Mellegut:
Nebhana
Bezlk
Chiba
L:tkmess
.sou Her1.'na
Iuumine
Lebna
Sisi Saad
Sidi Salem
Silianu
Marg'Jellil

Tunisia

Totai

42122

~rasri

UK

95 reservoi rs surveyed

l[~ iy

Csdore Valley
S<lbcr..::

r~l y

To~i

USA
USA
US A
USA
US A
USA

USA

Total

Morroco
Morroco
Morreco
Marreco
MOCTecO
Morroco
Morroco

NakhLa
Mohamed V
Lalla Takerkousc

~~O!"TCCO

EI Kansera

Moue KiTalcabi
Ibn Batouca
My. YaU.>sef
Mar.sour Ed D!lhoi
Bin EI Ou idane
Hassan Addakhi1
Y B Tachfine

Morroco
Morroco
Morr.x o
SM3 Abdeilah
Morroco
Mor.oc o o EI Makhazine
Mor.oco I Hnssanler
;':[c;m:;co 1 AI M:mira

508
8
150
4
32

I""

1965
:960
1966

1
I

1913
i959

236

33-9
12-6
35
3936-0
28-1
14-2
4297
42-1

55-8%
2150%
50-0%
296%
43-9%
358%
284%
371 %
394%

5377

597%

2900

203%

183

36'70

2408

1-6%
100-L%
i3%

1379
303
1464
3760
2364
752
519
1379
3i O
3388

65-8%

8
8

7
11 8
130
29
209
~55

iO
110

1776

I
I

12-38
2374
232-;6
54 12
6 23
743
1002
606
11 00
i 50
600
9 180
561 0
38-50
4400

60764

15 1%
91 3'70
701 %
626%

965%
945%
125-2%
83-9%
94%

5-8%
207%
439%
101'70
550%
40')%

1203 875

109 980

107
49920
1710
45oiO

13

084%

0-06
022
009

;49

G' ll

I0

1 ' 15%

2-5t %

Q4[0/0
I

020%
I

165
275
550
1360

0030
0050
0 101
0249

4228
1200
514
1918
1952
3516
2465
3811
1282
1196
5051
603
18L
3365
4911

0-427
03252
529
164
0 164
0225
0313
0202
0-5
05
05
5-4
J.3
35
55

042%
01 2%
()'05%
019%
020%
0-35%
025%
038%
0-13%
0 L2%
05 1%
006%
002%
034%
0-49%

27-79

Oo();%

0 10%
360
160

723
96
330
43
436
198
592
1484
369
320
51]9
807

I
I! 8

I
196 1
1967
1935
1927
198 1
1979
1970
1972

1974
197:2
;07 1
i979
271
1987
2724 I 1976

O - I~

143
235
209
23-6
l804
19 1

1066
2563

3-4%

002
0 15
097
449
2069
57-92
134-44-

186

424 1

3-')%

218-69

603
2569 1

468'70
354%

265
64-66
6-96
56
22
62-88
9932
2096
16-49

li 6O/o

023
- L1 17
050
122
0S7
058
1-10

22 97

4 59'9

4SB

i 1987
1990
1988
1980
1989
1989
1'790
1988

,990
1989
: 1985
1 1990
11 990
I :937

0
2
23
94

257%
193%
165%
1304%
91%
36%

196%
162%
13%
111 %
106<;;,
6-7%
5-7o/c

51%

33-99

4 '2L~

to

82-94

3-7%
3-0%

014
00 1 I

I
11
138
700
5331
29249
74499

6400

1208)

770
154
93-6

342%

178

1 1.50 %

I
1969
1925
1954
1965
1960
1'765
1966
1968
1976
1983
1986
1981
1981
1987
1990

82
26
332
86
6

IMO
!5 C:("J

1-31%

135

043

I 1996

no

1-64%
3-57')I>
147%
165%
0-90%
137%
083%
105 %
109 %

100

010

30%

Notes

Muann % vol _

0-10
9840
0-85
044
9-99
156
01 1
093

I
I

sedi m~n cJc ian

I M_m-'/~nl1

m-'/km%/yr

280

1 1

Annual

1 1990
1980
19~

425
1132
94 12
39 345
2265 i6
454897
477 14B

4570
3780
9SCO
1821)
1670
28500

Val. la st:

3-644

TOl;ll
sedilO~maci o n:
M.m!

J.6

380
93

190 reservo irs 0 co 12 -33 m'


257 reservoirs 1233 to L233 mJ
283 reservoirs 1233 to L233 mJ
176 reservoirs L233 co 12 330 m~
107 reser/oir5 12 330 to 123 300 mJ
69 reservoi rs 123 300 co I 233 (JOO m'
23 reservoirs> I 233 000 m'

USA

1998

Tunisi a
Tunisia
Tlmisia
Tunlsia
Tunisia
Tu nisia
Tu nisia
Tanisin
Tunisia

Italy

101
271
10 300
855
84
64
127
53
390
418
99
8950
18 250
1040
1110

T~~ ~si:\

1999
1999
1999
_1999
1999
1999
1999
1999
1999
1999

183
183
183
183

Ringlet
Ringlet

To~1

1965
1936
1965
1959
1966
1967
1956
1972
1965
1969

1974

14300

Torn I

M~ lolysia

61
6
7
5980
95
3:!
120U
148
10
71

End

9006

I Ring let

Malaysia

Survey dates
Smn

3093

3-22

I
I

131

0-97
1-77
3-09
333
i -54

356%
2-00%
102%
081%
043%
023%
01 6%

0-20%
1770/0

1-54%
052%
0-37%
2-02%
0 13%
056%
0-66% I
012% j
03 6%
030%
Q35 ',t..
0-38%
122%
1 0.28 %

159

EVACUATION OF SEDIMENTS

Table Al.2.

continued

Country

Reservoir

Mormeo
Morroeo

ldriss ler
I Abde1moumen

MOITOCO

Total

Algeria
Algeria
Algeria
Algeria
Algeria
Algeria
Algeria
Algeria
Algeria
Algeria
Algeria
Algeria
Algeria
Algeria
Algeria
Algeria
Algeria
Algeria
Algeria
Algeria

Ghrib
DjorfTorba
Bough7..0ul
Bouhanifa
5MBA
Chcurfas
K'Sob
Bakkhadda
Foum EI Gharza
Beni Bahdels
Oued Fodda
Ighil Emda
La Chaffia
Zardesas
Sarno
Foum E1 Gheiss
Hamiz 1& 2
Mef\'rouch
Reservoirs on the Atlas Mountains
Reservoirs all flatter land

Algeria

Tolal

TaiwlIn
Taiwan
Taiwan
Taiwan
Taiwan
Taiwan
Taiwan
Taiwan
Taiwan
Taiwan

Shihmen
TenKee
WuSheh
Tsengwen
Pai ho
Gen Shai Pei
A Kung Ticn
Lu Liao
Tapu
Ku Kuan

Catchment:
km 2

Stlllt

End

1972
1981

1986
1987

169
1120
55
73
235

1974

31

1939
1974
1952
1938
1974
1954
1965

1986
1986
1986
1986
1986
1986
1986
1986
1986
1986
1986
1986
1986
1986
1986
1986
1986
1986

3680
1300

1217
216

135195

10351

9099
34118
28571

1414
246
1132
443
1277
873
288
1453

309
255
150
713
25
7
36

3094
138

2944

1525

Indonesia
Ind{lIlesia
Indonesia

Karangkates
Selorejo
Wonogiti

2050
238
1262

343
62
730

Indonesia

Total

3550

1135

Sudan
Sudan
Sudan

Sennar - Blue Nile (dis of Roseirc~)


EI Girbll- Atbara river (Nile lrib)
Roseires - Blue Nile

Sudan

Total

77108

74%

1126
336
22-3
214
94
12
1316
81
205
65
923
349
73
165
07
14
665
J4

667%
30%
409%
293%
40%

425%
40%
477%
10-3%
636%
349%
43%
148%
32%
436%
348%
20%

Egypt

Aswan High Dam

Ethiopia

Koka - Awash & Mojo riven;

CYPl1ls
Cyprus
Cyprus
Cyprus
Cyprus
Cyprus
Cyprus

Galini
Petra
Kalokhorio
Lynlhroclhonda
Lymbia
Kophinou
Akrounda

Cyprus
Venezuela

221
023

)]1
85
21
59
213
41
198
122
433
i36
2098
2498
573
938
234
187
640
648
1100
2800

1982
1982
1985

930
1300
3354

167000
I

I 0-42%

JOI
290
060
056
118

060%
026%
110%
077%
050%

028
003
049
014
093
319
050
027
034
003
0-48
0-03

090%
001%
114%
021%
064%
319%
029%
024%
J55%
093%
249%
004%

1014

008%
060%
016%
031%

17%

4-41

I 039%

560
740
1127

602%
569%
336%

918
2846
5932

099%
219%
177%

2427

1 435%

9696

I 174% i

2100

13%

8077

0-05%

1986
1990
1985

1964

1990

1961

1981

]700

340

10

0216

0026

JI9%

000%

Sanlo Domingo

427
307
1202
148
2350
1400
1082
4S85

"

1976

1978

0575

15
50
8
143
109
32
229
II

1905
1913
1918
1932
1947
1937
1941
1960

1930
1959
1933
1961
J961
1965
1%5
1%7

0188
092
0138
85
03
38
4183
103

13%
18%
17%
59%
03%
118%
183%

56706

95%

192%

94<;i-

24

17
63
125
15
125
364
118

029

958%

001
002
OOJ
029
002
01.:f.
)78
0\5

005%
004%
012%
021%
002%
042%
078%
: ]35'k

242

i
I

i 040%

166

1
1
1

19

TOlal

597

204
010
227

410%
390%
303%
114%
7,13%
1433%
555%

30%
19%
12%

82%
78%
91%
57%
285%
430%
167%

mo

067%

167

000192
00024J
000628
000232
000739
0-00159
000370

067%
044%
072%
059%
151%
144%
143%
001%
254%
2-48%

0023
0031
0069
0041
0026
0004
0022

0-49%

26
37
24
9
33
12
26

l..esna
Pikhowkc
Lubachow
OtmuchQw
Turawa
Porabka
Roznow
I Myczkowce

0. 18 %
011%

1022
IJ6
907

1925
1964
1966

5584

% vol.

206
112
j07
423
038
010
052
000
023
0-42

I
1977
1970
1981

IMtlann

4327

13
2700
1900
4900
8800
14300
9400
16300
30
2200
600

1938

NOles

155%

409
1963

1998

M.mo/ann

m>lkm /yr

25%
06%

Total

Annual sedimentation
2

9
17

Total

160

1953
1939
1972
1937

Vol. lost:
%

M.m~

2633
763
592
219
481
27
11
32
8
104
708

Taiwan

1948
1978

203
43
63
145
100
171
111
22
3
19
70

Taiwan

Poland
Poland
Poland
Poland
Poland
Poland
Poland
Poland

Tota!
sedimentation:

Capacity:
M.ml

1
1

II

r"
.1

APPENDIX I

Table Al.2.

continued
Catchment:
kml

I
[

ICaDadty: I Survey
I

~l.m'

! I
Start

Avisio
P"rnegg
Pt.mtebbn
Steyerdurchbrudl
Tar:ento

Austria
Austria
Austria
Austria
Austnn
Austrit,
Austria
Austria

Wet;:man
Margarilze - study rt!f. further

ITotal

956
6150
10
575
62
324

085
0[5
0-60

8177

410

Switzerland
Switzerland

Kallna..:h
Petolle.

Switzerland

Total

2621

New Zealand
New Z:!\]and

Roxburgh
MutOlhina

8826

28+1

5740
34

i7:

::[~C~land ~lomi
Total

S'lutn Africa
South Africa
Suuth Africa
SO(1tb Africa
South Africa
South Africa
South Africa
South Africa

Afhasini

End

M.il,.'

1925

, 189()
1927

1892
ISS]

1908
1884

II: ~~~ !~~~


I
1913

1919

: 1872

1886

1961

1979

[966
193.;1.

1980
1959
1979

Vol.losl;
%

1000%
700%

035
025
071
015
06

840%
1000%
1000%

381

930%
556%
100,0%

Annual sedimentation

m)/km~/yr
261
3704
1545
003
193
1852

12$

56

M.m}l:mn

Notes

I Mtlanll

% vol.

025
023
002

12:50%
4670%
380%

00\
060

800%
10000%

114

2HI%

017
007

945%

713%

714%

IOH
497

475
01

i 1939

~~:

26
liS

A1!cmanskraal
Bt:ervld
Boegoeberg
Bon Accord
Bronkhorstpruit
Bulshoel(

9t

'9

Darlil1l!:ton
Drid -

years

2628
2-982
95
00296

ill

2~~~~~:

1923

1995
1995
1986
1995
1995
1986
1995

1917

1995

1935
1922

1995

98

1995

1399

1995
1995
1995
1993

1975
[929

1925

S9
6

1950

to

E!!monc
Elnndsdrift.
Erfenis
Aornkraal
Gamlmpoort

1952
1960

36

124
169
to

Somh Afri:a

7
208
50

42

i973
1937

1977
1960
1957
1969

H:utbefspoort

Hazdmere
K:Ilkfolltein
Kamn1<lnassie
Keerom

91

19~4

195
18
355
36

]9!5
1977

[(l:l%rric

Kommandodri ft
KGPpi~~

Kiommenellenboog

74
41
9
73

10.5
14

Leeu Gllmka

]g4

Marico
NooU!!edacht
Phalaborwa ba.crll"e
Pietersfont.t!in

'17
79
9

PUnl!alaDCOn

S~'u(h.Ar.';ca

Poo;;jie'

E'rinsrivier
Ril!tvld
Rust de Wimer
S(ompcirlf.

11
27

55
33
2536

V:lU Rr:-:e'.'e!c?u$
Welbedacht
Wentzel
Windsor

SQuth Africa

To!:!!

1923
[954

1960
1956
1911
1955
1970
1924
1959
1939
1933
196::!
1966
1963
1917
1933

1934
1965
1990
1938
1936

Vaaiha.ctZ

S"uth Airica
SO'Jtn Africa
SQuth Arrie:!

1938

196&
[973

"

1995

1995
1995
.1995

5673
22

South Afric:l
SoulhAfric:a
South Africa
South Africa
SDuth Africn
South Africa
South Africa
South Africa

Toml

~:~:

200%
210%
192%

165
75
66
22

[4600

139%

g:~~~~ I

043%

080%
053%

00007
20537

IO!% I

II'

13%
263
025
0l4%
0..()9%
396
314
08%
1
7920
36j0
9 rellrs I
2822
-----------4------+---~--~~--+_------+_--~----~----~--_4----~--009%
382
08%
3-39
30471125
8870 ! 3329

Bmzil

SuutnAfrica
S~'I.!th Africa
SomhAfricn
South Africa
Suuth Africa
South Africa
South Africa
South Arne:!
SoudlAfrica
SilUlh Afril.:a
South Afric:t
South Africa
South Africa
South Africa
South Africa
S')llth At'ricn
South Afric:!
SuuchAfric:!
Soui.hArrica
SourhAfrica
South Aftica
Seurh Afric:!
SomhAfrica
South Africa
South Africa
South Arrka
SOllth Africa
Scutt: Africa

! 18~2

1360
J261

~~: ~~::~~~ I ~~:~~~

B.'aZi!

2()O
050

dUlt:$

112
'6

19:!.5
1973
1934

1950

1995
1995
199.:5
1995
1995
1995
19~6

1995

J995
1995
1995
1995
1995

1986
1995

41
4C.~

6643
14J

\3
68913
I
I

35
25
::!31
16g
101
5.:l..59

29
412

H
09
0435
1.:1.7
12
23
114

111%

066
044
039
2::!75

24}%

9-6%
132%
453<;7.

113%

2-19%
062%
198%
0,32%

038%
093%
040%
049%
064%

015%

199%

06"
005
002
002
038

018%
014'1&
030%
029%
051%

295%

014

035%

2459

006
050

061%

122%
ij,}'e

169'7~

551%
59%
119%
07%

64

011
058

037%
066%
066%
061%
042%
033%
024%
027%
013%

029
034

73
22-656

Q033
210

016
192

3;14%

71-2%

5)
064.;1.3
1-1968

022
002
019
001
001
023
006
014

7j.S

3-2

060

79%
828%
.;!.g.!%
360%
357%

103%

3
06
55

1996

174'*
207%

34.:2%
103%
[00%

05516

1995
1995
1995
1995
1995
1995

117%

20

1995
1995
1995
1995
1986
1986
1995

399%
297%

61
366

1985
1995

160%
230%
73%

3J,,1~;

22,)%
22"%
90[%
663';7~

52t;~

449'c
IHj%
01%

107
0-22
0-48

005
002
010
002

250
006
007
001

002
021

001

190%

068%
102%
153%
013%
019%
003%
[20%

0&3%
010%
282%
08:5%
010%
008%
039%
002%
015%
065%

83~:tt

36~

41-(;

3g4%

071

3!4

397%

O57'V~

967
13

045
440

:mJ%

{JO::!,

O3Y'lc

~31';i

OOS

135~1:

107

066';:;'

1577

)93'ij,

034%

161

EVACUATION OF SEDIMENTS

Table Al.2.
Country

continued

Reservoir

Catchment:
km 2

ICapacity:
M.m)

Survey dales

Kenya
Kenya
Kenya
Kenya

Kindaruma on Tana
Kamburu on Tana
Gitaru on Tana
Masinga on Tana

9810
9520
9540
7335

Start
1967
1974
1978
198]

7
150
20
11560

End

Total
sedimentation:
M.m3

IVol.%lost:

117

1980

Annual sedimentation

78%

m;/km'/yr

M.m 3/ann

246

2-34

Kenya
Portugal
Portugal
Portugal
Portugal
Portugal
Portugal
Portugal
Portugal
Korea

Santa Luzia
Montargil
Campilhas
Burgaes
Idanha
Arade
ROKo

Lake Kariba

50
164
22
0
79
29
96

I Total

Lake Cabora Bassa

1960
1961
1966
1962
1963
1973
1980

156%

440
85

26 years

002
002
003
000
015
011
007

003%
002%
015%
026%
019%
039%
008%

05%
00%
16%
58%
32%
71%
10%

&239

1-4%

041

009%

077012

111%

003

043%

116000 M.m' - no re-survey data available. but estimates = 1600 to 16000 yrs to !ill dead storage

650000
1000000
(inc. Kariba)

310
21
289
33
404
497
215

025
005
0347
0019
254
205
0983

Sedimentation may
I

in

NOles:
1. Reservoir is dis of Lake Wular
2. Catchment area not permanantly covered by snow only
3. 12 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 ml/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

% vol.

11737

Beaggog

Notes

Mtlann

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 28 billion. In addition, with an
installed capacity of 3478 MW and a finn electrical energy of 148 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

The accumulation of sediment within the reservoir causes two major


problems.
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
]

APPENDIX 2

-l
:]

1
t
l

14 000

12 000

Rainfall

10 000
IJl

---I

8000

iti

3:
0

6000
4000

2000 I-

J
J

Month

Figure A2.1.

Case study - inflow hydrograph to Tarbela reservoir, Pakistan

A2.2.2. Indus River flows


The average Tarbela inflow hydrograph, see Figure A2.1, shovvs t.'1e contribution
to the run-off made by rainfall and snowmelt. It is estimated that the monsoon
(rainfall) contribution to the total run-off is approximately 10% of the whole. The
mean annual inflow into the Tarbela reservoir is 81 km3 Variability of mean
annual flows from year to year is small, wiLl. a coefficient of variation of only
15%.

A2.2.3. Sediment inflows


The mean annual sediment inflow is 240 NIt. Of this, approximately 40 Mt of
very fine sediment passes through the reservoir and 200 Mt -. of the coarser
fractions deposit within the reservoir. The vast majority of the annual sediment
load enters the reservoir during the high flow season, l\1ay to September.

A2.2.4. Sediment deposition


Since the reservoir was first impounded in 1974, a high proportion of the annual
sediment inflows into D.1.e reservoir have been deposited to fann a delta which has
been advancing towards the dam. This has been monitored by detailed surveys
which have been carried out annually since 1979, covering the whole of the
reservoir.
The delta profile is sensitive to the way the reservoir is operated, in particular
to the minimum pool level and the length of time the minimum pool level is

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.S. Numerical sediment modelling


A numerical model was used by HR Wallingford to simulate reservoir
sedimentation in the Tarbela reservoir. The model took the original reservoir
cross-sections and a 60-year sequence of water and sediment inflows into the
reservoir. It computed the sediment profile at each cross-section using equations
that relate sediment movement and flow for a range of sediment sizes. The output
included:

changes in bed topography as sedimentation deposits and erodes


volumes of sediment being deposited and eroded
changes to live and total storage curves as sedimentation progressed
daily discharges passing to the downstream reach
sediment loads passing to the downstream reach.

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.

A2.2.7. Reservoir operation policies


Several scenarios were developed for the operation of a flushing system at
Tarbela and appropriate runs of the model were carried out to explore the
influence of different assumptions about the way in which the reservoir and the
flushing system should be operated.

A2.2.B. Reservoir flushing


Five model runs were carried out to simulate reservoir flushing, in which the
flushing level, the flushing period and the date flushing commences were varied.
The results show that for the conditions pertaining at this particular site:
flushing provides a substantial long-term live storage with only a small annual
reduction

166

APPENDIX 2

455
435

c:

.2

415

ca>
Q)

395

"0

--1974observed
_. 1996 observed

Q)

co
375

- - - - 1984 mode!
1996 model

355
335

10 000

20 000

30 000

40 000

50 000

60 000

70 000

80 000

90 000

100 000

Distance from dam: m

Figure A2.2.
Pakistan

It

It

Case study -

veri,;ication of numerical model, Tarbela reservoir,

low-level flushing is more effective than high-level flushing


flushing over a 3D-day period is more effective than over a 20-day period.

A2.2.9, Typical numerical modelling results


Figure A2.3 shows the throughput of sediments to the year 2056 without the
introduction of a flushing system. It assumes an operating system which
gradually raises minimum water level year on year until the live storage of the
500
c::::J Silt 1 c:::J Silt 2 ~ Slit 3 ~ Silt.:;. :::::::: Siit 5 ~ Sand 1

Sand 2 _

DiSChaigel120 000

450
100000

i'! ::~

,..,
E

80 009

c::

'E
(Ii

60 000

250[

(3
CIl

200r

150

'6

40000

-<

ca::l
C

c:

-<

100
50
OL-~

__-L__~__~__L-~__~__~__~~~~__- L_ _~____~~_ _~I 0

1975

1985

1995

2005

2015

2025

2035

2045

2055

Year

Figure A2.3.
Pakistan

Case study -

sediment throughput without flushing, Tarbela reservai,;

167

EVACUATION OF SEDIMENTS

500

120000
c:::::J Silt 1 c:::J Silt 2 E::El Silt 3

Sand 1 _

Silt 4 c=J Silt 5

Sand 2 -

Discharge

450

400

'E

350

-0

Qi
':;"

100000

80 000

'"E
~

'E

300

a.;

C 250
OJ
E

60000 . ~

200

'5

'5
OJ

C/)

C/)

(i'l
::!

c
c

40000

150

(i'l

::!
C
C

100

20000

50
o~~--~--~--~~--~--~~~~--~--~~--~--~--~--Uo

1975

1985

1995

2005

2015

2025

2035

2045

2055

Year

Figure A2.4.
Pakistan

Case study -

sediment throughputs with flushing, Tarbela reservoir,

reservoir is reduced to approximately 20% of the value at first impoundment. The


sediments passing through the reservoir are mainly the finer fractions and these
either pass through the machinery during the generation of power or over the
spillways towards the end of the flood season.
Figure A2.4 shows the equivalent results with the introduction of a low-level
flushing system and an appropriate operating rule for reservoir levels. In this
case, the long-telID throughput of sediment matches the incolning sediment
quantities and the reservoir live storage settles to a value of approximately 50%
of the original live storage.
14000

____ D1 -1350 ft min

-L:r12000

D5 -

~ D4 -1500 ft min
- 0 - F4 - High flush from 2021

.......

1350-1500 ft min

--0- F6 - Flush from 2006

-+-

10000

85 -

11
10

F7 - Short flush from 2006

8asha in 2016

.......................... . . :. . . . :. . : i

4:ir.~~..-;-;-:-F:.~-~-.~.-.o-.

u.

0- - 0 -

,I

6000

.....~

~ -

4000

.() -

-0

0-

0 -

-0- -0- -

-----~-----x------x-~~~..~--~--~---..~
~
-&'--A-=--

2000

::J

-1

3
2
1

O~

__

1996

__

2001

____

2006

2011

__

____

2016

2021

__

__

2026

____

2031

2036

__

__

2041

____

2046

2051

__

~O

2056

Year

Figure A2.S.
Pakistan

Case study -

II

..-1

prediction oj Juture live storage, Tarbela reservoir,

I
j

168

~
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


A3.1. MANGAHAO RESERVOIR (New Zealand, 1924)
The Mangahao reservoir and 20 MW hydropower scheme "vas constructed in
1924 on the Mangahao River, in the south of New Zealand's North Island
(Jowett, 1984). It was recognised at the outset that sedimentation of the reservoir
would be problem, and a second reservoir was formed upstream to act as a
sediment trap. As it was originally intended to continue building more reservoirs
upstream for sediment trapping, no specific sediment management facilities were
installed in either of the reservoirs. No information is available regarding the
catchment area, basin size, annual run-off and sediment inflow.

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 highlevel 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 088 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.I.3. Downstream impacts


Large banks of sand and silt formed immediately downstream of the Mangahao
reservoir during flushing. Studies have shown that the effect of flushing on
substrate and invertebrates was minor, but noted the temporary increase in
turbidity that was disturbing to recreational users.

A3.2. GUERNSEY RESERVOIR (USA, 1927)


Guernsey reservoir on North Platte River, Wyoming, is used prinlarily for .
irrigation, but also provides hydropower. It is impounded by an earthfill dam,
which was completed in 1927 . The dam height is 41 nl, the reservoir length is
235 km and the original storage capacity was 91 M.m3 (Jarecki and Murphy,
1963), representing only 4% of the estiInated mean annual inflow.

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

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 08 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 02% 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.
.

AI2.3. Downstream impacts


No problenls are mentioned in the references, probably because the outflow
sediment concentrations and efficacy of flushing are low.

foJ.3. ZEMO-AFCHAR RESERVOiR (Former USSR, 1927)


The Zemo-Afchar hydropower reservoir, completed in 1927, is located just
downstream of the confluence of two rivers. No data are available with respect
to the original storage capacity of the reservoir, although the basin length is given
as 8 km along one tlibutary and 18 km along the other (UNESCO, 1985).

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 23 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

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
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 85
to 655 hours, with a mean of 185 hours, and was canied out mainly in the
month of April, Mayor November.
The volumes scoured each year ranged between about 05 and 2 M.m3, with an
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.
General conclusions drawn (UNESCO, 1985) from the flushing operation at
the ZelTIo-Afchar reservoir were:

l
JI
j

!
~
i

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
No information available.

A3.4. JENSANPEI RESERVOIR (Taiwan, 1938)


The Jensanpei reservoir was built in 1938 for the purpose of water supply to the
sugar cane industry. The reservoir had an original storage capacity of 7 M.m3,
which is probably of the order of a third or less of the average annual inflow to
the reservoir. This was raised to 77 M.m3 in about 1942 and to 81 M.m3 in about
1958, presumably by raising the impounding level.

A3A.I. Sedimentation
The erosion rate in south-west Taiwan is particularly high, due to climatic
conditions and geological conditions of soft and erosive rock. In an 18-year
period from 1938 to 1955, the storage depletion at the Jensanpei reservoir due to
silting was 426 M.m\ an average annual loss of 34% of the storage capacity
.
(Hwang, 1985).

174

- I

J
J

J
J
J

J
J

APPENDIX 3

A3.4.2. Flushing
L., 1955 a 15 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.

A3.4.3. Downstream impacts


No infonnation available.

A3.5. NAODEHAI RESERVOIR (China, 1942)

The N aodehai reservoir is a flood detention reservoir situated on the Liuhe River
in an arid region in the north-eastern part of China. As the main purpose of the
reservoir was for flood control, it was built initially with ungated outlets near the
river's original river-bed. Control gates were subsequently installed (apparently
1970) to preserve clearer water for irrigation in the non-flood season. The
reservoir was designed to attenuate a peak: inflow flood of 3500 m3/s to 1640 m3/s
(UNESCO, 1985). The original design flood storage volume was a little less than
the mean annual inflow.

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.

Quantity of deposits during flood peaks in 1963, Naodehai reservoir

Date

July 20-22

July 23-27

July 28-31

Max. water level (m)

8457

8862

8308

Max. inflow discharge (m3/s)

]928

7980

]160

Max. outflow discharge (m3/s)

760

2470

440

Max. silt discharge of inflow (tis)

617

3280

388

Max. silt discharge of outflow (tis)

67

168

]47

23 0

673

129

99

113

104

]31

563

25

Silt quantity in inflow (Mt)

J
I

Silt quantity

i~

outflow (Mt)

Quantity of sediment deposited (Mt)

-,

Between 1950 and 1958, for example, there was progressive erosion along the
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.

A3.5.3. Downstream impacts


No information available.

A3.6. GMOND RESERVOIR (Austria, 1945)


The Gnliind reservoir, used for hydropower, was formed by the construction of
the first arch dam in Austria (1943-45). The height of the dam, whose crest also
forms the spillway, is 37 nl, the maximunl impounded depth is 30 m (Rienossl
and Schnelle, 1982), the length of the reservoir is 940 m and the maximunl width
is 200 m, giving the reservoir a surface area of 124 ha and an original storage
capacity of 093 M.m3, which is only 05% of the average annual run-off.
The reservoir includes a bedload trap at the upstream end of the basin, leading
to a tunnel that bypasses the reservoir. This has required substantial periodic
repairs as a result of abrasion, firstly with a steel-plate lining and subsequently
with a lining of basalt slabs. Bedload reduced substantially afterthe completion
of Durlassboden reservoir in the upper catchment in 1967.

176

J.

1
~

APPENDIX 3

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 007 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 02 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 015 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:
the reservoir is drawn dovvn by 9 m one week before flushing is to commence
the newer bottom outlet is op~ned the evening before flushing day and water
is discharged until the drawdown is 14 m
\I the original bottom outlet is opened and the reservoir is emptied
~ 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
Durlassboden reservoir (25 m?/s) over a period of 3 to 5 hours.
s

It has been found that between 15 000 m 3 and 30 000 m 3 of sediment is


scoured in this "'lay, with a sedin1ent/water ratio as high as 5%.

177

EVACUATION OF SEDIMENTS
-r

J
The construction of Durlassboden reservoir upstream of Gmund reservoir has
-_been a key factor, reducing the incoming sediment load and aiding the flushing
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.

A3.6.3. Downstream impacts


No information available.

-- f

J
-,

J.

i
i

~,

A3.7. PALAGNEDRA RESERVOIR (Switzerland, 1952)


Palagnedra reserv()ir was completed in 1952 and forms part of a hydropower
scheme on the Melezza River in the southern slopes of the Alps. The water is
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
length of the reservoir is 26 km, giving the reservoir an original storage capacity
of 55 M.m3 , which is less than 3% of the average annual run-off.
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.

I
I

I
I

J.

A3. 7.1. Sedimentation


The average annual sediment inflow to Palagnedra reservoir is 008 Mt
(Atkinson, 1996), but this masks a wide annual variation. The sediment type is
largely silts with some coarser materials.
It is believed that there were no particular problems with sedimentation in the
first seven years of operation, although there are no specific records to confirm
this (SNCOLD, 1982). In 1961 a flood with a peak discharge of 500 m 3/s carried
about 05 M.IU3 of alluvium into the lake (9% of the storage capacity), after
which accurate routine observations were n1ade. These showed that sedimentation continued steadily, at a rate which was thought to have increased as a result
of the effects of the flood. By the end of 1968, the volume of the deposits had
reached 147 M.m3, or 270/0 of the original capacity.
Records of annual sedimentation are available for the period 1969 to 1976
(SNCOLD, 1982) and show annual deposition of between 10 000 and
240000 m 3
In August 1978 an unusually large flood originated in the catchment area of the
3
Melezza River. This flood, with a peak discharge of 1000 m3/s, caused 18 M.m
of deposition in the reservoir, which was equivalent to 33% of the original
storage voluine. The intakes of the two bottom outlets were covered by Inaterial
deposits 12 m and 26 In thick respectively and were temporarily out of service.

178

J,

1.
1.

APPENDIX 3

The materials deposited during the flood ranged from silt, sand and gravel to
14% wood.
In the period 1953-78 a total of 379 M.m3 had been deposited, representing
69% of the original volume and an average annual accretion rate of 015 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 03 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 15 m3/s by using part of the river run-off.
During this three-month phase it was estimated that approximately 21 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.

A3.7.3. Downstream impacts


No information available.

A3.8. GUANTING RESERVOIR (China, 1953)

Guanting reservoir, on the Yongding River upstream of Beijing in northern


China, was built to provide flood protection, river regulation and hydropower.
The river regulation supports a major downstream water supply abstraction for
the city of Beijing, together with downstream hydropower schemes.
It is reported (UNESCO, 1985) that the reservoir provided flood detention
from 1953, "vith the sluice gates normally partly open and water levels in the
basin fluctuating widely. From 1955 it has operated as an impounding reservoir,
with the
level varied to
releases to the downstream river and to
provide storage for impending floods.
The
is impounded by a 45 m high dam.
reservoir basin
comprises two aIms, the Yongding arm being about 30 km long and the Guishui

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 13 Mt in the period
1951-84 (Binnie and Partners, 1986). As shown in Table A3.2, there was a
Table A3.2.

Guanting reservoir, sedbnentation


Sediment deposition: M.m3

Mean annual
sediment
inflow: Mt

Mean trap
efficiency:
%

In period

Cumulative

73

73

352

352

1961-70

19

96

151

503

1971-80

11

96

85

588

1981-84

100

23

611

1953-84

28

611

611

Period

1953-60

I
!

8]

180

APPENDIX 3

progressive reduction in the estimated annual sediment inflow to the reservoir


betvveen the 1950s and the early 1980s.
During the first years of operation, up to 1957, it is reported (UNESCO, 1985)
that 268 Mt of sediment was deposited in the reservoir, cOlTesponding to a trap
efficiency of 63%. Sediment discharge during that period comprised erosion of
the emptied bed, which occurred prior to i.mpounding in 1955, and density
current venting thereafter.
The reduction in the amount of sediment entering Guanting reservoir since the
1950s is attributed to the construction of some 300 reservoirs, with a total storage
capacity of 1500 M.m3, within its catchment (UNESCO, 1985) and the warping
of agricultural land by the diversion of highly turbid irrigation flows.
By 1986 it was apprehended that the sediment deposition in the Yongding arm
of the reservoir basin was about to create operational problems, principally by
blocking the downstream end of the Guishui ann and isolating a major"part of the
residual benefit storage.

A3.8.2. Flushing and other remedial options


The original operating rules required that each of the eight bottom outlet gates
should be opened for a short time at least four times per year (Binnie and
Partrrers, 1986), to minirnise the build up of sediment in the immediate area. By
1986 t~e general sediment level had reached 17 m above the invert of the lower
gates within 250 m distance of the bottom outlet, and occasional blockage was
reported.
One incident of blockage occurred in 1962, resulting in minimal outflow for
the first few minutes after a sluice gate was opened, but eventually the area was
flushed clear of sediment. Since then, it is reported that the sediment level is
monitored and the gates are each opened for about 20 1I1jnutes in tum whenever
the level is more than 05 ill above the lower gates, with a minimum interval of
one month. These operations remove only local accumulations and no attempt is
made to flush a significant proportion of the sediment inflovv. Partial blockage of
the gate occurred in 1974, which was sluiced away after the gate was raised
above 1 m.
Atkinson (1996) reports what appears to have been a more substantial flushing
operation in October 1954, when a discharge of 80 m3/s was passed for five days
at a ponded depth of 8 ill and removed about 10% of the annual sediment inflow.
This was prior to the period of impounding and very early in the life of the
reservoir, so it cannot be taken as representative of what might now be achieved
by flushing.
NUIT1erOUS options have been considered for managing sedimentation of
Guanting reservoir (Binnie and Partners, 1986). Flushing would not be
acceptable, because of the ad verse effects on the small impoundments for the
downstream hydropower schemes and the intake for the water supply to Beijing,
together with the rise in flood levels which would occur due to sediment

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.
No information is available on whether these or other proposals were adopted,
or on sedimentation and flushing experience since 1986.

A3.B.3. Downstream impacts


As noted eariier, large-scale flushing could not be undeltaken (even if it were
technically possible), because of impacts on downstream works and flood
levels.

A3.9. SHUICAOZI RESERVOIR (China, 1958)

The Shuicaozi reservoir is located on the Yili River in south-west China and
forms part of a hydropower scheme involving four hydropower stations in
cascades on two adjacent river basins (IWHR, 1983). The reservoir is 6 Ian long
with a 37 m high dam, giving the reservoir an original storage capacity of
958 M.m3, which is less than 2% of the average water inflow to the reservoir of
514 M.m3

A3.9.1. Sedimentation
The mean annual sediment inflow was estimated at 063 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 36 M.m3 ,

A3. 9.2. Flushing


Between 1965 and 1981, six flushing experiments were conducted on drawing
down the water level in the Shuicaozi reservoir to erode sediment deposits, The
outflow Inust be flushed down the spillway, which has a crest elevation about
17 m higher than the original river bed (11 m below the maximum hnpounding
level).

182

"

1,

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 02 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.
Retrogressive erosion of the bottom of the basin commences at the upstream
end of the pool, whose position is controlled by the spillway level. This limits
the erosion capacity of the flow.
In the early stages of flushing, the sediment derives mainly from the deepening
of the main channel.
o Subsequently, the sediment derives mainly from collapsing of the banks.
1"1 the top portion of the delta, on deposition, the main channel is scoured
res lIlting in a progressive increase in accretion on the flood plains.
In the lower part of the delta and in the reach in front of the dam where the
river valley is narrow, unconsolidated silts migrate towards the main channel
and the level of the entire cross-section can be lowered.
$

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.
Flushing should con1IDence after achieving the maximum p~acticable. drawdown.
e In the early stages of flushing, ~he maximum practicable discharge should be
deployed to deepen the main channel.
The discharge should then be reduced, to lower water levels and to induce
more bank instability.
e Finally, the discharge should be raised again to flush out the sediment which
has accumttlated from bank sliding, etc., and enlarge the cross section of the
main channel.
C)

This procedure was adopted in the 1984fiushing operation and proved


successful in increasing the average sediment concentration and reducing the
water consumption.

183

J
EVACUATION OF SEDIf'lIENTS

A3. 9.3. Downstream impacts


No information available

}
A3.IO. HEISONGLIN RESERVOIR (China, 1959)

Heisonglin is a small reservoir, with an initial storage capacity of 86 M.m3 ,


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.
. Following severe sedimentation problelns at numerous reservoirs built in
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
appropriate sediment management techniques (MOITis and Fan, 1997).

A3.! 0.1. Sedimentation


The mean annual sediment load is reported as 071 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
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
typical of most reservoirs with the potential for significant sedimentation
problems.) The sediment is lnainly silt and derives largely from high rates of
gully erosion in the catchment.
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
siltation of 162 M.m3 , representing an average rate of 6% per annum.

J
J

I
!I

,1

J
,

II

AJ! 0.2. Flushing


The reservoir is equipped with a single 2 m x 15 m bottoln outlet with an invert
level 12 m below the original base of the reservoir basin and a discharge capacity
of 10 m3/s, for the purpose of making downstream releases.
Starting in 1962, the mode of operation changed to involve emptying the
reservoir during the flood season and impounding water only during the nonflood season. Turbid density ClUTents were also released. These measures
lnanaged to reduce the trap efficiency of the reservoir to 15%, but the reservoir
was still losing capacity. As a result of additional work, the lateral erosion
technique was developed and first used at this site in 1980 by Xia Mai-Ding and
his co-workers. This technique was found to be capable of arresting sediInent
accumulation and recovering a portion of lost storage.

184

J
J

J
J

APPENDIX 3

The stable long-term capacity which can be sustained at Heisonglin by


deploying all the sediment management techniques is estimated as 25-30 M.m 3 ,
representing about 20% of the mean annual inflow and about 30% of the original
storage capacity. Further descriptions of the flushing techniques follow.

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
68 months between 1980 and 1985. Using a flow of only 02 m 3/s, 816000 m 3
of deposits were eroded into the main channel. This equated to a remarkably
higher sediment/water ratio of 023.

A3.IOJ. Downstream impacts


A unique feature of Heisonglin is that sediment balance is achieved in the
reservoir, while both the water and sediment are diverted to beneficial use.
Between floods, the river flows freely through the reservoir outlet, and is diverted
into llTigation intakes downstream. Seasonal floods exceeding the capacity of the
irrigati;n intakes are temporarily impounded and released to irrigators at the
maximum capacity of the downstream irrigation diversions, which, along with

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
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
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
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:
the reservoir was to impound water continuously, but release 350/0 of the
sediment inflow as turbidity currents, through 12 outlets at an elevation of
300 m (40 m below top operating level)
"
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.
, These figures proved to be extremely optimistic (Monis and Fan, 1997).

1
1

1
I

,..l
t

1
J

, -'

A3.II.!. Sedimentation
The estimated sediment load prior to construction of the dam was 1600 Mt per
annum, with a median diameter of 003 mm, representing a mean concentration
of 38 gIl. Extreme concentrations of up to 940 gil had been measured in the
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
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
elevation and flood levels in the Yellow River as fat as 260 kin upstream of the
daln.

186

APPENDIX 3

A3.//.2. Development of flushing and sediment control


The severe sediment problems at Sanmenxia threatened agricultural lands and
riverside industrial areas and could have required the relocation of up to 1 million
more people, while abandonment of the project would require removal of the
dam and have severe impacts due to additional downstream sediment loads over
the subsequent years. Accordingly, the establishment of a sustainable and
acceptable sediment balance became a high priority objective, on which the
realisation of the other project benefits depended. The water and sediment
management objectives for Sanlnenxia can be summarised as:

controlling extreme floods


providing irrigation supplies, ice jam control and hydropower
limiting upstream backwater deposition and rises in flood levels
limiting the amount of deposition downstream of the dam
preserving the long-tenn effective storage capacity.

To achieve these objectives, Sanmenxia dam required extensive reconstnlction


to provide hjgh capacity bottom outlets, and the reservoir operation was changed
substantially; impounding during the non-flood seasons and emptying and
flushing during the flood season from July to October each year. Almost two
decades were required from first impounding to bring the sediment inflow and
discharge into balance, through the implementation of a series of staged sediment
control measures. The six stages described below outline the operational history
and development of Sanmenxia reservoir.

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 3355 m (45 m below the
planned impounding level). Sediment accumulation raised bed levels by 45 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 175%, 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)
In time for the start of the 1970 flood season, 8 of the original 12 river diversion
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 yearon-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
125 MW sets.
Stage 5 (1973-78)
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
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
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.
Stage 6 (1978 onwards)
Repair of serious abrasion in the bottoln outlet was carried out, decreasing the
cross-sectional area of the outlets and reducing the low-level flushing capacity.
To cOinpensate for this, two additional bottom outlets were opened in 1990.
Since 1980, to avoid turbine abrasion by high sediment concentrations,
hydropower generation has been halted during the flood season.
Long-term storage capacity
Since the sediment balance was brought under control from about 1975, the net
storage capacity below an elevation of 330 In (10 In below top water level) has
fluctuated between about 3000 and 3300 M.m3 , representing 50-55% of the
original capacity at that elevation.

188

[
[

[
[
r

I
I
I

L
[

1
r

APPENDIX 3

A3.ll.3. Downstream impacts


Between 1960 and 1964, when the reservoir was operating with a high trap
efficiency and releasing essentially clear water, there was degradation of the
river-bed in the lower Yellow River, posing a danger to the fioodbanks through
undelmining.
Between 1964 and 1973, when extensive modifications vvere being made and
sedimentation in the reservoir basin was being brought under control, the limited
discharge capacity of the bOttOlTI outlets meant that high concentrations of
sediment were being released, but that the discharges were generally insufficient
for its conveyance, leading to severe sediment accretion over a length of
300-400 kIn downstream of the dam, reducing the flood discharge capacity of
the main channel and increasing downstream flood risks.
Only after the excavation of the two bypass tunnels, the conversion of the four
power intakes and t..~e reopening of the eight diversion outlets, could both
sediment and water be discharged at the rates required to stabilise conditions
along the lower Yellow River. However, channel deposition continues to occur at
low discharges, and the regulation of sediment in the lower Yellow River is a
problelTI that is not fully solved (Mon-is and Fan, 1997).

A3.12. VVARSAK RESERVOIR (Pakistan, 1960)


Built in 1960 on the Kabul River, a tributary of the Indus, Warsak reservoir was
forn1ed for irrigation and hydropower and impounded by a 76 m high concrete
dam. The crest of the gated spillway is 12 m below the highest impounding
level.
reservoir basin is 42 km long and has an original storage capacity of
170 M.m3 The average fu"1nual inflow to the reservoir of 21 100 M.m3 is 124
tilnes the original storage capacity.

A3./2.1. Sedimentation
During the period 1961-70, the average measured suspended-sediment concentration was 727 mg/l, giving an average sediment inflow of 153 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 42 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
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.

A3.12.3. Downstream impacts


After a relatively short period while the reservoir filled with sediment and
concentrations in the downstream river were reduced, it appears that the
downstream effect of the presence of Warsak reservoir is currently minor. If
flushing were to be instigated, the local effects could be severe, but downstream
of the confluence with the much larger Indus River (which has been starved of
high sediment loads since the impounding of the Tarbelareservoir in the 1970s),
the effects could be beneficial in diminishing degradation of the river bed.

A3.13. OUCHI-KURGAN RESERVOIR (Former USSR, 1961)


Ouchi-Kurgan is a 17 Ian long reservoir, used for the purpose of inigation and
power production, which began impounding in October 1961 and had an original
storage capacity of 564 M.m3 (UNESCO, 1985). The catchment area is not
given, but the annual run-off is given as about 15000 M.m3 (Atkinson, 1996),
which is over 250 times the original storage capacity.

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.
r

190

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.

A3.13.3. Downstream impacts


No information available.

A3.14. SEFIO-RUD RESERVOIR (Iran, 1962)


Sefid-Rud reservoir in Iran was constructed in 1962 for irrigation and power
generation. The dam is a buttress-type concrete gravity structure, with a
maximum dam height of 106 m. There is a 'morning glory' service spillway and
a gated auxiliary spillway (}\IloITis and Fan, 1997).
The catchment area, which is largely semi~arid, with annual rainfall of
250-400 mmlyr, totals 56200 krn2 , yielding an average annual inflow to the
reservoir of 5008 M.m3 . The reservoir is located at the junction of a major and
minor catchment and the basin also has major and minor branches.
The maximum initial reservoir depth is 82 ill, the length of the major branch
is 25 km and the original storage capacity of the reservoir was 1760 M.m3,
representing 35% of the mean annual inflow. The normal impounding level is
27165 ill with a maximum of 27625 m (presumably the estimated peak flood
level). During normal operation the minimum drawdown level is 240 ill,
equivalent to a draw down of about 40% of the original basin depth.

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 365 M.m3 per annum, equivalent to an annual
rate of 21 %. The trap efficiency during this period was estimated as 730/0, with

19/

EVACUATION OF SEDIMENTS

most of the discharged sediment complising density CUlTents. The sediInent


composition was 33% sand, 47% silt and 20% clay.
The reservoir capacity reached a nlinimum of about 630/0 of the original in
1982-83, before recovering as a result of the flushing measures.

A3./4.2. Flushing
The reservoir was built with three bottom outlets on the right-hand side (total
discharge capacity 430 m3/s, elevation 1913 m) and two bOttOlTI outlets on the
left-hand side (550 m 3/s, 1938 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 19% 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

Table A3.3.

Summary of sediment flushing at Sefid-Rud reservoir

F1ushing
year

Dra\vdown flushing
duration: days

Empty flushing
duration: days

'Water volume
used: M.m'

Water volume used as


% of annual inflow

Sediment
removed: Mt

1980-81

61

536

10

24

1981-82

6S

390

11

12

1982-83

117

10

1513

26

52

1983-84

16

80

795

23

68

1984-85

19

138

1810

29

142

1985-86

18

129

1131

29

46

17

85

26

27

1987-88

24

86

1988-89

113

1989-90

103

681

351

744

10667

1986-87

Total

942

]812

22

1057

31

54

22

32

21

514

57

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.
Lateral erosion by piping

SediInents in the upper reaches of the reservoir were deposited in thick


alternating layers of cohesive and non-cohesive soils and the resistance to erosion
of the cohesive materials had been enhanced by the desiccation and compaction
caused by annual drawdown. However, collapses in the cohesive materials were
noticed near the areas of the main channel, which appeared to have been
triggered by piping and washout of the sand sediments beneath the cohesive
deposits . .
In the 1985-86 flushing season, field experiments were set up to induce and
enhance piping artificially. A pit 2 In in diameter and 4 m deep was dug into the
sandy deposits 40 m away from the main channel, which was kept full of water
by pumps during the flushing operation. After 15 days, the top cohesive layer had
collapsed into the main flushing channel, creating a gully 3-10 m deep and
6-15 m wide. It was concluded that a hydraulic gradient of 025 to 033 between
the pit and the main channel was sufficient to initiate piping in these deposits.
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 12 and 22 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
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
by drawdown flushing alone.
The pilot channel was 76 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
I

.--l

._- /

:--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.

_I

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
possible to recover lost storage and maintain a long-telID storage capacity of
about 90%, compared with about 75% by flushing alone.

A3.14.3. Downstream impacts


In the case of the Sefid-Rud reservoir, the intakes and canals used for diverting
reservoir releases to irrigators cannot tolerate high sediment loads. To prevent
downstream sedimentation problems, the sediment concentration should not
exceed 5 gil dudng the irrigation season. The sluices on irrigation ba.7ages
remain open during the flushing season, thus passing the sediment-laden flow
with minimum interruption, while the irrigation intakes remain shut to exclude
the flushed sediment from the delivery canals.
No information is given in the literature regarding potential environmentat
impacts downstream of the reservoir.

A3.IS. KHASHM EL GIRBA RESERVOIR (Sudan, 1964)


The Khashm EI Girba dam, sihlated on the Atbara River in Sudan, was
completed in 1964 and is used for power generation, irrigation and water supply.
The reservoir had an original storage capacity of 950 M.m3 , but no information
is available regarding the catchment area or mean annual inflow.

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

Table A3.4.

Khashm El Girba reservoir, sediment flushing

Flushing period

Water used:
M.rn 3

Silt inflow:
Mt

Net sediment
release: M.rn 3

11-14 July 1971

612

35

175

29 July-2 Aug. 1973

545

33

125

According to EI Hag (1980) and EI Faith Saad (1980) (quoted in UNESCO,


1985) the sediment outflow each July, including the flushing operation periods,
was 85 Mt, which is about the same as the average estimated annual sediment
inflow.
Unfortunately, the 1980 references have not been obtained, so no fUlther
details are available of the apparently successful flushing operations carried out
at the K.l1ashm EI Girba reservoir.

A3.15.3. Downstream impacts


No information available.

A3.16. HENGSHAN RESERVOIR (China, 1966)


The Hengshan reservoir is used for flood control and in-igation. It is a small
gorge-type reservoir, 1 to 2 kIn in length (the references disagree on this detail),
located in an arid zone of scarce water supplies (Mon-is and Fan, 1997). The
concrete arch dam is 69 m high and the maximum water depth is 65 m (Atkinson,
1996), giving the reservoir an original storage capacity of 133 M.m3 (UNESCO,
1985), which is rather less than the reported mean annual nln-off of 158 M.m3

A3.16.1. Sedimentation
From 1966 to 1973, the first eight years of the reservoir's operation, 319 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 002 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, 26 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

Ij

i
i

I
I

Ij
I

discharge 145 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, 08 M.m 3 of sediInent was
removed from the reservoir. The reservoir was then impounded for five years to
June 1979, before flushing for the second time for a period of 52 days during the
flood season. The second flushing period removed 103 M.m3 of sedilnent,
reducing the volume of sediment in the reservoir to 262 M.m3 (20% of the
original storage capacity). Emptying and flushing were subsequently undertaken
in 1982 and 1986.
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.

A3. I 6.3. Downstream impacts


No information available.

A3.17. CACH! RESERVOIR. (Costa Rica, 1966)


Cacm hydropower reservoir, located on the Reventazon River, was the first major
hydropower scheme on the river and several others are planned to be built
downstream. The reservoir was completed in 1966 with the construction of a
76 m high concrete arch dam, it has a surface area of 324 ha, is 6 kIn in length,
with a maximum depth of 69 ill, giving an original storage capacity of 54 M.m 3
The mountainous catchlnent of 785 km2 produces a mean annual run-off of about
1500 M.m 3 , or 25-30 tirrles the original storage capacity. The catchment area is
heavily vegetated, about 55% is forest and most of the remaining area is
agricultural.

197

J
EVACUATION OF SEDiMENTS

,J
A3./7./. Sedimentation
The average annual sediment inflow is 081 Mt, which would have a deposited
volume of the order of rather over 1% of the original storage volume. The annual
load is estimated to be distributed as follows:
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 reservoir allowed to refill, typically taking between 16 and 21 days. The amount of

198

J
J

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
Table A3.5.

Sediment released by 14 flushing events at Cachf reservoir


Quantity flushed: t

Date

Slow
drawdown

Rapid
drawdown

Freefiow

Total

Oct. 1973

Aug. 1974

186200

225200

411400

Oct. 1975

Oct. 1977

40700

44000

19500

5000

Oct. 1981

14600

348900

113400

476900

Oct. 1982

5800

111 600

250900

386300

402400

114300

May 1980

Sept. 1983

I!

28700

84700

24500

545400
I

665500

Oct. 1984

23300

604600

June 1985

July 1987

Sept. 1988

61600

627 000

577100

Sept. 1989

42400

144300

482200

r---moo

Gsoo

278 700

347 100

653000

Oct. 1990

32600

1265700

Note: Dashes indicate no data available; italics indicate an estimate

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.

A3.173. Downstream impacts


During each flushing operation, in which peak concentrations exceeding 400 gil
have been measured, it was observed that sub.stantial amounts of sediment were
deposited on lower floodplain areas and bars between the dam and the Caribbean
and a stratified plume of turbid water was observed in the sea. It was expected
that the riverine deposits would be eroded by subsequent floods. No studies have
apparently been can-ied out on the effect of the sediment on downstream biology,
although anecdotal reports from local observers suggest that the concentrated
sediment releases ,cal:lse extreme mortality to all types of river biota.

A3.18. GEBIDEM RESERVOIR (SWitzerland, 1968)


Gebideln hydropower reservoir is situated in the Swiss Alps on the Massa River,
a tributary of the Rhone. The dam comprises a thin 122 m high arch. The
reservoir is 1 5 km long, with a Inaximum depth of 113 m and storage capacity
of 9 M.m3 The catchment is 200 km2, of which 65% is occupied by a glacier.
The average annual inflow to the reservoir of 429 M.n13 is almost 50 times the
impounding volume. The annual discharge is seasonal, being dominated by
snowmelt, glacial melt and sumnler storms between May and October, with
negligible flows in the winter months (Morris and Fan, 1997).

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:
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).
I)

In preparation for flushing, the reservoir level is lowered to the minimum


operating level by releasing water through the turbines, which are then closed ..
Drawdown flushing is then initiated over a period of two hours by first opening
one gate then the second gate, progressively raising the discharge from 10 m3/s
up to about 60 m3/s. It takes between 3 and 6 hours for free flow conditions to
be achieved at the outlet, typically at discharges of between 10 m 3/s and 20 m 3/s.
In some years, the outlet gates are interrrJttently closed and the reservoir is
allowed to fill for 20 minutes, then the' gates are fully reopened; the resulting
raising and lowering of the water level facilitates flushing of deposits in the
vicinity of the dam.
. .-.
Tables A3.6 and A3.7 provide a summary of the flushing carried out at
Gebidem dam from 1982 to 1993 and a sediment balaIlce for 1990-91.

A.3. f 8.3. Downstream impacts


The combination of sediment release and reduced discharge due to diversion for
power production has heavily impacted the Massa gorge downstream of the dam.
By
end of the 1992 flushing, it was found that the lined conveyance channel
on the gently sloping reach just upstream of the confluence with the Rhone had
completely filled with coarse
This was attributed to the fact that the
flushing flows Vlere typically half those originally anticipated during the design
studies. It was therefore recornmended that the flushing flows be increased.

201

EVACUATION OF SEDIMENTS

Table A3.6.
Year

Summary of the flushing at Gebidem, dam, 1982-93

Duration:
h

Water
volume used:
M.m 3

Mean
flushing
flow : m 3/s

Gate
operations *

1982

56

238

11 8

1983

48

338

196

Sediment
removed:
M.m 3

Solids
concentration: t

0143

60

0175

52

0178

60

1984

68

297

121

1985

49

250

142

0150

60

1986

45

353

218

0212

60

1987

45

320

198

13

0192

60

1988

79

293

103

13

0176

60

1989

49

249

141

0150

60

1990

40

318

. 221

12

0191

60

1991

96

235

68

0270

115

1992

151

328

60

61

0197

60

1993

101

248

68

0260

105

* Number of intennittent closures and reopenings used


t Where value is 6%, this is an assumed value used to estimate the sediment outflow

Table A3. 7.

Sediment balance at Gebidem dam, 1990-91 (12 months)

Parameter

Volume: m3

Annual sediment release


Sediment released by flushing (955 h)
Sediment passing through turbines

270000
70000

. ~

Fate of sediment flushed

Deposited in gorge

81000

Deposited in aggregate works*

32000

-it

Delivered to Rhone
Delivered to Rhone via turbines
Total

* Aggregate supplier's extraction operation in river-bed downstream of dam

202

157 000
70000
340 000

f[

tf

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.

A3.19. SANTO DOMINGO RESERVOIR (Venezuela, 1974)


The Santo Domingo hydropower reservoir was formed by the construction of an
80 m high arch dam at the confluence of the Santo Domingo and Aracay rivers.
The reservoir basin has two branches of similar length, both tributary valleys that
have steep slopes, with river-bed gradients of the order of 4% to 6%. The valley
bottoms are typically 30 m to 50 m wide (Krumdieck and Chamot, 1979).
The maximum water depth is about 65 m, the longer branch is about 1 Ian long
and the surface width is typically 100 m, giving a gross original storage capacity
of about 30 M.m3 The combined 427 km2 catchment of the two rivers, which is
mostly covered by tropical vegetation and agriculture, produces a flow regime
that allows the power station to operate continuously between April and October
with an average inflow of 20-25 m3/s. In the dry season this is reduced to 5 m 3/s,
so the. reservoir storage is used to support generation to suit the daily pe3.J."<:
demand.

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 058 M.m3 of deposition had occurred over two flood seasons,
suggesting a rate of accretion of 029 M.m3 per annum, which was rather greater
tt~an the design estimate of 024 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 32 x 25 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 freeflowing conditions, which_ are described below.
Pressurised flushing. In this condition, the bottom outlets are submerged by
sediment and flushing is started by inducing a 'piping' failure of the overlying
sediments, which are discharged through the outlet at a high concentration;
This causes sufficient sediInent to be eroded in order to provide a clear
pathway to the bottom outlet. Special attention should be paid during the
operation to check on the size of sediment load and to ensure that an adequate
flow is discharged, capable of transporting the material to the river-bed
downstream.
Free-flow flushing. This is used when the outlets are clear of deposited
sediment and usually begins when the level of the reservoir is already low and
the sediment load is moving towards the central channel of the reservoir. Freeflow conditions are capable of flushing greater sediment loads, but at the cost
of consuming greater volumes of water. Experiments showed that 12 000 n13
to 15 000 m 3 of sediment could be passed through the bottom outlets per day.
., First flushing operation. The first flushing of Santo Domingo reservoir took
place in May 1978, after four years of operation, when the powerhouse was
closed to enable a complete inspection of the reservoir and the plant to be
undertaken. It was estimated that the bottoin outlets flushed between 50% and
60% of the deposited sediment in a period of only three or four days of freeflow flushing, at a time when the inflow was 8-10 m 3/s. The bottom outlets
were covered by deposits by the fourth day, so the reservoir was allowed to fill
overnight, following which flushing under pressure cleared the botton1
outlets.
Some three weeks later, with inflows remaining low in both rivers, it was
decided to attempt to accelerate the flushing operation by using two bulldozers
to move deposits out of reach of active erosion towards the main strealns of
each river.
The entire flushing operation was sufficient to relTIOVe the majority of the
sediment which had been deposited in the reservoir basin over the four years
of operation. A subsequent topographic survey established that 062 M.m 3 of
deposits had been removed, restoring the storage volume to 2 85 M.m3

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

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 freeflow 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).
e

A3.19.3. Downstream impacts


It was found that7 for relatively low flushing discharges, the concentration of
sediments released could exceed the capacity of the downstream channel to
convey theIn, resulting in accretion starting to obstruct the outlets. No
information is given regarding environmental and other downstream impacts.

A3.20. NANQiN RESERVOIR (China, 1974)


Nanqin reservoir, used for flood detentiQfl and irrigatioDLis situated i..11 the hillymountainous Shaanxi Province in southelTI China. The lnaxilnum depth (up to an
impoundment elevation of 124 m) is 29 ill and the length of the reservoir is
45 km. The original storage capacity of the reservoir was 102 M.m3 (Chen and
Zhao, 1992), which is 8% of the mean annual inflow.
The reservoir's history can be divided into three phases:
(a) between 1974 and 1976 it served solely as a flood-detention reservoir
(b) betw'een 1976 and 1983 flows were impounded to a middle level of 110 In
a..l1d released for irrigation
(c) since 1984, with an improved regime of sediment management.

A3.20.1. Sedimentation
The mean annual suspended-sediment load inflow is given as 053 Mt, the vast
majority of which enters during the flood season, July to September. The annual
average covers a wide annual variation, of between 012 Mt and 134 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.

Density current venting


Due to the steep bed slope of the Nanqin reservoir, density currents can easily
"- -re'ach the dam and form a secondary reservoir of turbid water beneath the clear
water layer. Sediment sluicing by density current venting began at Nanqin
reservoir . in 1977. Between 1977 and 1984, out of the 243 Mt of suspended
sediment entering the reservoir, it is reported that 156 Mt (64%) was discharged
successfully by this method.
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
072 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.
It was estimated that, if these principles are observed, a long-term storage
capacity of the order of 75 M.1n3 (74% of the original) can be sustained. No
details were given of the success of subsequent flushing operations.

206

APPENDIX 3

A3.20.3. Downstream impacts


No information available.

A3.21. ICHARI RESERVOIR (India, 1975)


Ichari dam is a 60 m high concrete gravity dam constructed across the River
Tons, a tributary of the Yamuna, in 1972 (Mohan et a!., 1982). The reservoir,
which is used for hydropower generation and was first impounded in 1975, is
113 kIn long, has a maximum depth of 37 m and an original storage capacity of
11 55 M.m3 The mean annual inflow of about 5300 m}/s is about 450 times the
original storage capacity.

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 049 an.d 29 M.m3, with a median value of
22l\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
(June to
May)

Total
water
inflow:
M.m3

1975-76

Total
sediment
inflow:
M.m3

Sediment
trapped:
M.m3

262

1976-77

5049

163

061

1977-78

6455

371

012

1978-79

7825

2902

186

207

114

capacity: M.m3

Dead

Live

Total

393

500

893

340

492

832

480

820

425

634

160

360

520

209

.l

1979-80

3420

f
::

4585

480

060

114

346

460

5445

109

-{)41

108

393

501

1982-83

4716

049

035

108

358

466

1983-84

5148

243

023

120

323

443

1980-81
1981-82

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.

A3.21.3. Downstream impacts


It was reported that there is neither silting nor appreciable scouring in the

downstream reach of the liver. The sediment flushed from the reservoir is carried
by the water discharged. No details are given of possible environmental
impacts.

A3.22. BAtRA RESERVOIR (India, 1981)

Baira reservoir forms part of a hydropower project which utilises the combined
flow of three tributaries of the River Ravi in the nOl1h west of India. The 51 m

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 01 % 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 0092 M.m3 , but in the first 18 months of operation, a silt
volume of 045 M.m3 had accumulated, representing allnost 20% of the original
capacity and suggesting an annual sediment load of at least 03 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
e

1/

CD

the reservoir was drawn down to the minimum normal operational water level
(1113 m), following which flows to the powerhouse were stopped
the diversion tunnel was opened to allow a discharge of 150 m 3/s until
reservoir was almost empty
the diversion service gate was opened fuHy
water was fed to t.~e reservoir from,the Suil and Bhaledh to clear silt from
around the associated structures
flushing ceased when the concentration in the discharge had decreased to
about 10 gil
the diversion tunnel gate was closed and the reservoir refilled.

The duration of the flushing operation (from ceasing to resl1nung power


generation) was about 40 hours, with a maximum sediment concentration of
380 gil. The volume of
removed was estin1ated to be 038 lYLm3 ,
representing over 80% of the
which had occurred since impounding.
The recommendations made
future flushing were that it should be carried
hours and that it would be more
in
out once a year for a period of
APlil or iVlay, when the discharge from the Baira is about 100 m3/s (Jaeggi and

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.

A3.22.3. Downstream impacts


No info1111ation available.

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


per rainfall: mm/hr

Number of falls of rain

Average erosion: kg/m2

40

037

255-508

6]

06

509-762

40

118

763-1016

19

114

1017-1270

l3

342

1271-152-4

363

1525-1778

387

177 9-2540

479

0-254

],
j
i

Many studies have been carried out to determine a relationship between


erosion and rainfall and SOine of these are illustrated in Figure A4.1.
The relationship proposed by Langbein and Schumln (1958) between annual
sediment yield and effective precipitation has been widely documented and
utilised (Walling and Webb, 1983). Maximum sediment yields occur at an annual
effective precipitation of approximately 300 nun (i.e., semi-arid regions). In
areas with higher rates of effective precipitation, vegetation growth is increased
and the sUlface is protected. In more arid areas there is insufficient rainfall to
move materiaL The relationship uses the term effective precipitation, defined as
the annual precipitation required to generate the given annual run-off at a
standardised mean temperature of 50F and not the standard definition of
precipitation minus evapotranspiration. There are many possible sources of error
in this study, the main one being that the results are based on only 94 sampling
points in the USA.
A number of alternative relationships have been deIived. Results produced by
Judson and Ritter (1964) were based on the average suspended-sediment yield
and annual run-off for the seven major drainage regions in the USA. Work by
Dendy and Bolton (1976) used the group-averaging technique to generalise the
relationship based on .data from 500 locations within the USA. The peak
sediment yield occurred at a mean annual run-off of 25 lnrnto 75 nun which is
equivalent to an effective rainfall at 50F of 450 nun to 500 mm, higher than the
300 mm at which peak yields were predicted by the Langbein and Schunun
curve.
Relationships based on global data show less similarity to the Langbein and
Schumm curve. There is a general increase with precipitation at the lower end of
the scale but values increase again when annual precipitation and lun-off exceed
1000 mnl and 500 nml respectively. The relationship produced by Wilson (1969)
has two peaks at 750 mm and at 1750 mm of annual precipitation. These coincide

212

~.

APPENDIX 4

with sub-hulnid and tropical conditions and contradict the_peak demonstrated by


Langbein and Schumm for semi-arid regions. The relationship proposed by ,
Tabuteau (1960) demonstrates a more complex pattern with a wide range of
yields in all regions.
A study by Walling and Kleo (1979) based on data fron1 1246 global
measuring stations showed no clear pattern between mean annual precipitation
and mean annual suspended-sediment yield (see Figure A4.2).
A relationship between precipitation and sediment yield group-average data
for drainage basins with an area less than 10000 km2 can be seen in Figure A4.3

800

>-

C)'

><'t'

600

.:s::

~
Qi

Qi

';:;"

"E
CD
E

E
'5 200
CD

's:,

400

"E

tlJ

'6

U)

en

250

500

750

1000 1250

Effective precipitation: mm
(i)

1250

>-

c::::

-s:,

a:;

1::
tlJ
E

1:: 500
tlJ
E

a:; 100

10

'5

1'

tlJ
(j)

150

100

10

100

1000 10000

400

800

1200

1600

(ii)

L 1200
>-

~ 800

-c
Q.i

':;'

'5-

C
<ll
E

50

en

25:

(i)

><'t'

'6

750

';:;"

-L

-c
a:;

l-

tlJ

1200

1000

;;'1000

oX

en

800

-L

1000

'6

600

Mean annual run-off: mm


(ii)

(a)

10 000

C\I

400

200

'E 400
<ll
E

'0
Q)
(j)

200

400

600

800

1000

400

(iii)

Figure A4.1.

800

1200

1600

2000

2400

Mean annual precipitation: mm

Mean annual run-off: mm


(b)

(iv)

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 .
Run-off. Run-off results from an excess of precipitation over the sum of
infiltration and evapotranspiration and is the quantity of water available to
convey the products of erosion. Factors that encourage infiltration and thereby
reduce run-off will decrease the quantity of erosion. Run-off may provide a
better correlation with rates of erosion than precipitation does. Low run-off
rates indicate aridity and hence poor vegetation cover compared with high
rates that indicate an excess of water and therefore dense vegetation cover. A

10000
'0

....
:.

>-

.
. ~:
: ..

..

00

1000

i:i

Q)
':;"

c
Q)

E
:0
(!)
C/)

"C

Q)

100

"C

c:
Q)
c..
C/)

:J
C/)

(ij
:J

c:
c:

ns
ns
Q)
:2
c:

10

1L-__

L-~

200

__

400

__

600

__- L_ _- L_ _- L_ _

800

1000

1200

__

1400 1600

__

1800

~~~

____

2000 3000 4000

____

6000

__

__

8000

Mean annual precipitation: mm

Figure A4.2.
tion

214

Mean annual suspended-sediment yield versus mean annual precipita-

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 evapotranspiration 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-

j\ .. .. \. //

4DO

\.of

;/0 0 \./

Qi

's;.

400

800

Q)

ca::l

1600

1200

2000

2400

2800

3200

3600

4000

Mean annual precipitation:mm .

'6
<ll
en

o~

0/\ ./

800

1200

c::

c::

ctI

Q)

800

400

------------

~.~./.

0\./ . 0
1L-____- L______L -____

400

800

______

1200

1600

____

______

2000

2400

____

____

2800

~~

3200

Mean annua! run-off: mm

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:


m3/km 2/yr

Resistant: conglomerate, limestone


and resistant sandstone
Medium: fliable sandstone
Soft: shale and gypsum

Sediment loss, New Mexico/


Arizona: m3/km 2/yr

143

95-143

571

523

1237

761

A4./J. Geology
Rock type
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
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 14 (Dedkov and Mozherin, 1992).

J
.J
J

Volcanic and tectonic activity


Recent tectonic activity influences the intensity of erosion especially in
mountainous areas. An increase of approximately two fold was found for an
increase of earthquake activity by 1 point in USSR (Dedkov and Moszherin,
1992). High rates of erosion are seen in zones of young mountains, e.g.
Himalayas and New Zealand.

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
influencing erosion rates are the texture, structure, organic matter content, shear
strength and infiltration capacity.

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

.,.....L

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

,...

A4.1.S. Catchment characteristics


Slope
gradient and length of the slope directly influence the velocity of run-off and
its erosivity. The energy and, therefore, the erosive potential of flowing
water increases with the square of the velocity (Goldman et al., 1986). Long,
continuous slopes allow run-off to build up momentum and the base of the slope
becomes more susceptible. On a flat surface the raindrop splash is random,
nTt-,,,,,,.,::l-:lC on sloping ground more raindrops are splashed downSlope than upslope
moving sediment with them.
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.

Drainage basin area


An inverse relationship has been demonstrated between sediment yield per unit
area and catchment area (see Figure A4.4). In larger catchments there is a lower
overall slope, smaller percentage of erodible rock and more opportunity for
sediments eroded from steeper slopes to be deposited in the floodplain (White,
1982). In China approximately 50% of the sediment load from the Yangtze River
is produced by 13% of the catchment and 43% of the sediment in the Yellow
River comes from 7 % of the area.
A number of recent articles have questioned the standard relationship between
suspended sediment and drainage area. Church and Slaymaker (1989) suggested
that in British Colon1bia specific sediment yields increased downstream in
catchment areas of up to 30 000 km2 due to remobilisation of quaternary
sediments stored in the valley and channel systems (Walling and Webb, 1996).
Dedkov and Moszherin (1992) proposed that river systems are characterised by
positive or negative relationships depending on the relative importance of
channel and slope erosion. Where channel erosion dominates, in areas with dense
vegetation cover, erosion rates increase downstream. There is a positive
relationship with drainage basin areas due to greater entrainment and
transportation of sediment. Where slope erosion is dominant, erosion is
concentrated in the headwaters and a proportion of the mobilised sediment will
be deposited during transport through the system. There is therefore an inverse
105

x
x

104

i::i
05
";;:..
C
~
'5
Q)

103

o
o
o

102

1::..

o
+OX

10 1

o
+

q.
0

C/)

1::..0

-k

1::..

10oL-~--~---L--~--~~~~--~--~--~--~~

10-3

10-2

10-1
Drainage area: 106 km2

Figure A4.4.

218

Sedilnent yield versus drainage area

100

101

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 rainfallare 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 OJ:) and 4
experienced mean annual erosion rates of 0015, 002 and 0003 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 08, 26 and 90 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
In semi -arid regions the
response and recovery may take a long period of
recovery time is four times that of humid areas (Walling and Kleo, 1979).
on erosion rates can
The effect of cultivation or lack of soil cover by
be seen in Table A4.3.
to bare ground can also
The increase erosion rates from natural
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)

China'

Natural

Cultivated

<02

15-20

Bare soil

28-36
I

USA

05-17

0003-03

04-90

0003-002

001-90

10-750

Nigeria

005-01

001-3 5

03-15

India

005-01

003-20

10-20

Belgium

001-005

03-30

07-82

UK

001-005

003--03

10-45

Ivory Coast

Human impact

It is estimated that human activities have degraded 15% (200 million ha) of the
land between 72N and 57S. 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:

Ungrazed thicket

Soil lost by erosion,


t per acre

Ot
19%

rI

Water lost by run-off


per cent of rainfall

260%

50-4%

Figure A4.5.

220

Soil erosion at Mpl,vanga, Tanzania

J.
. i

APPENDIX 4

Table A4.4.

Erosion rates for different land use categories (Morris and Fan, 1997)

Land Llse

Forest

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 14 times; however, larger increases in lowland areas
have reduced the difference between mountainous and low land regions to 32
. 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.

A4.2. C.A.SE STUDIES OF EROSiON RATES

A4.2.1. Erosjon rates in Africa


The annual sediment yield of rivers in Africa for drainage basins around
10 000 kIi? is between 1 tJkrn2 and 4000 t1km2 Four-fifths of the surface area of
Africa produces less than 100 t/lan2/yr, highlighting L,e regional variation in
erosion rates (Shahin, 1993).

The Upper Tana basinJ eastern Kenya


The area of the Upper Tana basin is 9250 km2 and can be subdivided into three
areas depending on altitude:
e
$

10

Mount Kenya volcanic summits and the Aberdare range


slopes of J\.1ount Kenya
11wea-M~singa plains.

221

I'

EVACUATION OF SEDIMENTS

Table A4.S.
1983)

Increases in sediment yield due to land use changes (Walling and Webb,

Region

Land use change

Factors for
increase in
sediment yields

Source

Rajasthan, India

Overgrazing

4-18

Utah , USA

Overgrazing

10-100

Noble (1965)

Oklahoma, USA

Overgrazing and cultivation

50-100

Rhoades et al. (1975)

Oklahoma, USA

Cultivation

5-32

Rhoades et al. (1975)

Texas, USA

Forest clearance and cultivation

340

Chang et al. (1982)

Northem California,
USA

Conversion of steep forest to


grassland

5-25

Anderson (1975)

Mississippi, USA

Forest clearance and cultivation

10-100

Southern Brazil

Forest clearance and cultivation

4500

Bordas and Canali (1980)

Westland,
New. Zealand

Clearfelling

O'Loughlin et al. (1980)

Oregon, USA

Clearcutting forest

39

Fredriksen (1970)

Sharma and Chattelji (1982)

Ursic and Dendy (1965)

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.
Altitude: m

Vegetation

Rainfall: mm

Soil

>1800

1800

Dense evergreen forest

Clay loam

>1400

1400-1800

Steep cultivated slopes

Clay loam

1100-1400
<1000

222

Variations in rainfall, vegetation and soils with altitude (Ongweny, 1979)

900-1400

Farming
Marginal farming

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.

E.rosion rates in the grazing lands of Kenya


The Athi-Kapiti plains within a 50 km radius to the south and east of Nairobi
consist of a dissected plateau developed on cernozoic tuffs and lavas. The mean
annual rainfall is 5 to 700 rom with grasses that cover 40-90% of the ground
below a sparse (5%) canopy cover. The soils vary from planosols on the ridges
through phaeozems to vertisols on the footslopes.
180 km south east of Nairobi, quaternary lavas extend northwards as a stepped
plateau from the slopes of Mount Kilimanjaro where the meatl arulual
precipitation is 450 mID. Soils are 10 cm to 200 cm thick sandy-clays. The
ground cover is less than 10% with a sparse (10-30%) canopy of dry woodland
and bush.
Between the two volcanic plateaus lies a belt of precambrian basement schists
north of the Amboseli basement. The region receives 300 mm of rainfall and the
vegetation cover is grassland (10-40%) and bush with a canopy cover of up to
40%. The soils are sandy clays between 50 cm and 150 cm deep.

-- --

__~-L__~~~~~~--~~--~--~~--~
100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500
Mean annual run-off: mm

1~~~

III 0 Forest

r:.

Forest> Agriculture
Agriculture> Forest

Grazing
Scrub lorest

1-UpperTang values plotted on Dunne's curve


2-Dunne's values

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 01. 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).

Sediment loads in the Orange River, South Africa


The sediment load is predominantly small particles less than 02 mm. The
average sediment yield from sub-basins of the Orange River vaIies,from less than
10 t/km2/yr to lTIOre than 1000 t/km2/yr. The highest yields are from the
solonetzic soils and valleys filled with sediments on the westelTI side of Lesotho
with lowest yields from the sandy permeable regions in the north-west drainage
region. Soils ~roIP- areas with a high silt and:-Iow clay content are generally the
most erodible (Rooseboom and Van Harmse, 1979)

A4.2.2. Erosion rates in Thailand


The rates of sediment production in Thailand range from a value of 8 t/km2/yr to
3874 t/km 2/yr with the highest rate experienced in the Lam Dome Noi River
catGhment, a tributary of the Mun-Chi River. The lowest rate was in Huai Pa Tao
River Basin, a tributary of the Chi River (see Table A4.7).
The highest rates seen in the north-eastelTI region are due to the highest rates
of deforestation (Jantawat, 1985).

A4.2.3. Erosion rates in China


The erosion rates in China are highly variable. High rates occur in northern
China where silt contents are high, the climate is dry and there are extensive
areas of loess. The climate in southelTI China is generally \vet and warm
producing a better vegetation cover and therefore lower erosion rates (Lagwanker
et a/., 1995).

Table A4.7.
Region

224

Erosion rates, Thailand


Rate of erosion: tlkm 2/yr

NOlthern

12-2045

Central plains

20-570

East

27-356

South

30-1787

. - :!

. i

-~.

APPENDIX 4

Table A4;8.

Sediment contribution for sections of the Yellow River (Tal Wei Soong and
Yean Zhao, 1994)

Upper

Area: km'

385000

Length of
channel: km

Middle

Contribution:
% water

Contribution:
% sediment

111000

48-7

9-0

36-6

893

345000

1206

111400

22000

786

1/8000

Lower

3472

Average
slope: m/m

11

17

Lower Yellow River


The Yellow River is the second largest river in China at 5464 km long with a
drainage area of 752400 k..rn? The sediment contlibution is mainly from t.l}e
middle reaches that pass through the loess plateau (see Table A4.8).
The rainfall is seasonal with the majority occurring between July and October.
The run-off during this period can account for 60% of the annual discharge and
sediment inputs 85% of the annual sediment totals. The rainstorms are generally
of high intensity and short duration and sediment delivered from the catchment
reaches 911 kg/m3 (Tai VVei Soong and Yean Zhao, 1994).

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 16 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).

A4.2.4. Erosion rates in India


A relationship between drainage basin area and average annual silt deposits was
calculated by Lagwanker as follows:
Catchment area above 2500 km2
Catchment area between 2500 and 100 km2
Catchment area below 100 lan2

S = 0-065 A
S = 0278 AO- 8i5
S = 0200 A 0-887

where S = average annual silt deposits in ha per m


A = catchment area in km2
The relationship was based on selected reservoir sites in various areas of India
(Lagwanker et aI., 1985). '

225

EVACUATION OF SEDIMENTS

Erosio-n and sediment yield in Krishna River Basin

The Krishna River rises in the Western Ghat mountains at an elevation of 1400 m
and flows 1400 k:m to the Bay of Bengal. 40% of the area is mountainous. 80%
of the basin is formed on archaen and younger crystalline rocks and 20% on
Deccan Traps and recent sedin1ents. The discharge is 3 m 3/s to 3400 m 3/s with the
majority of the flow occurring during the monsoon in July and August. The
maximum sediment yield occurs between July and October with 95% of the
annual sediment load derived froln the monsoon period. The erosion rate is
highest in small sub-basins of the Krishna dver where rates are up to
4000 t1km2/yr (Subramanian, 1982).

A4.2.S. Erosion rates in the Himalayas


The Sapt Kosi is the third largest river with a source in the Himalayas. The three
main tributaries are the Sun Kosi, Arun and Tamur. Precipitation falls 89% as rain
and 11 % as snow with 80-85 % of the total rainfall in the monsoon months from
June to August. 85% of the run-off occurs between June and September with
sediment load 98% of the annual total. The area to the gauging station is
59 000 km2 with an average annual run-off of 53 000 M.m3 The average annual
sediment yield is 2800 t/km2/yr composed of 16% coarse sand, 29% medium
sand and 55% silt and clay (Mahlnood, 1987).

A4.2.6. Erosion rates in Pakistan


The River Indus carries large alnounts of sediment particularly in the spring and
summer when melting snow leads to high rates of erosion. The source of the river
is in the Tibetan Plateau at a height of 5500 m above sea level. There are two
distinct hydrological regions in the 169 650 km2 drainage basin upstream of the
Tarbela dam. About 90% lies between the Karakoram and Himalayan mountain
ranges, from which the meltwaters contdbute a significant proportion of the flow.
About 10% iImnediately upstream of the dam is subject to monsoon rainfall
between July and September. The average annual rainfall is 890 mm with twothirds occurring between June and October. The climate is subtropical and
semi-arid in the headwaters. The summer season is from April to September and
the winter. from October to March. Average daily temperatures range from 7C
in January to 41C in June. The mean annual sediment inflow to Tarbela reservoir
is 200 Mt equating to a sediment yield of 1179 t/km2/yr (Attewill et al., 1998).

A4.2.7. Erosion rates in Puerto Rico


The climate is wet and tropical and the topography is rugged and hilly with
elevations up to 1337 m. The rainfall is 1500 mm/yr to 2800 mmlyr and is
variable, being orographic in nature. The total sediment load is 37 600 t1yr of
sediInent based on 650 tlkIn2/yr for a 58 km2 drainage basin. The geology is

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 25C. 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).

A4.2.B. Erosion rates in Switzerland


65 % of the 200 krn 2 catchment above Gebidem reservoir is occupied by
d'Aletsch glacier, the largest in Europe. The glacial activity leads to denudation rates of 25 mmlyr, more than an order of magnitude greater than
unglaciated catchments in the same area. The sediment load is 400 000 m3/yr or
2000 m 3/km2jyr of cohesionless material from fine to gravel. The discharge is
seasonal and dominated by snowmelt, glacial melt and summer stonns (Morris
and Fan, 1997).

A4.2.9. Erosion rates in Turkey


Seventeen dams in Turkey were studied to produce estimates of sediment yield
depending on catchment characteristics like drainage area, soil type, erosion
intensity, slope, annual average precipitation, water discharge, kinetic energy of
rainfall and stream power. To find a relationship the actual rates of sediment yield
'were determined either by using hydrographic resurvey results or using data from
the sediment gauging stations of the Electrical Power Resources Survey and
Development Administration. Sediment yield rates in t/km2/yr were converted to
m 3/km?/yr by dividing by the density of sediment, taken to be 18 tim3 This
figure was factored by 12 to allow for 20% bedload.
The relationships were (Gogus and Yener, 1997);

Qs

= 13959*A

1,213

(r

=0946)

Qs == 0024*A 1.002*Ei}509*p~.994*Q~185 (r == 0-96)

Qs

=0036*A I.039*S~O.I08*Ei).73*SO061*p~.855*Q~122*p~.063 (r =0-961)

where A == catchment area (km2)


St == soil type

E: == erosion index
S == slope of telTain
P r = annual average precipitation (mm)

227

EVACUATION OF SEDIMENTS

Qw = water discharge x 106 (m3/yr)


P s =stream power (kgf/yr)

A4.2. f O. Erosion rates in Spain


Sixty drainage basins were studied, ranging in area from 31 km2 to 16 952 km2
with specific sediInent yields from 8-4 t/km2/yr to 2703 t/km 2/yr, with an average
of 429 tJkm 2/yr. The basins are divided into groups.
Group I
Specific sediment yield less than 150 t/km2fyr. Twenty catchments in zones
where erosion processes are slow, or where carbonate lithologies predominate.

Sf = 617 AO- 67

(r

=077, n =20)

Group 2
150-1000 t/km 2/yr. This group contains over 500/0 of the reservoirs.

SY =202 A I -07

(r

=092, n =33)

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.

Sf =3137 Ao- s7

(r=O91,n=7)

These relationships describe the sediment yield at the reservoirs, which is not all
the material eroded. The sediment delivery ratio varied from 08% 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)

A4.2.11. Erosion rates in Canada


The 4000 km2 Oldman river basin extends eastwards from the Rocky Mountains
in south-west Alberta. The area can be divided into mountains, foothills and high
plains. The suspended-sediment yield averages 70 tlkm2/yr cOlnpared with a
range of 20 to 100 elsewhere in Alberta and up to 350 t/km2fyr in Canada
generally (Neil and Mollard, 1982)

228

APPENDIX 4

A4.3. CLIMATE CLASSIFICATION

A4.3. I. Koppen classification


Of the many climatic classifications produced, one of the most durable is the
Koppen classification which was originally based on eight climatic regions.
Others have since refined it and a recent version published by UNESCO in 1990
is illustrated in Figure A4.7.

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 evapotra...l1spiration.
The classification calculates an index I which expresses the relationship
between surplus moisture, 5, moisture deficiency, d,and potential evapotranspiration, 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).

A4.3.4. Map

of aridity

A map of aridity based on the Thornthwaite classification was compiled by


UNESCO. The aridity indices were determined in this case by examining the
ratio of annual precipitation, P, to annual potential evapotranspiration, ETP. Four
aridity classes were identified:
Hyper-arid
Arid
Semi-arid
Sub-humid

(PIETP
(PIETP
(PIETP
(PIETP

< 003)
< 02)
< 05)
< 075)

In addition to the four aridity classes, temperature was used to further define
the arid regions. The subdivisions were warm winter, mild winter, cool winter and

229

IV

<

()

(:)

--I
0
Z

-n
l/)

0
3:

--I

l/)

A. T,op"=-,
,ainy clim.',"

8. Dry
cli~l"

C. Humid
nwtOlMtm.1
cllmal"

O. Humid
mlcrotMfma'
ctifNl"

~ T,opiul fllinlorn,IAf. Ami


~ T,opic.1 .....nN IAwl
~ Steppe 18&1
mo.-I18WI

D
'0

f11M

Wafm wilh dry win'., lewl

Imon_ upllllld .....Mll


IC.I
IMtdil..,_enl
Humid I....".,.'. ICII

W..m whh dry

ILl""""

Cold ",ilh main w'nter lOt)

lmmCoId with lily win.er 10wi

I.,-t-J~~

~ ImonJOOf\ IYpe)

. f'ollf

clime..,

Figure A4.7.

_1

Koppen climate classification

.J

' ]

APPENDIX4

Table A4.9.
Symbol

Climatic classification by Thornthwaite (UNESCO, 1990)


Moisture province

Thermal province

Annual pe: em

Index: I

Perhumid

Megathermal

>114

>100

B4
Humid

I
I

80-100
998-114

60-80

Bz

856-998

40-60

B,

713-855

20-40

B3

C2

Meso thermal

Moist sub-humid

Microthennal

0-20

42-8-570
I

Dry sub-humid

286-427

(-33)-0

Semi-arid

Tundra

143-285

(-66)-(-33)

Add

Forest

0-142

(-100)-(-66)

60

60

120

120-"

180

5
60~----~----~~~-------4------~~~------r---'------+--------~

60

1. Equatorial
2. Subequatorial
3. Tropical
4. Subtropical

Figure A4.B.

60

120

180

120

5. Temporate

6. Subpolai
7. Polar

Climatic zones a/the earth (after B. P. Alisov)

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.

A4.3.6. Vegetation classification


Vegetation depends on a variety of factors including te~perature, rainfall, soils
and topography. Different regions of the world have di~tinct vegetation types
called biomes. These may be used as a surrogate classification of cliInate and
lnay offer an alternative basis for classification of i global erosion rates.
Combinations of temperature and rainfall have been used ~o identify nine biomes,
as shown in Figure A4.10.
'
Tundra
This area occurs around the North Pole, mainly north of the Arctic Circle. The
ground is pennafrost as there are long cold winters and ~hort 'warm' summers.
There are also 'tundra' like regions, known as Alpine regions, found on the peaks
of the tallest mountains at all latitudes. Trees and tall perennial plants are usually
absent and the ground is covered by mosses, lichens, grasses and perennial
herbs.
Northern coniferous forest (taiga or boreal forest)
This zone is found in North America and Eurasia. It is ch~ractelised by very cold
winters, more precipitation than the tundra and longer, w~nner winters. The soil
thaws and the vegetation grows abundantly. The principal plant life is drought
resistant needle-leaf conifers and some deciduous trees li~e paper birch.
Deciduous forests
Temperate areas with abundant rainfall. The summers ~re relatively long and
warm and the winters are cold. Broad-leafed deciduqus trees dominate the
canopy.

232

APPENDIX4

:!

~
0
~
0

:!
0

..

<0

0
'Of

~
0
0

..-."

Cl

CI

~.

:..

0
~

:....
~

--

Co

c..::
c..::

-e

C)"

)...,>

.::;

'';::

.~
'\::

,..~

0\
~
~

::::

CJ:

t;:

233

EVACUATION OF SEDIMENTS

Table A4.10.

Climatic classification by Budyko (UNESCO, 1990)

Climatic index

Vegetation type

Run-off coefficient

< 1/3

Tundra

07

1/3-1

Forest

03-07

1-2

Steppe

01-03

2-3

Semi-deselt

<01

>3

Deselt

<01

30~----~--------~~~~-----------

25

20

15
Warm temperature

Cold temperature

-5

-10

-15L-----~~-----L------~------~----~------~------~------~----~

50

100

150

200

250

300

350

Total annual rainfall: em

Figure A4.10.

234

Biomes based on temperature and rainfall (after Budyko)

400

450

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

A4.4. TABLES OF COUNTRY DATA

The following tables (Tables A4.11-A4: 16) provide, country by country, data on
climatic conditions and rates of sediment yield" Examples from case studies are
included where available.'-The tables.also.contain data on the number and storage
capacity of dams in each country together. with .estimates of the loss of storage
due to sedimentation.

235

Table A4.11.
Country

Afghanistan

Country data for Asia


Climatic
classification
(Koppen)

BwklBsk

n
Rates of sediment
yield: tlkm 2/yr
(Walling and Webb,
1983)

Example rates:
tlkm 2/yr
(from literature)

No. of
dams
data for

Total
capacity
of dams:

% Li:lP,lL Il Y
lost to
sediments

--I
o

Annual
loss of
storage

M.m}

100-250/250-500

28

8000%

."

00%

Vl

Armenia

DfiDs

CJ

100-250

3:

----

Azerbaijan

DflDs

100-250

Bahrain

Bwh

<50

Z
--I
Vl

-----

Bangladesh

Am/Cwa

1128

500-750/> 1000

Bhutan

Bs

>1000

Brunei

Af

<50

Burma

Am

500-750/750-10001
>1000

Cambodia

Aw

250-500

China

ET/BwkIBs/Cwbl
Cf/CfalCw/Dwl

Dwa/Dwb

50-100/250-5001
500-750/7 50-1 0001
>1000

Df

100-250

Hong Kong

Cwa

250-500

India

Aw I AflAmIAs/Cwa

] 00-2501250-5001
500-750/>] 000

3130/310115

47

784]2

96%

05%

250-500/500-750

6250112 000111 200

1135

17%

04%

616

804012957/1 8001
1620/1400124611571
25 600/2] 7001
16300/2300

28

42804

28}%

]2%

--

TncioI1P"i::J

---,----BwkIBsk

50-100

Iraq

Bwh

50-100125()-500

l~rael

Bwh

50-100

DfC/D IlJ/Cfa

<50/50-tOOI
250-500

Tnm

-50

. - -------- -----

--

--

Japan

729

17322

69%

02%

111%

0-4%

.Jordan

Bwh

50-100

Kazakstnn

DI11/Bsk/Bwk

<50/50-100

Korea (North and


South)

ClhlOwa

500-750

Kuwait

Rwh

250-500

Kyrgy:-:tan

Df

100-250

Laos

Aw/Cf

250-500

Lebanon

Bwh

50-100

Malaysia

Af

250-500

Maldives

Csb

Island

Mongolia

Bwk/Bsk

50-100

Cwa

500-750

Cwa

>1000

Om:m

[3wh

50-100

Pakistan

Bw

50-1 OOt:!50-500

.-

--

~~---~

--------

Myanmar
-------~---

Nepal
----~---

~-~

..

2800
~

~---

-~

-u

iJ

2498/454/1179

14300

Pnpu:1 New
Guinea
.

AI'

250-500/500-750

2581/492/11 126
..

-------~

203%

08%

m
Z

o
X
..t:>.

Table A4.11.

continued

~
()

Climatic
classification
(Koppen)

Rates of sediment
yield: tlkm2/yr
(Walling and Webb,
1983)

Philippines

Am

500-750

Qatar

Bwh

50-100

DfclDtbfDfalDfdF-
DwclDwdlET

<50/50=100

Saudi Arabia

Bwh

50-100

Singapore

Af

250-500/500-750

Sri Lanka

Af

100-250

Syria

Bwh

50-100

Taiwan

Cwa

>1000

Country

Example rates:
tlkm2/yr
(from literature)

No. of
dams
data for

% capacity
lost to
sediments

Total
capacity
of dams:
M.m3

Annual
loss of
storage

~
o
z
o
-n
Vl

Russia-

..

-~

~----

----

~241128/39/141B/9/

CJ
3:

. ...

6/5/5

Z
-I

Vl

.-~

31 700

1525

10
......

Tajikistan

Ds

250-500

Thailand

Aw

<501250-500

Turkey

DslDf

100-250/500-750/
>1000

Turkmenistan

Bwk

50-lO0/lO0-250

United Arab
Emirates

Bwh

50-100

Uzbekistan

BwkfBsk

<50/50-100

Vietnam

Aw

250-500

Yemen

Bwh

50-100

........

07%

.......

76-3874
16

9006

..
-------

--------

--------

1083/203

597%

15%

Table A4.12.

Country data for Africa

Country

Climatic
classification

(Koppen)

Rates of sediment
yield: tllcm 2/yr
(Walling and Webb,
1983)

Example rates:
tlkm1/yr
(from literature)

No. of
darns
data for

Total
capacity
of dams:
M.m)

% capacity
lost to
sediments

Annual
loss of
storage

L7

26328

155%

05%

124%

OO~%

-Algeria

BWh/Csa

<50

Angola

Aw/Cwa

]00-250

Benin

Aw

50-100

Botswana

BSh

100-250

Burkina Faso

Bsh/Aw

50-100

Burundi

AflAw

500-750

Aw/Am

100-250

lslmul

Is1ancl

Aw/Am

<50

BwhlBsh/Aw

<50

Comoros

ls1and

Is1and

Congo

Aw/Af

<50

Congo,
Democratic
RepUblic

Aw

<50

Djibouti

13sh

50-100

Egypt

BWh

<50

--

--

--------Carneroon
Verde

Central Africa
Republic

Chad

--

--

38

'1

1689000

Table A4.12.
Country

continued

Climatic
classification
(Koppen)

Rates of sediment
yield: tlkro2/yr
(Walling and Webb,
1983)

Aw/Am

50-100

ExampJe rates:
tlkm2/yr
(from literature)

No. of
dams
data for

()

TQtal
capacity
of dams:
M.m3

% capacity
lost to
sediments

Annual
loss of
storage

,!

Equatorial Guinea

---.

Eritrea

CwlBsh

50-100

Ethiopia

Bshl
Bwh

5<J-I 00/500-750

Am/Aw

100-250

Gambia

Bsh

50-100

Ghana

Aw

50-100

Guinea

Aw

50-100

Guinea-Bissau

BWh

50-100

Ivory Coast

Aw

50-100

Kenya

Bsh/C

50-100/250-500

Lesotho

Ctb

250-500

Liberia

Am

50-100

Libya

BWh

<50

Madagascar

Af/Aw

250-500/500-750

Malawi

Aw/Cwa

50-100

Mali

Bwh/Bsh/Aw

<50

"Tl
Vl

3:

Z
-I

-----

Vl

---

'-------------

19 520/20117 600/
20-3000/2500-17 600

--------

-----

o
z
o
o

-----

Gabon

?:j

150-0

780%

1-56%

Mauritania

DwhlBsh

Mauritius

Isl.and

<50
Island
~

Morocco

Bsh/Bw

Mozambique

Aw

<501750-1000

100-2501250-5001

10 351-0

17

7-45%

0-42%

4346%

174%

80/17

500-750

Namibia

BwlBsh

50-100

Niger

BWh

<50

Aw

50-100

At'

<50

SnoTome and
Principe

Island

Island

Senegal

Aw/l3sh

50-100

Seychelles

Islands

Nigeria

33

~-~.

Rwanda

-----

Island

--

--.--Sierrn Leone

Am

50-100

Somalia

Bsk

100~-250

South Africa

Csb/Cll)/I3W

<50/50-1001

17/10-1000

100-250/250-5001
5,00-750

----

-----

--

Slldan

BWh/Bsh/Aw

<50

Swaziland

eft)

250-500

55840

TLltlZania

Bs/Aw

250-500/500-750

Aw

50-100

-Togo

94

iJ
-u

rn
Z

o
x

,.tI..

()

o
z
o

."

c.n
m

3:
m
Z

Table A4.12.
Country

Tunisia

-I

continued

c.n

Climatic
classification
(Koppen)

Rates of sediment
yield: tlkm2/yr
(Walling and Webb,
1983)

BWh/Csa

<501750-} 000

Example rates:
tlkm2/yr
(from Ii terature)

No. of
dams
data for

Total
capacity
of dams:
M.m3

% capacity

lost to
sediments

Annual
loss of
storage

15

17760

3421%

0-07%

---

Uganda

Aw

50-100/500-750

Zambia

Cwa

50-100

Zimbabwe

Aw

100-250

-~--

---

.1

'j

Table A4.13.
Country

Country data/or-Australasia
Climatic

classification
(Koppen)

Rates of sediment
yield: t/km2/yr
(Walling and Webb,
1983)

Example rates:
tlkm2/yr
(from literature)
28

Australia

Aw/Bsb/Cfh/CsICsa/
Bs/Bwh

<50/50-100

Fiji

Af

Island

Kiribati

Af

Island

Marshall Islands

Af

Island

Micronesia

Af

Ishmd'

Nauru

AI'

Island

New Zealand

Ctb

250-500/500-7501
>1000

Palau

Af

ISland

Solomon Islands

Af

Island

Tonga

AI'

Island

Tuvnla

Af

Island

Vanuatu

At"

Island

6982119 9701
1'7 340/13 8901
17 070113 3001
12736

No. of
dams
data for

Total
capacity
of dams:
M.m3

% capacity

Annual

lost to
sediments

loss of
storage

202

192%

10%

---Western Samoa

~
w

Af

I----- .

Island

iJ
iJ

ITI

o
x

Table A 4.14.

Country data for Europe

Climatic
classification
(Koppen)

Example rates:
t/km2/yr
(from literature)

Albania

Ctb

500-750

4150/3590

AndolTa

Ctb

250-500

o
z
o
"m

250-500/500-750

3:

Country

No. of
darns
data for

Total
capacity
of dams:
M.m3

()

Rates of sediment
yield: tlkI1l2/yr
(Walling and Webb,
1983)

% capacity
lost to
sediments

Annual
loss of
storage

Vl

Ctb

Austria

930%
-

Dtb

<50

Belgium

Cfb

<50

Bosnia and
Herzegovina

Cfb

500-750

Bulgaria

Dfa/Ctb

100-250/250-500

Croatia

Ctb

100-250

Cyprus

Csa

50-100

Czech Republic

Ctb

<50/100-250

Denmark

Cfb

<50

Estonia

Dtb

<50

Finland

DfclDtb

<50

France

Ctb

<50/50-100/
100-250/250-500

GelTOany

Ctb

<50/100-250

Greece

CsalCfb

500-750

Belarus

279%

-l
Vl

119%

39%

236

30%

02%

-,-~-~

"J

"\
!

'I

~--

--

111

Hungary

Ctb

100-250

Icchmd

efelET

<50

Ireland

cn)

50-100

Italy

Cfa/Csa

50-1001100-250/
250-500

Latvia

Dtb

<50

Liechenstcin

Cfb

100-250

Lithuania

Dfb

<SO

Luxembourg

Cfb

<50

Macedonia

cnl

500-750

Malta

Csa

250-500

Moldova

Of

<50

Monaco

Cfb

250-500

Netherlands

Ctb

<50

Norway

orc

50-100

Poland

Otb

<50

597

95%

04%

Portugal

Ctb

100-250

440

14%

01%

Romania

CfblDtb

< 5011 00-250

18

207

324%

32%

San Marino

Ctb

250-500

--_._--

45701214

83

"'0

.--

Serhia and
Montenegro

------

Ctb

500-750

"'0

rn

><
..r:...

~
()
C

o
z
o

II
(f)

Table A4.14.

continued

Climatic
classification
(Koppen)

Rates of sediment
yield: tlkm2/yr
(Walling and Webb,
1983)

Slovakia

Ctb

100-250

Slovenia

Cfb

500-750

Spain

Cfb

100-250

Sweden

DfclDfb

<50

Switzerland

Cfb

50-100

Ukraine

Dfb

<50

United Kingdom

Cfb

<50/50-100

Vatican City

Cfb

250-500

Country

Example rates:
tlkm2/yr
(from literature)

84-2703

No. of
dams
data for

Total
capacity
of dams:
M.m3

% capacity
lost to
sediments

Annual
loss of
storage

53

28-4%

0-9%

91

23323

3-7%

0-1%
--

71-4%

8-6%

-----

95

-)

0-1%

3:

m
Z
-I

(f)

Table A4. J5.

Country data for North America

Country

Climatic
c1nssi1ication
(Koppen)

Rates of sediment

Antiglla amI
Barhuda

At"

1.00-250

Bahamas

Af

100-250

Barbados

At'

100-250

Belize

Af

lOO-250

Cl.1nadn

ETICfc/C1b/Dfcl
Dlb

< 50/50-] 001


100-250/250-5001
.500-7501750-10001

yield: tlkm2/yr

(Walling amI Webb,


1983)

Example rates:
t/km2/yr
(from literature)

No. of
dams
data for

Total
capacity
of dams:
M.m3

_..__.

91/55/4/0-350

>1000
Costa Rica

Aw

100-250

Cuba

Af

100-250

Af

100-250

Af

100-250

EI Salvador

Aw

100-250

Greenland

Hf

50-100

Af

100-250

Guatemala

Aw

100-2501250-500

Hniti

Af

100-250

Dominica

Dominican
Republic

-------.---

Grenada

--

----

% capacity

lost to
sediments

Annual
loss of
storage

()

Table A4.15.

continued
Climatic
classification
(Koppen)

Rates of sediment
yield: tlkm 2/yr
(Walling and Webb,
1983)

Honduras

Aw

]00--250

Jamaica

Af

100-250

Mexico

Bw/Cw/Af/Aw

50-100/100-2501
250-5001
500-750

Nicaragua

Aw

100-250

Panama

Af

100-250

Saint Kitts and


Nevis

Af

100-250

Saint Lucia

Af

100-250

Saint Vincent and


Grenadines

Af

]00-250

Trinidad and
Tobago

Af

100-250

USA

DfblDfaiB w/Cfal

< 50/50-1 001


100-2501250--5001
500-750/7 50--10001
>1000

Country

Example rates:
tlkm2/yr
(from literature)

No. of
dams
data for

Total
capacity
of dams:
M.m 3

% capacity
lost to
sediments

Annual
loss of
storage

3:

-I

(/)

211

.-.. ---..

(/)

--------_.

CsblBs/Df

o
z
o
"
m

407212374122921
1167/500114511 071
71150112

1105

--~-----

109980

39%

02%

Table A4.16.

Count}")1

data for South America

Climatic
classification

Country

(Koppen)

Rates of sediment
yield: tlkm 2/yr
(Walljng and Webb,
1983)

Example rates:
tlkm 2/yr
(from literature)

130/33

Argentina

Bw/Cfa/Bs

< 501250-500/
750-1000/> 1000

l10livia

AwlBsc

100-250/250-500/
> 1000
<50/100-250

Cfa/Cwa/ Awl AI'


Brazil
- - - - - - - - -1 Ct11lCsIBsIB w
Chile

No. of
dams
data for

Total
capacity
of dams:

% capacity
lost to
sediments

Annual
lo~s of
storage

M.m3

]46/9

3829

0-8%

--0-1%

192%

--96%

100-2501>1000

----~--~----.

Colombia
------~---

50-100/> 1000

Af/Aw
..--.-

917

-,

Ecuador

Aw/Af

100-250/> 1000

French Guiana

A1'

<50

Guyana

AI'

<50

Paraguay
---

Cw..uAw

<50

Peru

Af/AwlBsc

100-250/250-5001
>1000

Af

<50

2000

----~---

Suriname

---UllIgtluy
Venezuela

~
~

Cra

Aw/Af

<50
<50/100-250

212

-0
-0

o
x

Index

Page numbers in italics refer to illustrations.

Africa
country data 239-242
erosion rates 221-224
hydropower 27
irrigation 27
population 26-27
Alisov classification 229, 231
Americas
country data 247-249
hydropower 24-26
irrigation 24-26
population 24-26
annual precipitations 106, 213-214,213
Aracay River, Venezuela 202
areas
erosion rate 95-100
suited to flushing 93-124
aridity maps 229,232
Asia
country data 236-238
hydropower 27-28
irrigation 27-28
population 27
Atbara River, Sudan 195
Athi-Kapiti plains 223
Australasia 243
Austria 68,72-74, 76
case study 176-178
downstream impacts 178
flushing 177-178
sedimentation 177
autumn 1998 precipitation 108,111

Baira, India 68, 72-75, 81


case study 208-210
downstream impacts 209-210
flushing 209-210
sedimentation 209

bank failure 193


basin areas, erosion rate 121
basin outlet distances 122
basin shapes 10, 84,88-89
bed width, incised channel 44, 44
bibliography 143-148
biomes 232,234-235,234
boreal
climates 115-116
forests 232
broad reservoirs 84
Budyko classification 232-235, 233-234

Cacm, Costa Rica 68, 72-75, 79


case study 197-200
downstream impacts 200
flushing 198-200
sedimentation 198
Canada, erosion rates 228
capacity criterion, geometry 49,49
case studies
B aira, India 208-210
Cacm, Costa Rica 197-200
erosion rates 221-228
flushing 71-81, 171-210
Gebidem, Switzerland 200-202
Grofind, Austria 176-178
Guanting, China 179-182
Guernsey, USA 172-173
Heisonglin, China 184-186
Hengshan, China 196-197
Ichari, India 207-208
Jensanpei, Taiwan 174-175
Khashm El Girba, Sudan 195-196
Mangahao, New Zealand 171-172
N anqin, China 205-207
Naodehai, China 175-176
numerical models 163-170
Ouchi-Kurgan, former USSR 190-191
Palagnedra, Switzerland 178-179
Sanmenxia, China 186

251

EVACUATION OF SEDIMENTS

Santo Domingo, Venezuela 202-205


Sefid-Rud, Iran 191-195
Shuicaozi, China 182-184
Tarbela Dam, Pakistan 163-170
Warsak. Pakistan 189-190
Zemo-Afchar, former USSR 173-174
catchments
characteristics 123,217-221
delivery ratio 122
hydrology 81-82, 123
orientation 217-218
sedimentology 9, 82
size 122
slope 217
Central AmeIica
hydropower 25-26
imgation 25-26
population 25
channel erosion 193
China 68, 72-80,88
case study 175-176,179-186,196-197,

205-207
density current venting 206
downstream impacts 176,182-186, 189,

coniferous forests 232


constraints, flushing 60
constrictions, flushing flow 51-52
constIuction reservoirs 6-7, 28-30
continental variations
erosion rate 94-96
sediment yield 94-96
cool Mediterranean climates 115
Costa Rica 68,72-75, 79
case study 197-200
downstream impacts 200
flushing 198-200
Reventazon River 197
sedimentation. 198
country classification, climate 118-120
country data tables
Africa 239-242
Asia 236-238
Australasia 243
Europe 244-246
North America 247-248
South America 249
cross-sections, flushing channel 48, 48

197,207
erosion rates 224-225
flushing 175-176, 181-189, 196-197,

206
lateral erosion 185
Liuhe River 175
sedimentation 175, 180-186, 196,
205-206
storage capacity 188
Yellow River 186, 225
Yeyu River 184
Yili River 182
Yongding River 179-180
climate 211-215
classification 229-235
combining homogenous groups 117
erosion rate 93, 115-120
Tarbela Dam, Pakistan 164-165
climatic zones 12
country classification 118-120
earth 231
erosion rate 115-120
river basins 118-120, 118
sediment yield 118, 118
world 101-120
cold steppe climates 115
concentration inflow 56

252

DDR see-drawdown ratio


deciduous forests 232
delivery ratios
catchment size 122
depositional features 122
demand comparisons 30
demand distribution 4-7
storage 23-28
density cun-ent venting 66-69
Nanqin, China 206
deposition
features 122
potential 9, 83-84
sediment size 53-57
Tarbela Dam, Pakistan 165-166
design considerations 127-130
detention flushing 185
discharge flushing 45
distIibution
construction reservoirs 28-30
sediment rate 33
storage demand 4-7
storage loss 7,31-34
distIibution of demand 23-28
diversion channels 194
downstream impacts 10, 86-88

INDEX

Baira, India 209-210


Cachi, Costa Rica 200
Gebidem, Switzerland 201-202
Gmund, Austria 178
Guanting, China 182
Guernsey, USA 173
Heisonglin, China 185-186
Hengshan, China 197
Ichari, India 208
Jensanpei, Taiwan 175
Khashm El Girba, Sudan 196
Mangahao, New Zealand 172
Nanqin, China 207
Naodehai, China 176
Ouchi-Kurgan. fonner USSR 191
Palagnedra, Switzerland 179
Sanmenxia, China 189
Santo Domingo, Venezuela 205
Sefid-Rud, Iran 195
Shuicaozi, China 184
Warsak, Pa..\istan 190
Zemo-Afchar, former USSR 174
drainage basin areas 218-219,218
erosion rate 121
sediment yield variations 95
drainage density 219
drawdown 10, 85-86, 88
eachf, Costa Rica 198
flushing 40-41,41
Heisonglin, China 185
incomplete 51
N anqin, China 206
drawdown ratio (DDR) 51
dryness, radiational index 232-235
Durlassboden reservoir 176-178

earth. climatic zones 231


econoIilic assessment, flushing 60
economic factors analysis 14, 129
efficient flushing 43, 58-61
hydraulic conditions 7-8, 58-59
empty flushing 66
enhancements, flushing 10, 70, 86
environments, Tarbela, Pakistan 164-169
equiiibrium conditions 42
erosion
affecting factors 211-221
controls 220-221
definition 211
human impact 220-221

land use 219-220


processes 122
vegetation 219
erosion rates 11
Africa 221-224
Canada 228
case study 221-228
China 224-225
continental variation 94-96
drainage basin area effects 121
geology effect 121
geotechnics 93
Himalayas 226
human impact 93, 121

India 225-226
Kenyan grazing lands 223-224
Krishna River, India 226
land use effect 93, 121
Pakistan 226
precipitation effect 121
Puerto Rico 226-227
slope effect 121-122
soil effect 121-122
Spain 228
Switzerland 227
Taiwan 96
Thailand 224
topography 93
Turkey 227-228
vegetation effect 93, 121-122
worldwide 93
Europe
_
country data 244-246
hydropower 24
irrigation 24
population 23-24
evaluation
flushing criteria 50

sediment br..lance ratio

45-46

financial analysis 129


flood control 3-4
flushing
Baira, India 209-210
Cacm, Costa Rica 198-200
case study 71-81,171-210
constraints 60
criteria 42-58, 87-88
discharge 45
economic assessment 60

253

EVACUATION OF SEDIMENTS

efficiency 43, 51-52


enhancements 10, 70, 86
equilibrium conditions 42
flow constrictions 51-52
Gebidenl, Switzerland 200-201
Gmiind, Austria 177-178
Guanting, China 181-182
Guernsey, USA 172-173
Heisonglin, China 184-185
Hengshan, China 196-197
!chari, India 207-208
influences 39-62
Jensanpei, Taiwan 175
Khashm E1 Girba, Sudan 195-196
long-term equilibrium conditions 42
Mangahao, New Zealand 171-172
mechanisms 40-42
Nanqin, China 206
N aodehai, China 175-176
numerical models 61-62
operation duration 45
optimum locations 12-13
Ouchi-Kurgan, former USSR 191
outlets 127
Palagnedra, Switzerland 179
periods 57-58
Sanmenxia, China 187-189
Santo Domingo, Venezuela 202-205
Sefid-Rud, Iran 192-195
Shuicaozi, China 182-183
site-specific factors 60
suitable geographical areas 93-124
Tarbela Dam, Pakistan 166-167
techniques 66-70
value 10
\yarsak, Pakistan 190
water available 8, 59
worldwide experience 65-89
Zemo-Afchar, former USSR 173-174
flushing channels
cross-sections 48, 48
narrow reservoirs 53
wide reservoirs 53
flushing width ratio (FWR) 53
forests 232, 235
fOlmer USSR 68, 72-74, 76, 78
case study 173-174, 190-191
downstream impacts 174, 191
flushing 173-174,191
sedimentation 173,190
free flow

254

Cachf, Costa Rica 198


Santo Domingo, Venezuela
full drawdown, flushing 41
FWR see flushing width ratio

204

Gebidem, Switzerland 68, 72~75, 79


case study 200-202
downstream iInpacts 201-202
flushing 200-201
sedimentation 200
geographical areas, flushing 93-124
geographical distribution 19
geology 216
erosion rate effect 121
geometry, capacity criterion 49,49
geotechnics, erosion effect 93
global sediment yields 93-100
variation maps 100, 102-105
global water resources 18-19
Gmiind, Austria 68, 72-74, 76
case study 176-178
downstream impacts 178
flushing 177-178
sedimentation 177
gorge-like reservoirs 84
graded sediments, effect 54-55
grasslands 235
growth, world population 20, 23
Guanting,China 68,72-74,77,88
case study 179-182
downstream impacts 182
flushing 181-182
sedimentation 180-181
Guernsey, USA 68, 71-74, 88
case study 172-173
downstream impacts 173
flushing 172-173
sedimentation 172

Heisonglin, China 68, 72-74, 77-78


case study 184-186
downstream impacts 185-186
flushing 184-185
lateral erosion 185
sedimentation 184
Hengshan, China 68, 72-73, 75, 79
case study 196-197
downstreanl impacts 197
flushing 196-197

INDEX

sedimentation 196
high annual precipitation 10 1
high erosion rate areas 95-99
Himalayas, erosion rates 226
historic growth .
hydropower 21-22,23
reservoir 30
homogenous climatic groups I 17
human impacts
erosion control 220-221
erosion rate 93, 121
humid climates 115
hydraulic conditions, flushing 7-8, 58-59
hydraulic modelling 13-14, 128-129
hydrology
catchment 81-82, 88, 123
characteristics 12, 123
investigations 13, 127-128
sedimentology 9
Tarbela Dam, Pakistan 164-165
hydropower 3
Africa 27
Americas 24-26
Asia and Oceania 27:-:-28
Europe 24
historic growth 21-22, 23
potential 22-23

Iehari, India

68, 72-75, 80, 88

case study 207-208


downstream impacts 208
flushing 207-208
sedimentation 207
ICOLD
World Register of Dams 17-19
world storage volume data 152-155
incised channels 44, 44
incomplete drawdown 51
India 68,72-75,80-81,88
case study 207-210
downstream impacts 208-210
erosion rates 225-226
flushing 207-210
Krishna River 226
Ravi River 208
sedimentation 207, 209
Indus basin Pakistan 164
Indus River flows 165
infiltration rate, soils 217
inflows 127-128

concentration 56
hydro graph 165
Tarbela Dam, Pakistan 165
insufficient drawdown 41
insufficient flushing flows 51-52
investigations
sediment 128
site-specific 13
Iran 68, 72-73, 75, 78-79
bank failure 193
case study 191-195
channel erosion 193
diversion channels 194
downstream impacts 195
flushing 192-195
lateral erosion 194
long-term predictions 195
sedimentation 191-192
sheet erosion 193
irrigation 3, 21
Africa 27
Americas 24-26
Asia and Oceania 27
Europe 24

Jensanpei, Taiwan. 68, 72-76


case study 174-175
downstream impacts 175
fl ushlng 175
sedimentation 174

Kabul River, Pakistan

189

Kenyan grazing lands 223-224 .


Khashm E1 Girba, Sudan 68, 72-75, 79
case study 195-196
downstream impacts 196
flushing 195-196
sedimentation 195
Koppen classification 12, 229-230, 230
climate 108, 112-115,116
world climate 116-118,116
Krishna River, India 226

land use
erosion control 219-220
erosion rate effect 93, 121
lateral erosion 185, 194
Liuhe River, China 175

255

EVACUATION OF SEDIMENTS

Loiza, Puerto Rico 227


long-term capacity ratio (LTCR) 47-50
long-term equilibrium conditions 42
long-telm storage capacity 188
longitudinal energy gradient 44
loss
rate trends 34
soil 212
storage 7, 31-32, 35-36, 35
low annual precipitation 101
low erosion rate areas 99-100
low-level outlets 10, 85
lower Yellow River, China 225
LTCR see long-term capacity ratio

major rivers, sediment yield 97-98


Mangahao, New Zealand 68,71-74
case study 171-172
downstream impacts 172
flushing 171-172
sedimentation 171
maps
alidi ty 229, 232
global sediment yield 100,102-105
Massa River, Switzerland 200-201
mean annual precipitations 213
mean annual run-off 213,215
mean suspended-sediment yields 214-215,
223
mechanisms, flushing 40-42
Mediten-anean climates 114-115
Melezza River, Switzerland 178
mid-latitude
summer dry 114
wet 113
winter dry 113-114
mobility, sediments 8-9,59-60
modelling
hydraulic 13-14
system simulation 14
Tarbela Dam 166-169,167
Mount Kenya 221-222
Mount Kilimanjaro 223
Mpwanga, Tanzania 220
Mwea-Masinga plains 221

Nanqin, China 68,72-75, 80


case study 205-207
density CUlTent venting 206

256

downstream impacts 207


flushing 206
sedimentation 205-206
Naodehai, China 68, 72-74, 76
case study 175-176
downstream impacts 176
flushing 175-176
sedimentation 175
nalTOW reservoirs 84
flushing channels 53
New Zealand 68, 71-74
casestudy 171-172
downstream impacts 172
flushing 171-172
Mangahao River 171
sedimentation 171
North America
country data 247-248
hydropower 25
irrigation 25
population 24-25
Northern coniferous forests 232
numerical models
case study 163-170
flushing 61-62
sediment 166
Tarbela Dam, Pakistan 166-169,167

Oceania
hydropower 27-28
irrigation 27-28
population 27
operation duration, flushing 45
operation policies, Pakistan 166
operational considerations 85-86, 88
operational limitations 10
optimum locations, flushing 12-13
Orange River, South Africa 224
organic content, soil 217
orientation, catchments 217-218
Ouchi-Kurgan, fOlmer USSR 68,72-74, 78
case study 190-191
downstream impacts 191
flushing 191
sedimentation 190

50 percentile size
river-bed matelial 55
transported sediment 55

INDEX

Pacific Asiatic-Australian sector, erosion 96


Pakistan 72-74, 78, 88
see also Tarbela Dam
case study 163-170,189-190
climate 164-165
downstream impacts 190
environment 164-169
erosion rates 226
flushing 166-167, 190
history 163
hydrology 164-165
Indus basin 164
Indus River 163-165
Indus River flows 165
Kabul River 189
numerical sediment modelling 166, 167
operation policies 166
sediment 163-164
sediment deposition 165-166
sediment inflows 165
sediment throughput 167-168
sedimentation 189-190
storage prediction 168-169,168
Palagnedra, Switzerland 68, 72-74, 76-77
case study 178-179
downstream impacts 179
flushing 179
sedimentation 178-179
parameters, sediment ratio 55
permeability, soils 217
polar
desert 114-115
wet and dry 114
population
Africa 26-27
Americas 24-26
Asia and Oceania 27
Europe 23-24
world 19-21
precambrian basement schists 223
precipitation 211-215
erosion rate effect 121
seasonal variation 101-108
precipitation distribution
autumn 1998 108, 111
spring 1998 108, 109
summer 1998 108, 110
winter 1998 107
pressurised flushing 66, 204
Puerto Rico 227
erosion rates 226-227

radiational index, dryness 232-235,233


rain forests '235
rainfall intensity 212
rapid drawdown 198
rates
construction reservoirs 28-30
loss of storage 7. 31-32, 35-36, 35
Ravi River, India 208
references 133-140
regional sedimentation rates 31
register of dams 17-18
Reventazon River, Costa Rica 197
Rhone flow 201
ringlet sedimentation 34
river-bed material 55
rivers
Aracay, Venezuela 202
Atbara, Sudan 195
bank 193
basin 118-120,118
climatic zones 118-120, 118
Indus, Pakistan 163-165
Kabul, Pakistan 189
Krishna, India 226
Liuhe, China 175
Mangahoa, New Zealand 171
Massa, Switzerland 200-201
Melezza, Switzerland 178
Orange, South Africa 224
Ravi, India 208
Reventazon, Costa Rica 197
Santo Domingo, Venezuela 202
sediment yield 97-98
Tons, India 207
Yellow, China 186, 225
Yeyu, China 184
Yili, China 182
Yongding, China' 179-180
rock types 216
routing techillques 66-70
run-off 214-215,215,223

Sanmenxia, China 68, 72-74, 78, 225


case study 186
downstre~m impacts
189
flushing 187-189
sedimentation 186
storage capacity 188
Santo Domingo, Venezuela 68, 72-75,
79-80

257

EVACUATION OF SEDII'1ENTS

case study 202-205


downstream impacts 205
flushing 202-205
sedimentation 203
SBR see sediment balance ratio
SDR see sediment delivery ratio
seasonal variations, precipitation 101-108
sediment balance ratio (SBR) 42-43
evaluation 45-46
parameters 55
sediment delivery ratio (SDR) 122
sediment loads 224
sediment size 53-57
sediment size ratio (SSR) 54,55
sedimentation, rate data 156-162
Sefid-Rud,Iran 68,72-73,75,78-79
bank failure 193
case study 191-195
channel erosion 193
diversion channels 194
downstream impacts 195
flushing 192-195
lateral erosion 194
long-term predictions 195
sedimentation 191-192
sheet erosion 193
shapes, basin 10
shear strength soils 217
sheet erosion 193
Shuicaozi, China 68,72-74, 77
case study 182-184
downstream impacts 184
flushing 182-183
sedimentation 182
site-specific factors 9, 60
site-specific investigations 13, 127-130
sizes, reservoir 35-36, 35
slopes
catchments 217
erosion rate effect 121-122
slow draw down 198
sluicing 66-69
soils 216-217
erosion rate effect 121-122,220
infiltration rate/permeability 217
loss 212
organic content 217
shear strength 217
structure 217
texture 216
South Africa, Orange River 224

258

South America
country data 249
hydropower 25-26
irrigation 25-26
population 25
Spain, erosion rates 228
spring 1998, precipitation 108, 109
SSR see sediment size ratio
steppe climates 115
storage 3-4
capacity 9, 83, 188
construction vs demand 30
demand diSllibution 4-7, 23-28
distribution 31-34
gross requirements 35-36
increase distribution 30
loss rate 7,31-36
lost to sedimentation 33
requirements 7, 35-36
trends 34
volume disllibution 19
volume predictions 168-169,168
world demand 19-23
world total 17-18
worldwide distribution 18-19
structures, soil 217
Sudan 68, 72-75, 79
Atbara River 195
case study 195-196
downstream impacts 196
flushing 195-196
sedimentation 195
summer 1998, precipitation 108,110
suspended-sediment yields 214-215,223
world maximum 96
sustainable reservoir capacity 47-50
Switzerland 68, 72-77, 79
case study 178-179,200-202
downstream impacts 179, 201-202
erosion rates 227
flushing 179, 200-201
Masse River 200-201
Melezza River 178
sedimentation 178-179,200
system simulation modelling 14, 129

tables, country data 235-249


taiga forests 232
Taiwan 68, 72-76
case study 174-175

INDEX

downstream impacts 175


erosion rate 96
flushing 175
sedimentation 174
Tarbela Dam, Pakistan
case study 163-170
climate 164-165
environment 164-169
flushing 166-167
history 163
hydrology 164-165, 165
Indus basin 164
Indus River flows 165
modelling 166-169, 167
numerical sediment modellimz 166
operation policies 166
...,
sediment 163-164
sediment deposition 165-166
sediment inflows 165
sediment modelling 166
sediment throughput 167
storage prediction 168-169
tectonic activity 216
temperatures 215
textures, soil 216
Thailand, erosion rates 224
Thomwaite classification 229,231
throughputs sediment, Pakistan 167-168
Tons River, India 207
top width ratio (TWR) 53
topography, erosion rate 93
transportation 11
transported sediment 55
transporting capacity 42-44
empirical equation 43
trapping efficiency 45
trends, storage loss rate 34
tropical
desert 112-113
forests 235
wet 108,112
wet and dry 112
tundra 232
Turkey, erosion rates 227-228
TWR see top width ratio
Upper Tana basin, eastern Kenya
upstream depositation 71
USA 68, 71-74, 88
case study 172-173
downstream impacts 173

flushing 172-173
sedimentation 172
USSR see former USSR

value, flushing 10
vegetation
climate classification 232, 235
erosion control 219
erosion rate effect 93, 121-122
Venezuela 68,72-75, 79-80
Aracay River 202
case study 202-205
downstream impacts 205
flushing 202-205
sedimentation 203
volcanic activity 216

warm humid climates 115


vVarsak, Pakistan 72-74, 78, 88
case study 189-190
downstream impacts 190
flushing 190
sedimentation 189-190
water levels
flushing 8, 59
incomplete drawdown 51
wide reservoirs 53
wind direction 215
wind speeds 215
winter 1998, precipitation 107
world
climatic zones 101-120
ICOLD data 152-155
Koppen classification 116-118,116
maximum suspended ...gediment yield 96
population 19-2 t
storage 4
storage volume 152-155
worldwide
erosion rates 93
reservoir construction 28-30
sediment flushing experience 65-89
storage distribution 18-19

221-223
Yellow River, China 186, 225
Yeyu River, China 184
yields, sediment 93-105, 213
YiIi River, China 182

259

EVACUATION OF SEDIMENTS

Yongding River, China

179-180

Zemo-Afchar, former USSR


case study 173-174

260

68, 72-74, 76

downstream impacts 174


flushing 173-174
sedimentation 173
Zemo-Afchar reservoir, former USSR

72-74, 76

68,