MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT WORLD METEOROLOGICAL ORGANIZATION

WORLD METEOROLOGICAL ORGANIZATION
OPERATIONAL HYDROLOGY REPORT No. 47
MANUAL ON SEDIMENT MANAGEMENT
AND
MEASUREMENT
By Yang Xiaoqing
WMO-No. 948
Secretariat of the World Meteorological Organization – Geneva – Switzerland
THE WORLD METEOROLOGICAL ORGANIZATION
The World Meteorological Organization (WMO), of which 187* States and Territories are Members, is a specialized agency
of the United Nations. The purposes of the Organization are:
(a)
To facilitate worldwide cooperation in the establishment of networks of stations for the making of meteorological
observations as well as hydrological and other geophysical observations related to meteorology, and to promote the
establishment and maintenance of centres charged with the provision of meteorological and related services;
(b)
To promote the establishment and maintenance of systems for the rapid exchange of meteorological and related information;
(c)
To promote standardization of meteorological and related observations and to ensure the uniform publication of
observations and statistics;
(d)
To further the application of meteorology to aviation, shipping, water problems, agriculture and other human activities;
(e)
To promote activities in operational hydrology and to further close cooperation between Meteorological and
Hydrological Services; and
(f)
To encourage research and training in meteorology and, as appropriate, in related fields and to assist in coordinating
the international aspects of such research and training.
(Convention of the World Meteorological Organization, Article 2)
The Organization consists of the following:
The World Meteorological Congress, the supreme body of the Organization, brings together the delegates of Members
once every four years to determine general policies for the fulfilment of the purposes of the Organization, to approve longterm plans, to authorize maximum expenditures for the following financial period, to adopt Technical Regulations relating
to international meteorological and operational hydrological practice, to elect the President and Vice-Presidents of the
Organization and members of the Executive Council and to appoint the Secretary-General;
The Executive Council, composed of 36 directors of national Meteorological or Hydrometeorological Services, meets at
least once a year to review the activities of the Organization and to implement the programmes approved by Congress;
The six regional associations (Africa, Asia, South America, North and Central America, South-West Pacific and Europe),
composed of Members, coordinate meteorological and related activities within their respective Regions;
The eight technical commissions, composed of experts designated by Members, study matters within their specific
areas of competence (technical commissions have been established for basic systems, instruments and methods of observation, atmospheric sciences, aeronautical meteorology, agricultural meteorology, marine meteorology, hydrology,
and climatology);
The Secretariat, headed by the Secretary-General, serves as the administrative, documentation and information centre of
the Organization. It prepares, edits, produces and distributes the publications of the Organization, carries out the duties
specified in the Convention and other Basic Documents and provides secretariat support to the work of the constituent
bodies of WMO described above.
________
* On 30 November 2003.
WORLD METEOROLOGICAL ORGANIZATION
OPERATIONAL HYDROLOGY REPORT No. 47
MANUAL ON SEDIMENT MANAGEMENT
AND
MEASUREMENT
By Yang Xiaoqing
WMO-No. 948
Secretariat of the World Meteorological Organization – Geneva – Switzerland
2003
Copyright in this electronic file and its contents is vested in WMO. It must not be altered,
copied or passed on to a third party or posted electronically without WMO's written
permission.
© 2003, World Meteorological Organization
ISBN: 92-63-10948-6
NOTE
The designations employed and the presentation of material in this publication do not imply the
expression of any opinion whatsoever on the part of the Secretariat of the World Meteorological
Organization concerning the legal status of any country, territory, city or area, or of its authorities,
or concerning the delimitation of its frontiers or boundaries.
CONTENTS
Page
Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
vii
Summary (English, French, Russian and Spanish) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
ix
CHAPTER 1 — ECOLOGY AND ENVIRONMENT RELATED TO SEDIMENTATION . . . . . . . . . . . . . . . . . .
1.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.2
Impacts of soil erosion on ecology and environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.2.1
Desertification and degradation of agricultural production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.2.2
Sediment-related disasters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3
Impacts of river sedimentation on ecology and environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3.1
River sediment and flood disasters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3.1.1 Conveyance capacity of rivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3.1.2 Fluvial process and instability of river channel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3.1.3 Safety of training works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3.1.4 Sediment deposits by floods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3.1.5 Variation of groundwater level and salinity by river sedimentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3.2
Environment of sediment-laden rivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3.2.1 Deposition in irrigation systems and desertification at irrigation system heads . . . . . . . . . . . . . . . . . . . . . . . .
1.3.2.2 Impacts of river channel shifting on environment and ecology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.4
Reservoir sedimentation and environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.4.1
Loss of reservoir storage capacity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.4.2
Water pollution by reservoir sedimentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.4.3
Rise of groundwater level and salinity by deposit extension in reservoir backwater regions . . . . . . . . . . . . . .
1.4.4
Problems of downstream reservoir . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.4.4.1 Flood plain collapse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.4.4.2 Downstream navigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.4.5
Case studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.4.6
Guanting Reservoir in China . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.4.7
Aswan High Dam . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.5
Utilization of sediment resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1
1
1
2
2
2
2
3
3
3
4
4
4
4
4
4
4
6
6
6
6
6
7
7
7
8
9
CHAPTER 2 — SOIL EROSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2
Natural erosion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2.1
Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2.2
Wind . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2.3
Freeze-thaw . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2.4
Living organisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.3
Accelerated erosion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.4
Factors affecting soil erosion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.4.1
Meteorology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.4.2
Geology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.4.3
Topography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.4.4
Soil characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.4.5
Vegetation cover . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.4.6
Human activities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.5
Degree and intensity of soil erosion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.5.1
Soil loss tolerance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.5.2
Soil erosion intensity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.6
Sediment yield in a basin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.7
Monitoring of soil erosion and sediment yield in a basin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.7.1
Runoff plots and experiments in the laboratory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10
10
10
10
12
12
12
12
12
12
13
13
14
15
15
15
15
15
15
16
16
iv
CONTENTS
Page
2.7.2
Measurements of soil and water losses on pilot watersheds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.7.3
Measurement with Cs-137 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.7.4
Dynamic monitoring by remote sensing and GIS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.8
Prediction of soil erosion and sediment yield . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.8.1
Prediction of soil erosion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.8.2
Prediction of sediment yield . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.8.3
USLE and RUSLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.8.4
Empirical regression statistical model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.8.5
Deterministic sediment yield models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.9
Soil erosion control and watershed management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.9.1
Soil and water conservation planning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.9.2
Measures for soil and water conservation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.10
Summary on global soil erosion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
16
16
17
17
17
17
18
19
21
23
23
24
26
27
CHAPTER 3 — SEDIMENT TRANSPORT IN RIVERS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.1
Patterns of sediment transport in rivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.1.1
Bed material load and wash load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.1.2
Bed load, saltation and suspended load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.1.3
Continuity of sediment movement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.1.4
Relative importance of bed load and suspended load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2
Bed load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.1
Incipient motion of sediment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.1.1 Stochastic property of incipient motion of sediment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.1.2 Condition of incipient motion for non-cohesive uniform sediment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.1.3 Condition for incipient motion of cohesive sediment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.2
Bed form and resistance in fluvial streams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.2.1 Development of bed forms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.2.2 Flow resistance in alluvial streams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.3
Bed load transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.3.1 Transport of uniform bed load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.3.2 Transport of non-uniform bed load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.3.3 Characteristics of transport of gravel bed load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3
Suspended sediment transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3.1
Mechanism of sediment moving in suspension . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3.2
Diffusion equation and vertical distribution of suspended sediment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3.3
Transport rate of suspended load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3.4
Non-equilibrium transport of suspended sediment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4
Total sediment load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4.1
Einstein’s bed load function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4.2
Colby’s method (1964) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4.3
Bagnold’s work (1966) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4.4
The Engelund-Hansen formula (1972) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4.5
The Ackers-White formula (1973) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4.6
Yang’s approach (1996) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4.7
Formula of the Wuhan University of Hydraulic and Electric Engineering (WUHEE) . . . . . . . . . . . . . . . . . . .
3.4.8
Estimation of total sediment load including wash load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4.8.1 Annual sediment load evaluated by the relationship between flow discharge and sediment transport rate . . . .
3.4.8.2 Estimation of sediment load based on factors in river basins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4.8.3 Estimation of sediment yield of a watershed from reservoir deposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.5
Hyperconcentrated flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
29
29
29
29
29
29
30
30
30
30
31
32
32
33
35
35
37
38
40
40
40
42
43
44
44
45
45
45
46
46
47
47
47
48
48
49
51
CHAPTER 4 — FLUVIAL PROCESSES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2
Categories of rivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2.1
Mountainous and upland rivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2.2
Plain and piedmont rivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
54
54
54
54
54
CONTENTS
v
Page
4.3
4.3.1
4.3.2
4.3.3
4.3.4
4.3.5
4.3.6
4.3.6.1
4.3.6.2
4.3.6.3
4.3.6.4
4.3.6.5
4.3.7
4.3.7.1
4.3.7.2
4.4
4.4.1
4.4.1.1
4.4.1.2
4.4.1.3
4.4.1.4
4.4.2
4.4.2.1
4.4.2.2
4.4.3
4.4.3.1
4.4.3.2
4.4.3.3
4.4.3.4
4.4.3.5
4.4.3.6
4.5
4.5.1
4.5.2
4.5.3
4.5.4
4.5.5
4.5.6
4.5.7
4.5.8
4.5.9
4.5.10
4.5.10.1
4.5.10.2
4.5.10.3
4.5.10.4
4.5.11
4.5.11.1
4.5.11.2
4.5.12
4.5.12.1
4.5.12.2
4.5.12.3
4.6
4.6.1
4.6.1.1
4.6.1.2
4.6.2
Classification of river patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
River patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Methods for classification of river patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Characteristics of rivers with different patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Causes for formation of river patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Transformation of river patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Critical relationships between different river patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Relationships between longitudinal slope and river patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Relationships between longitudinal slope and mean discharge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Relationships between longitudinal slope and maximum discharge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Wandering index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Relationships between longitudinal slope, bed sediment and discharge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Indexes of river stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Longitudinal stability of river channels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Transversal stability of river channels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Morphology of rivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Dominant discharge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Determination of dominant discharge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Bankfull discharge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Empirical expression for bankfull discharge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Bankfull discharge estimated by recurrence intervals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Longitudinal profiles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Geometric expressions of longitudinal profiles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Empirical relationships between longitudinal slope and watershed factors . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Cross-sectional morphology of rivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Hydraulic geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Hydraulic geometry along rivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Relationships between factors of watershed and hydraulic geometry along rivers . . . . . . . . . . . . . . . . . . . . . .
Analytic solution of hydraulic geometry along rivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Hydraulic geometry of gravel rivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Hydraulic geometry for canals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Fluvial processes of meandering rivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Plane morphology of meandering rivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Relationships between meander wavelength and discharge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Relationships between central angle and curvature radius . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Relationships between meander elements and width of straight (crossing) reaches . . . . . . . . . . . . . . . . . . . . .
Relationships between configurations and cross-sectional geometry of meanders . . . . . . . . . . . . . . . . . . . . . .
Crossings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Dynamic line of flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Transversal slope of water surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Longitudinal slope of water surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Transversal circulating flows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Distribution for transversal velocity (radial) of circulating flows (Rozovski, 1957, 1965) . . . . . . . . . . . . . . . .
Relative intensity of circulating flows (Xie, 1987) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Vortex intensity of circulating flow (Xie, 1987) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Transversal slope of bed surface and distribution of sediment particles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Sediment transport in meandering rivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Transport of suspended load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Bed load transportation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Characteristics of fluvial processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Collapse of concave banks and growth of convex banks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Migration of meanderings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Cutoffs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Fluvial processes of wandering rivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Flow and sediment transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Characteristics of river flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Characteristics of sediment transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Morphological features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
55
55
57
57
57
57
57
59
59
59
59
59
59
59
60
60
60
61
61
62
62
63
63
63
64
65
65
66
67
68
68
69
69
69
69
70
70
70
70
71
71
71
71
72
72
72
73
73
73
74
74
74
74
74
75
75
75
76
vi
CONTENTS
Page
4.6.2.1 Static features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.6.2.2 Dynamic features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.6.2.3 Node points . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.6.3
Channel degradation and aggradation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.6.3.1 Characteristics of degradation and aggradation for wandering rivers with high sediment concentration . . . . .
4.6.3.2 Degradation and aggradation for wandering rivers with relative low sediment concentration . . . . . . . . . . . . .
4.6.4
Degradation and aggradation in hyperconcentrated floods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.6.4.1 Features of hyperconcentrated floods in the Lower Yellow River . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.6.4.2 Flow patterns and transport modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.6.4.3 Features of degradation and aggradation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.6.5
Shrinking of river channel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.7
Fluvial processes of anabranched rivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.7.1
Morphological characteristics of anabranched rivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.7.1.1 Classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.7.1.2 Morphological indices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.7.2
Morphology of cross-sections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.7.3
Ratio of discharge and sediment diversion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.7.4
Fluvial processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.7.4.1 Main features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.7.4.2 Channel deformation for different anabranched rivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.8
Fluvial processes of straight rivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.8.1
Morphological features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.8.2
Features of flow and sediment transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.8.3
Features of fluvial processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.9
Stabilization and rectification of river channels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.9.1
Parameters of river training planning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.9.1.1 Determination of design discharge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.9.1.2 Determination of channel width . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.9.1.3 Alignment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.9.2
Structures of river training works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.9.2.1 Structures of training works for moderate and low flow channels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.9.2.2 Structures of training works for flood channels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.9.2.3 Dredging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.9.3
River training of meandering rivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.9.3.1 Measures of river training for stabilizing river channels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.9.3.2 Cutoff . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.9.4
River training of wandering rivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.9.5
River training of anabranched rivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.9.5.1 Measures for stabilizing flow diversion ratio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.9.5.2 Works of fork-channel blockade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.9.6
River training of straight rivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.9.7
Regulation of shoal reaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.9.7.1 Parameters for designing navigation courses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
76
76
76
77
77
77
77
77
78
78
78
78
79
79
79
79
79
79
79
79
80
80
80
80
80
81
81
81
81
81
81
82
82
82
82
83
83
83
83
83
83
83
83
84
CHAPTER 5 — RESERVOIR SEDIMENTATION AND IMPACT ON RIVER PROCESSES . . . . . . . . . . . . . . .
5.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.1.1
Dam construction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.1.2
Rate of loss of storage capacity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.1.3
Sustainable development of reservoirs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.1.4
Prediction of reservoir sedimentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.1.5
Issues related to reservoir sedimentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.2
Processes of deposition in reservoirs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.2.1
Movement of sediment in reservoirs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.2.2
Basic characteristics of reservoir deposits (Qian, et al., 1987) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.2.2.1 Longitudinal profiles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.2.2.2 Delta . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.2.2.3 Lateral distribution of deposits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
87
87
87
87
88
88
88
88
88
88
88
89
91
CONTENTS
vii
Page
5.2.2.4
5.2.2.5
5.2.2.6
5.3
5.3.1
5.3.2
5.3.2.1
5.3.2.2
5.3.3
5.3.3.1
5.3.3.2
5.3.3.3
5.4
5.4.1
5.4.2
5.4.3
5.5
5.5.1
5.5.2
5.5.2.1
5.5.2.2
5.5.2.3
5.6
5.6.1
5.6.2
5.6.3
5.6.4
5.6.4.1
5.6.4.2
5.6.4.3
5.6.4.4
5.6.4.5
5.6.4.6
5.6.5
5.6.5.1
5.6.6
5.6.6.1
5.6.6.2
5.6.6.3
5.6.6.4
5.6.6.5
5.6.6.6
5.7
5.7.1
5.7.1.1
5.7.1.2
5.7.1.3
5.7.1.4
5.7.1.5
5.7.1.6
5.7.1.7
5.7.2
5.7.2.1
5.7.2.2
5.8
5.8.1
5.8.2
5.8.3
Spatial distribution of deposits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Headward extension of backwater deposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Physical characteristics of deposits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Sediment release from reservoirs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Sediment release during flood detention . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Density current venting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Phenomenon and formation of density current . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Venting of density current . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Erosion in reservoirs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Retrogressive and progressive erosion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Erosion in the fluctuating backwater region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Empirical method of erosion prediction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Empirical estimation of long-term deposition in reservoirs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Method of trap efficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Method of rate of storage capacity loss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Process of depletion of reservoir storage capacity (lifespan of a reservoir) . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Numerical modelling of reservoir sedimentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Basic equations (for unit width) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Continuity equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Momentum equation of one-dimensional sediment-laden flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Supplementary equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Reservoir sedimentation management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Universality of reservoir sedimentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Indicators of reservoir sedimentation problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Basic operating rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Sediment design of hydrological projects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Collection and evaluation of basic data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Sediment input . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Sediment design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Prevention of sediment problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Prediction of the fluvial processes below a project . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Planning for sediment measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Methods of reducing sediment input in reservoirs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Soil conservation practice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Overview of remedial measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Drawdown flushing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Reservoir emptying . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Lateral erosion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Siphon dredging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Dredging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Design of sediment sluicing facilities of reservoirs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Fluvial processes below reservoirs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Fluvial processes below impounding reservoirs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Changes in flow regime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Drastic reduction in sediment load and concentration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Erosion below dams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Armouring of bed sediment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Adjustment of longitudinal profile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Adjustment of cross-sectional shape . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Adjustment of channel pattern . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Fluvial processes below detention reservoirs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Changes in flow and sediment regimes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Aggravation of deposition below dams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Case studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Liujiaxia Project . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Sanmenxia Project . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Heisonglin Project . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
91
92
92
93
93
94
94
94
95
95
95
95
96
96
97
97
97
97
98
98
98
98
99
99
100
100
100
101
101
101
101
102
102
102
102
103
103
103
103
103
103
103
104
104
104
104
104
104
105
105
106
106
106
106
106
107
107
108
viii
CONTENTS
Page
5.8.4
Shuicaozi Project . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.8.5
Guanting Reservoir . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.8.6
Tarbela Dam Project . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.9
Measurement of erosion and deposition in the reservoir . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.9.1
Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.9.1.1 Contour method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.9.1.2 Range-line method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.9.1.3 Composite method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.9.2
Instrumentation for positioning and depth sounding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.9.2.1 Depth sounding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.9.2.2 Positioning of sounding points . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.9.2.3 Surveying system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.9.2.4 Positioning by the Global Positioning System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.9.2.5 Measuring sediment thickness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.9.3
Measurement of bed material composition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.9.3.1 Undisturbed sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.9.3.2 Radioisotope density probe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.9.3.3 Selection of sampling points . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.9.4
Data processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.9.4.1 Computation of reservoir capacity or amount of deposition or erosion in river reaches . . . . . . . . . . . . . . . . . .
5.9.4.2 Computation of capacity from topographic surveys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.9.4.3 Unit weight of sediment deposits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
109
110
110
111
111
111
111
112
112
112
113
113
113
113
113
113
114
114
114
114
115
115
116
CHAPTER 6 — OPERATIONAL METHODS OF SEDIMENT MEASUREMENT . . . . . . . . . . . . . . . . . . . . . . .
6.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.1.1
Type of sediment load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.1.2
Network for measurement of sediment transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.1.3
Classification of hydrometric stations for sediment measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.1.4
Total load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.1.5
Sedimentation surveys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.1.6
Parameters to be collected for a complete sediment data set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.2
Measurement of suspended sediment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.2.1
Method of measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.2.1.1 Measurement of suspended sediment discharge in a vertical . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.2.1.2 Measurement of sediment discharge in a cross-section . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.2.1.3 Sampling for size analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.2.1.4 Frequency and timing of sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.2.2
Computation of sediment discharge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.2.3
Measuring devices and instrumentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.2.3.1 Sampler for taking representative samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.2.3.2 Basic requirements for an ideal sampler . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.2.3.3 Some developments in mechanical devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.2.3.4 Some developments in the in situ measurement of sediment concentration . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.2.3.5 Intercomparison of measuring devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.3
Measurement of bed load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.3.1
Direct measurement of bed load discharge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.3.1.1 Characteristics of bed load movement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.3.1.2 Frequency of measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.3.1.3 Selection of sampling verticals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.3.2
Indirect method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.3.2.1 Sedimentation process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.3.2.2 Dune tracking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.3.2.3 Tracer method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.3.2.4 Investigation of the lithologic properties of sediment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.3.3
Measuring devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.3.3.1 Technical requirements for an ideal bed load sampler . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.3.3.2 Various kinds of bed load samplers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
118
118
118
118
118
118
119
119
119
119
119
121
123
123
124
125
125
125
125
126
126
127
127
127
128
128
129
129
129
129
129
130
130
130
CONTENTS
ix
Page
6.3.3.3
6.3.4
6.3.4.1
6.3.4.2
6.3.5
6.4
6.4.1
6.4.1.1
6.4.1.2
6.4.1.3
6.4.2
6.4.2.1
6.4.2.2
6.4.2.3
6.4.2.4
6.4.3
6.5
6.5.1
6.5.1.1
6.5.1.2
6.5.1.3
6.5.1.4
6.5.2
6.5.2.1
6.5.2.2
6.5.2.3
6.6
6.6.1
6.6.1.1
6.6.1.2
6.6.1.3
6.6.1.4
6.6.2
6.6.3
6.7
6.7.1
6.7.2
6.7.2.1
6.7.2.2
6.7.2.3
6.7.2.4
6.7.2.5
6.7.2.6
6.7.3
6.7.4
6.7.5
6.8
6.8.1
6.8.2
6.8.3
6.8.4
6.8.5
6.8.6
6.8.7
6.8.8
6.8.9
6.8.10
6.8.11
New developments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Calibration of samplers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Direct field calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Laboratory calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Computation of bed load discharge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Measurement of total sediment discharge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Direct methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Measurement of suspended sediment and bed load discharge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Measurement by means of turbulence flume . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Measurement by sediment accumulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Computation of total sediment load from measured suspended sediment discharge data at a hydrometric station . .
Theoretical background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
The modified Einstein procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Correction coefficient . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Ratio of bed load discharge to suspended-sediment discharge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Comments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Laboratory procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Determination of sediment concentration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Evaporation method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Filtration method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Displacement method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Accuracy requirement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Size analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Methods for size analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Treatment of samples for size analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Measurement of physical properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Data processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Data processing for suspended load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Computation of sediment discharge and cross-sectional average sediment concentration . . . . . . . . . . . . . . . .
Computation of average daily sediment discharge and concentration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Sediment transport curve . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Data processing for suspended sediment size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Data processing for bed load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Examination of processed data and data processing using computers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Assessment of accuracy and reliability in measurement of sediment transport . . . . . . . . . . . . . . . . . . . . . . . . .
General description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Major factors influencing the reliability of measurement of sediment transport . . . . . . . . . . . . . . . . . . . . . . . .
Apparatus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Characteristics of measuring sections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Sampling frequency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
In situ measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Measurement of concentration and size analysis in the laboratory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Computation method and data processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Major factors influencing the reliability of bed load measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Analysis of systematic errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Analysis of random errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Summaries and recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Fundamental concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Implementation of measuring programmes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Measuring site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Measurement of suspended sediment discharge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Corrections for transport in the unmeasured zone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Frequency of measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Sampling apparatus — suspended sediment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Sampling apparatus — bed sediment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Computation of total load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Size analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Method of size analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
131
131
131
132
132
132
132
132
133
133
133
133
134
135
135
135
136
136
136
136
137
137
138
138
140
141
141
141
141
141
142
142
143
143
144
144
144
144
144
144
145
145
146
146
146
147
148
148
148
148
149
149
149
149
149
149
149
149
x
CONTENTS
Page
6.8.12 Data processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.8.13 Assessment of accuracy and reliability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.8.14 Monitoring for sediment quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
149
150
150
150
CHAPTER 7 — WATER QUALITY RELATED TO TRANSPORT OF SEDIMENT AND TOXIC MATERIAL . .
7.1
Effects of sediment and heavy metals on water quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
7.1.1
Absorption of heavy metals in sediment particles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
7.1.2
Effects of sediment particles absorbing heavy metals on water quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
7.2
Effects of sediment and toxic organic material on water quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
7.2.1
Absorption of toxic organic material on sediment particles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
7.2.2
Effects of sediment particles absorbing toxic organic material on water quality . . . . . . . . . . . . . . . . . . . . . . . .
7.3
Water quality model of sediment and toxic organic material and heavy metal . . . . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
153
153
153
154
155
155
157
157
158
FOREWORD
Water resources are the most indispensable resources for human life. As the primary enhancing agent for the sustainable development of
societies and economies, the development and utilization of water resources are becoming more and more important. In the development
of water resources, sediment and related problems have always presented a great challenge. Increasing attention is being focused on a
better understanding of the processes of erosion and sedimentation and their relationship to the surface runoff component of the hydrological cycle. To provide a basic understanding of these processes, WMO published Operational Hydrology Report No. 16, Measurement of
River Sediments (WMO-No. 561) and Operational Hydrology Report No. 29, Manual on Operational Methods for the Measurement of
Sediment Transport (WMO-No. 686). However, there was a great need to provide a manual or report to describe the comprehensive
processes of erosion, sediment transportation, fluvial processes and reservoir sedimentation, etc. Therefore, WMO decided to publish an
updated manual on sediment measurement and management.
The tenth session of the Commission for Hydrology (CHy-X) in 1996 requested Ms Yang Xiaoqing (China), the expert on sediment
of the Working Group on Basic Systems of CHy, to undertake the task of preparing a manual on sediment management and measurement.
With the support of the Ministry of Water Resources of China, an expert team was organized to undertake the work. Ms Yang Xiaoqing,
Dr Long Yuqian, Dr Wan Zhaohui, Dr Zhou Zhide, Messrs Zhou Wenhao, Hua Shaozu, Weng Jianhua, et al., high-level Chinese experts,
were included in the team. It is hoped that this Manual will provide a guide for water resources engineers, planners, managers and
hydrologists.
The authors of the individual chapters are as follows:
Chapter 1: Ms Yang Xiaoqing
Chapter 2: Messrs Hua Shaozu, Liu Xiaoying, Ms Yang Xiaoqing, Wu Deyi
Chapter 3: Dr Wan Zhaohui
Chapter 4: Mr Zhou Wenhao
Chapter 5: Dr Zhou Zhide, Dr Long Yuqian
Chapter 6: Dr Long Yuqian, Ms Zhu Xiaoyuan, Mr Zhou Gangyan
Chapter 7: Messrs Weng Jianhua, P. Literathy (Hungary)
It is with great pleasure that I express my gratitude to Mr B.J. Stewart (Australia), the Chairperson of the Working Group on Basic
Systems of CHy, as well as Messrs G. Leeks (United Kingdom) and G.D. Glysson (United States) for their review and useful recommendations and suggestions.
(G.O.P. Obasi)
Secretary-General
SUMMARY
This report covers a wide range of issues related to sedimentation. Its objectives are to present to readers a basic
understanding of operational methods of sediment transport
measurement, and serve as a practical reference in dealing with
sedimentation engineering.
Ecological and environmental concerns are increasingly
affecting the sustainable development of human societies worldwide. In Chapter 1, the impacts of soil erosion and river and
reservoir sedimentation on ecologies and environments are
discussed, as are potential benefits of sediment as a resource.
Chapter 2 presents soil erosion in detail, including its
basic characteristics, monitoring and prediction of erosion and
sediment yield in a basin, soil and water conservation, and watershed management. Finally, an overview of the global issue of soil
erosion is presented.
In Chapter 3, the contents of sediment transport in rivers
are discussed. The basic concepts of patterns of river sediment
transport form the basis on which to deal with river sediment.
They are elucidated concisely and thoroughly. Following this is a
discussion on bed load, suspended load and total sediment load,
using authoritative papers. Based on a large amount of data and
papers, mainly developed in China, hyperconcentrated flow is
discussed briefly at the end of this chapter.
Chapter 4 elaborates on fluvial processes. The main
points include classification of patterns of alluvial rivers, fluvial
processes of each basic river pattern, and stabilization and rectification of river channels. In this report, the alluvial rivers are
classified into four basic patterns: meandering, wandering,
anabranched and straight. In many literatures, three basic river
patterns are differentiated: meandering, braided and straight. Such
a difference may be induced by the large amount of sediment load
transported by some Chinese rivers.
Reservoirs play a significant role in human society,
including flood control, water supply, power generation, irrigation,
navigation improvement, recreation, etc. With the passage of time,
many reservoirs, particularly those built on sediment-laden rivers,
lose a certain percentage of their storage capacity due to sedimentation. In Chapter 5, the subject of reservoir sedimentation and its
impacts on river processes are expanded upon. Deposition
processes in reservoirs are presented first. Then, methods of estimation of long-term deposition in reservoirs, both empirical and
numerical, are briefly discussed. A discussion of reservoir
management follows, emphasizing the possibility of preserving
long-term reservoir capacity for permanent usage. Six case studies
show the reality of reservoir sedimentation problems.
Accurate sediment data are the basis of every aspect of
sediment management and numerical (computer) modelling of
sedimentation. In Chapter 6, operational methods of sediment
measurement, including measurements of suspended sediment,
bed load and total sediment load, are discussed. Also, laboratory
procedures, data processing and assessment of accuracy and reliability in sediment measurement are presented. Finally, some
recommendations for sediment measurement are given.
Water pollution is an increasingly important issue in many
places, particularly in developing countries. In Chapter 7, water
quality related to the transport of sediment and toxic materials, the
main source of water pollution, is elucidated briefly. To quantify
such impact, a water quality model is introduced.
RÉSUMÉ
Le présent rapport couvre un large éventail de questions relatives à la
sédimentation. Il permettra au lecteur de se familiariser avec les
méthodes de mesure des transports solides et servira de référence
pratique pour tous les aspects scientifiques et techniques de la
sédimentation.
Le développement durable des sociétés humaines est de plus
en plus soumis à des impératifs écologiques. Le chapitre 1 traite des
incidences de l’érosion des sols et de la sédimentation des cours
d'eau et des réservoirs sur l’environnement et les écosystèmes et des
avantages que peuvent présenter les sédiments en tant que ressource.
Dans le chapitre 2, les mécanismes d’érosion des sols sont
décrits en détail, on y évoque aussi la surveillance et la prévision de
l’érosion et des apports solides dans un bassin donné, la conservation
des sols et des eaux et la gestion des bassins versants. Le chapitre se
termine par un bilan général de la question de l’érosion des sols.
Le chapitre 3 traite des transports solides dans les cours
d’eau et des principes de base qui régissent les mécanismes en jeu.
La description de ces processus est à la fois concise et exhaustive.
S’appuyant sur des études faisant autorité en la matière, le rapport
aborde ensuite la question de la charge de fond, de la charge solide
en suspension et de la charge solide totale. Enfin, le chapitre se
termine par une présentation succincte de l’écoulement
hyperconcentré, qui a fait l’objet de nombreuses études, en
particulier en Chine.
Le chapitre 4 est consacré aux processus fluviatiles. Parmi
les principaux thèmes abordés, on citera la classification des rivières
alluviales, les processus fluviatiles propres à chaque type de rivière
ainsi que la stabilisation et la rectification du lit des cours d’eau. Le
rapport distingue quatre types principaux de lit fluvial : lit à
méandres, lit divaguant, lit anastomosé et lit rectiligne. Or, dans la
littérature scientifique, on ne distingue le plus souvent que trois
catégories : lit à méandres, lit tressé et lit rectiligne. Cette différence
est peut-être due au fait que certains cours d’eau chinois charrient
une grande quantité de matières solides.
Les réservoirs revêtent une grande importance pour la lutte
contre les inondations, l’approvisionnement en eau, la production
d’énergie, l’irrigation, l’amélioration de la navigation, les loisirs, etc.
Avec le temps, de nombreux réservoirs, en particulier ceux qui ont
été construits sur des cours d’eau à forte charge solide, ont perdu une
partie de leur capacité de stockage à cause de la sédimentation. La
xiv
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
question de la sédimentation des réservoirs et de ses répercussions
sur les processus fluviatiles est traitée en détail dans le chapitre 5, qui
débute par une description des mécanismes de dépôt dans les
réservoirs, avant d’exposer brièvement les méthodes — empiriques
et numériques — d’estimation de ces dépôts considérés sur de
longues périodes. Les auteurs s’intéressent ensuite à la gestion des
réservoirs, envisagée dans la perspective de préserver durablement
leur capacité. Les problèmes de sédimentation des réservoirs sont
illustrés par six études de cas.
La gestion des sédiments et la modélisation numérique de
la sédimentation doivent s’appuyer à tous les niveaux sur des
données précises. Le chapitre 6 est consacré aux méthodes de
mesure des sédiments, notamment de la charge solide en
suspension, de la charge de fond et de la charge solide totale. Il est
aussi question des pratiques de laboratoire, du traitement des
données ainsi que de la précision et de la fiabilité des mesures
relatives aux sédiments. Le chapitre se termine par quelques
recommandations dans ce domaine.
La pollution de l’eau est un problème qui devient de plus en
plus préoccupant, en particulier dans les pays en développement. Le
chapitre 7 évoque brièvement la question de la qualité de l’eau dans
le contexte du transport de sédiments et de matières toxiques,
principale cause de la pollution de l’eau. Un modèle de la qualité de
l’eau est utilisé pour quantifier les effets de cette pollution.
РЕЗЮМЕ
В настоящем отчете охватывается широкий круг вопросов,
касающихся отложения наносов. Его задачи – представить
читателям основы понимания процесса переноса наносов и
оперативных методов его измерения, а также послужить
практическим справочником в решении прикладных задач,
связанных с наносами.
Обеспокоенности по поводу экологии и окружающей
среды во все возрастающей степени влияют на устойчивое
развитие человеческих сообществ по всему миру. В главе 1
рассматриваются воздействия эрозии почвы, а также
заиления рек и водохранилищ на экологические аспекты и
окружающую среду, так же как и потенциальные выгоды
использования наносов как ресурса.
В главе 2 подробно представлены сведения об эрозии
почвы, включая ее основные характеристики, мониторинг и
предсказание эрозии и твердого стока в бассейне, сохранение
почв и воды, а также регулирование водосборов. И наконец,
представлен общий обзор глобальной проблемы эрозии почв.
В главе 3 обсу ждается су ть процесса переноса
наносов в реках. Базовые концепции процесса переноса
речных наносов формируют основу, на которой
рассматриваются проблемы, связанные с речными наносами.
Они разъясняются сжато, но тщательно. В соответствии с
этим, а также с использованием авторитетных работ,
рассматриваются вопросы, касающиеся донных и
взвешенных наносов, а также суммарный твердый сток. На
основе использования большого количества данных и работ,
подготовленных главным образом в Китае, в конце данной
главы кратко рассматриваются потоки с очень большим
содержанием наносов.
Четвертая
глава
посвящена
процессам,
происходящим в реках. Основные вопросы включают:
классификацию типов аллювиальных рек, флювиальные
процессы в реках каждого основного типа, а также
стабилизацию и спрямление речных русел. В настоящем
отчете аллювиальные реки классифицируются по четырем
основным типам: меандрирующие, блу ждающие,
разветвляющиеся на рукава и прямые. Во многих
литературных источниках классификация рек проводится по
трем основным типам: меандрирующие, разветвляющиеся на
рукава и прямые. Такое различие может быть вызвано тем,
что некоторые реки в Китае переносят большое количество
наносов.
Водохранилища играют значительную роль в
человеческом обществе, включая регулирование паводков,
водоснабжение, выработку энергии, ирригацию, улучшение
навигации, отдых и т.д. Со временем многие водохранилища,
в особенности те, которые построены на реках, несущих
много наносов, теряют из-за отложения наносов
определенный процент своего полезного объема. В главе 5
подробно излагается вопрос о заилении водохранилищ и его
воздействиях на речные процессы. Сначала представлены
процессы отложения наносов в водохранилищах. Затем
кратко рассматриваются как эмпирические, так и численные
методы оценки отложения наносов в водохранилищах за
длительные периоды времени. Затем следует описание
регулирования водохранилищ с основным вниманием к
возможности долгосрочного сохранения полезного объема
водохранилища с целью его постоянного использования. На
примере шести конкретных исследований показана
реальность проблем заиления водохранилищ.
Точные данные о наносах являются основой каждого
аспекта регулирования стока наносов и численного
(компьютерного) моделирования процесса отложения
наносов. В главе 6 рассматриваются оперативные методы
измерения наносов, включая измерения взвешенных и
донных наносов и общего их количества. Также представлены
применяемые в лабораториях процедуры, обработка данных
и оценка точности и надежности при измерении наносов. И
наконец, приводятся некоторые рекомендации, касающиеся
измерения наносов.
Загрязнение воды становится повсеместно и во все
возрастающей степени важным вопросом, в особенности в
развивающихся странах. В главе 7 кратко освещаются
вопросы качества воды, связанные с переносом наносов и
токсичных материалов, являющихся основным источником
загрязнения воды. Для количественного описания такого
воздействия приводится модель качества воды.
SUMMARY
xv
RESUMEN
Este informe, que abarca gran número de cuestiones relacionadas
con la sedimentación, tiene dos objetivos: dar al lector una idea
básica de los métodos utilizados a nivel operativo para la medición
del transporte de sedimentos, y servir de referencia práctica en
materia de ingeniería de la sedimentación.
Las preocupaciones ecológicas y ambientales inciden cada
día más en el desarrollo sostenible de las sociedades humanas en
todo el mundo. En el Capítulo 1 se examinan los efectos para la
ecología y el medio ambiente de la erosión de los suelos y de la
sedimentación en ríos y embalses, así como los posibles beneficios
de los sedimentos aprovechados como recurso.
En el Capítulo 2 se aborda en detalle el tema de la erosión
de los suelos, incluidas sus características básicas, la vigilancia y
predicción de la erosión, el aporte de sólidos en una cuenca, la
conservación de suelos y aguas, y el manejo de cuencas. Por
último, se presenta un panorama general del tema global de la
erosión de los suelos.
En el Capítulo 3 se examina el tema del transporte de
sedimentos en los ríos. Se examinan de manera concisa y detenida
los conceptos básicos del comportamiento de los sedimentos
transportados en los cauces fluviales, que forman la base del análisis
del sedimento de los ríos. Esto va seguido de un análisis del arrastre
de fondo, la carga en suspensión y del arrastre total, en la que se
hace referencia a documentos de autoridades en la materia. En la
última sección del capítulo se estudia brevemente el flujo
hiperconcentrado, sobre la base de un importante volumen de
información y de documentos técnicos, provenientes mayormente
de China.
En el Capítulo 4 se analizan detenidamente los procesos
fluviales. Los principales puntos incluyen la clasificación del
comportamiento de los ríos aluviales, los procesos fluviales
asociados con cada comportamiento básico, y la estabilización y
rectificación de los canales de los ríos. En ese informe, los ríos
aluviales se clasifican en cuatro categorías básicas: sinuosos,
tortuosos, divergentes y rectos. Muchos autores emplean una
clasificación basada en tres categorías: sinuosos, trenzados y rectos.
La diferencia puede obedecer a la elevada carga de sedimentos que
transportan algunos ríos de China.
Los embalses desempeñan un papel importante para la
sociedad en campos como control de crecidas, suministro de agua,
generación de energía hidroeléctrica, riego, mejora de la navegación,
recreo, etc. Con el paso del tiempo, en muchos embalses, en especial
los construidos en ríos que arrastran gran volumen de sedimentos, se
ha observado una cierta reducción de su capacidad de
almacenamiento debido a la sedimentación. En el Capítulo 5, el tema
de la sedimentación en los embalses y sus efectos sobre los procesos
fluviales es objeto de un análisis más detallado. Se presentan primero
los procesos de deposición en los embalses. A continuación se
examinan los métodos, tanto empíricos como numéricos, de
estimación de la deposición a largo plazo en los embalses. Más
adelante se aborda la cuestión de la gestión de los embalses,
haciendo hincapié en la posibilidad de preservar su capacidad a largo
plazo para el uso permanente. Seis estudios de caso muestran la
realidad de los problemas relacionados con la sedimentación en los
embalses.
La exactitud de los datos es esencial para todos los aspectos
de la gestión de los sedimentos y de la modelización numérica de la
sedimentación con empleo de computadoras. En el capítulo 6 se
examinan los métodos operativos de medición de los sedimentos,
incluida la medición de la carga/sedimento en suspensión, del
arrastre de fondo y del arrastre total. Asimismo, se presentan los
procedimientos de laboratorio, el procesamiento de datos y la
evaluación de la exactitud y fiabilidad de las mediciones de los
sedimentos. Por último, se hacen algunas recomendaciones en
cuanto a la medición de los sedimentos.
La contaminación de las aguas es una cuestión que cada
día cobra mayor importancia en muchos lugares, sobre todo en los
países en desarrollo. En el Capítulo 7 se analiza brevemente la
calidad de las aguas en relación con el transporte de sedimentos y
materiales tóxicos, la principal fuente de contaminación de las
aguas. Se introduce un modelo de calidad de las aguas con el fin de
cuantificar esos efectos.
CHAPTER 1
ECOLOGY AND ENVIRONMENT RELATED TO SEDIMENTATION
1.1
INTRODUCTION
Sedimentation impacts many aspects of the environment — soil
erosion, water quality, water supply, flood control, river regulation, reservoir lifespan, groundwater table, irrigation, navigation,
fishing, tourism, etc. It has attracted increasing attention from the
public and engineers in the field. In this Manual, the authors try to
describe the main problems and issues related to river and reservoir sedimentation to help the reader understand them better.
1.2
IMPACTS OF SOIL EROSION ON ECOLOGY
AND ENVIRONMENT
Soil and water conservation is one of the most critical environmental issues facing many countries, especially developing
countries. Water is the source of life and soil is the root of existence. Water and soil resources are the most fundamental materials
on which people rely for existence and development. The development of society is determined by its capacity to use its resources.
Some of these resources may in time become exhausted or deteriorate. Soil has been defined by the International Science Society
as ‘a limited and irreplaceable resource’, and the growing degradation and loss of soil means that the expanding population in many
parts of the world is pressing this resource to its limits. In its
absence, the biosphere environments of man would collapse, with
devastating results for humanity.
Soil and water conservation is a multidisciplinary
applied science studying soil and water loss and control measures,
in order to protect, improve and support rational uses of soil and
water resources in mountainous areas which suffer water and wind
erosion. Conservation also helps maintain and increase land
productivity.
Soil and water loss causes land resource destruction and
reduction in soil fertility, which leads to the deterioration of the
environment and the loss of ecological balance, causing natural
disasters and constraining the development of agriculture, consequently increasing poverty.
China, for example, is seriously affected by soil erosion.
Its total erosion area is 3.67 million km2, being 38.2 per cent of
the total territory, of which 1.79 million km2 is eroded by water
and 1.88 million km2 by wind. Soil loss is 5 billion tons per year.
The land is lost at a rate of about 0.13 million ha per year. In some
eroded areas, land destruction and deterioration have even threatened people’s existence. The Loess Plateau, one of the most
seriously eroded areas in China, contributes a large amount of
sediment to the Yellow River. According to long-term statistics,
1.6 billion tons of sediment are lost annually into the Yellow River.
Two thirds of the total sediment is transported by the river in
suspension and poured into the near sea and deep sea. The remaining one third of the sediment load is deposited in the lower
reaches of the river. As a result, the river bed rises by 8 to 10 cm
each year to create an unfavourable situation in which the river
bed is 4 to 10 m higher than the ground elevation outside the
levee. This has brought flood and drought disasters and poverty,
and has greatly threatened the safety of the population. It is also
the main constraint upon the development of agriculture and the
economy in the river basin. The 1.6 billion tons of sediment
contain 40 000 tons of nitrogen, phosphate and potash fertilizers
(N, P and K fertilizers). In north-east China, 7 million tons of N, P
and K are lost each year due to soil erosion.
The objectives of erosion control are to protect the two
most valuable natural resources, i.e. soil and water, and to prevent
the occurrence of the unfavourable consequences of such a loss.
Erosion control measures must be harmonized with agricultural production and water resources conservation. Such
measures should cover the following aspects:
(1) Comprehensive treatment. Soil and water conservation
requires the unified planning of water systems, forests, farmland, and roads in mountainous and hilly areas, to achieve
integrated management and comprehensive development.
(2) Principal body of construction. Soil and water conservation
is trans-sectoral and multidisciplinary. It should insist on
adopting a combination of vegetative measures to protect
land surfaces, structural measures to reduce and disperse
runoff on land surfaces, and tillage measures to prevent soil
loss caused by agricultural activity.
(3) Watershed management planning and activities. These
should bring ecological, social and economic benefits to
stakeholders, so as to ensure sustainable development of
watershed management.
In a river basin, soil erosion causes the deterioration of
ecology and environment and the degradation of agricultural
production. Even more seriously, it makes farmland forever
useless by reducing the fertility and productivity of soil. Sediment
deposited in river channels raises the water level of floods, and
therefore brings a series of ecological and environmental problems
and aggravates flood disasters, not only by the flood itself but also
by the sediment carried by the flood. On the other hand, the scouring of river channels lowers the water level and causes problems
for water supply and navigation, and also threatens the safety of
river training works. Reservoir sites are limited, precious, and not
renewable resources. Reservoir sedimentation reduces the storage
capacity and impacts the functions designed for reservoirs, such as
water supply, flood control, irrigation and power generation.
Downstream from reservoirs, scouring of river channels occurs
and also has a number of negative impacts on ecology and environment. In this chapter, the impacts of sediment on ecology and
environment will be introduced.
Sediment in water has two opposite effects on water
quality and environment. On the one hand, sediment particles in
water, especially the fine ones, absorb some pollutants and thereby
improve water quality to a certain degree. On the other hand, sediment also serves as the major pollutant, carrier and storage agent
of other pollutants, such as pesticides, residues, absorbed phosphorus, nitrogen, organic compounds, pathogenic bacteria and
viruses, and affects the water purity, transparency and quality. The
details of the impacts of sediment on water quality are described
in Chapter 7.
Soil erosion and sedimentation are among the greatest of
the world’s modern environmental concerns. In many parts of the
2
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
world, soil erosion has not only caused land deterioration and
hampered the development of agriculture and industry, but also
increased sediment yield from the watershed. Soil erosion thus
often devastates topsoil and causes nutrient loss. Some mountain
and hilly regions have become bare areas, causing environmental
degradation. Emergency events of debris flow, bank collapse and
landslides are often disastrous for people’s lives and property, as
well as infrastructure. The main impacts of soil erosion are
described below.
1.2.1
Desertification and degradation of agricultural
production
One of the most serious consequences of worldwide soil erosion is
desertification. Population growth in some developing countries,
inappropriate land use, deforestation, soil erosion by human activities, and inappropriate water resources utilization cause severe land
desertification. At present, the total area affected by desertification
in the world is 45.61 million km2, accounting for 35 per cent of
total global land. It accounts for 55 per cent and 75 per cent of the
total area of Africa and Australia, respectively. Areas affected by
desertification increase by 50 000 to 70 000 km2 annually, including about 6 million ha of farmland (Zhu, 1992). China is a large
country, but only one tenth of its land is cultivable. Desertification
has reached 334 000 km2 and is increasing at an annual rate of
1 560 km2. In Mongolia, the total land area is 156.5 million ha,
among which 5.0 million ha is already covered with sand.
Erosion damages soil structure, causing the loss of fertility, and consequently reduces agricultural production. Much of
South-West Asia, China, India, South-East Asia, North Africa,
Central America and Mexico suffer from severe land degradation.
In South America, land degradation is most acute on the cultivated
lands of the Andes Mountains. Water and wind erosion has
damaged some Argentine farmland. In China, the soil-eroded area
has reached 1.79 million km2, accounting for 18.7 per cent of the
total territory, with an increase of 2 460 km2 per year. In the last
50 years, 2.6 million ha of farmland has been lost due to soil
erosion. About 5 billion tons of eroded sediment enters rivers,
lakes and seas. In India, it is estimated that about 6 billion tons of
soil is lost each year as a result of sheet erosion. In addition, gully
and ravine erosion damages 8 000 ha of farmland annually.
1.2.2
Sediment-related disasters
Sediment-related disasters, such as debris flow, landslides and
slope collapses, often induce huge damage to people, economies
and the environment. Debris flows exist to some extent in the
mountainous areas of more than 70 countries. China is a mountainous country, of which 69 per cent of the territory is composed
of mountains and hills. Owing to a peculiar natural and humangeographic environment, almost all provinces, autonomous
regions and municipalities are endangered and troubled by debris
flows, landslides, and other sediment-related disasters. Incomplete
statistics show that, in China, there are more than 8 500 debris
flow ravines and 100 000 places susceptible to landslides, which
threaten the safety of 36 main train lines, 36 per cent of the roads
and more than 200 medium and small cities. Debris flows occur
far more frequently and forcefully than in other countries, and
caused a loss of more than US$ 12 billion in 1990. In 1953, a
glacier-induced debris flow occurred at Guxiang Ravine, Bomi,
Tibet, with a peak discharge of 28 600 m3 s–1. Taiwan is a mountainous area. The characteristics of geography and climate — i.e.
broken rock, steep slope, torrential and concentrated rain, and
short and rapid flow — cause debris flows, landslides and slope
collapses to occur frequently.
Indonesia has about 17 active volcanoes. It experiences
not only direct disasters due to frequent eruptions, pyroclastic
flow and nuce ardente, but also indirect disasters due to secondary
lahar caused by rainfall after eruptions have occurred. Many lives
have been lost. Also, huge amounts of volcanic product such as
ash, sand and gravel are deposited loosely on the slope around a
crater during the eruption. According to records, approximately
300 million m3 of volcanic product were produced by the eruption
of Mt. Agung in 1963, 22 million m 3 by the eruption of Mt.
Merupi in 1969 and 53 million m 3 by the eruption of Mt.
Galunggung in 1982 (Sabo, 1995).
In Japan, mountainous areas account for 74 per cent of
total territory. Earthquakes, debris flows and volcanic eruptions
occur often. During the torrential rains of August 1993, total rainfall exceeded 800 mm. These heavy rains caused a series of
overbank floods and debris flows in the Kagoshima area, including one that struck a train. These floods and debris flows caused
the interruption of transportation in the region due to the submergence of roads, and seriously interrupted the lives of local
residents by cutting power lines or breaking water supplies. The
successive heavy rains left 141 people dead or missing and about
150 000 houses damaged. The total losses, including damage to
public facilities, agriculture, forestry, and fishing was estimated at
1 trillion Japanese Yen. Another type of debris flow in Japan is
caused by volcanic eruption. Large amounts of rock, earth, sand
are released from volcanic eruptions and loosely pile up on slopes.
When heavy rain comes, the volcanic materials form debris flows
with a huge damage capacity. After Mt. Unzen Fugendake erupted
in 1990, a pyroclastic flow occurred in June 1991. Forty-three
people were reported dead or missing, nine were injured, and 179
buildings burned down.
1.3
IMPACTS OF RIVER SEDIMENTATION ON
ECOLOGY AND ENVIRONMENT
Deposit and scour are common in rivers because of the difference
between sediment load and the real sediment transportation capacity of flow. Deposition in river channels raises the elevation of
river beds. Consequently, it enhances the water level at the same
discharge, and increases the occurrence and the damage of floods.
On the other hand, scour brings some safety problems for river
training works, lowers water levels, and therefore affects water
supply and navigation along rivers.
1.3.1
River sediment and flood disasters
Owing to serious soil erosion in the river basin, a large amount of
sediment load enters the Yellow River and is deposited in the
lower reaches. The river bed rises about 5 to 10 cm annually. The
river bed below Zhengzhou City, the capital of Henan Province, is
higher than the ambient ground, a so-called suspended river
(Figure 1.1), and the river channel serves as the watershed boundary of the Haihe and Huaihe Rivers. If the river dikes were to
break along the lower reaches, the maximum area affected by
floods would be 250 000 km2 north to Tianjin City and south to
the Huaihe River, an area among the most economically developed in China. The maximum population affected by floods
would be 100 million. Such floods have occurred a number of
times in history.
CHAPTER 1 — ECOLOGY AND ENVIRONMENT RELATED TO SEDIMENTATION
3
L (km)
Figure 1.1 — Suspended river of the Lower Yellow River of China.
1.3.1.1 CONVEYANCE CAPACITY OF RIVERS
The conveyance capacity of a river changes due to deposition and
scouring of the river channel. On the Lower Yellow River, the river
channel is complicated, composed of the main channel and flood
plains, and the total width may reach 10 to 20 km. There are
0.22 million ha of cultivated farmlands, and about 1.5 million
people live on the flood plains. Middle and low floods are artificially constricted within the main channels and 85 per cent of the
deposit is in the main channel. Therefore, the conveyance capacity
of the main channel significantly decreases, which is called river
channel shrinking. From May 1986 to May 1994, the flow areas of
the main channel of 36 cross-sections on the Lower Yellow River
decreased by about 27 per cent. The water stage under the flow
discharge of 3 000 m3 s–1 rises by 0.12 to 0.15 m annually. The
bankfull discharge was reduced to between 2 800 and
3 700 m3 s–1. The number of occurrences of flow over flood plains
therefore greatly increased in recent years. In 1996, the flood
discharge was only 7 860 m3 s–1 at Huayuankou Station near
Zhengzhou City, which was much less than the flood of 22 300 m3
s–1 in 1958, but the flood stage was 94.73 m, the highest recorded,
0.91 m higher than that of 1958. The inundated land on the flood
plain was about 250 000 ha, with 1.07 million people affected.
The direct loss was about US$ 800 million (Hu, 1996).
1.3.1.2 FLUVIAL PROCESS AND INSTABILITY OF RIVER CHANNEL
The fluvial processes in both planar and longitudinal directions
significantly affect river behaviour and stability, especially for
large rivers, which play very important roles in a country’s
sustainable development of its economy, ecology and environment. The fluvial processes may cause or aggravate the disasters.
The 1998 flood in the Middle Yangtze River of China was a good
example of this.
The middle reaches of the Yangtze River are a river-lake
system composed of the Jinjiang River (i.e. the Middle Yangtze
River), Dongting Lake and other lakes (Figure 1.2). During floods,
part of the water is delivered to Dongting Lake through three
connecting river channel, mitigating the peak flood water passing
through the Jinjiang River channel. Owing to sediment deposition
at the end reaches of the three connecting channels, their
conveyance capacities have greatly decreased. The Lower Jinjiang
was once a typical meandering river, with 12 sharp bends. Two
bends, Zhongzhouzi and Shangchewan, were artificially cut off in
1967 and 1969, respectively, and the Shatanzi was naturally cut
off. In 1972, the cutoffs of the three bends reduced the river length
by 81 km. Therefore, the bed slope, flow and sediment
conveyance capacities increased. This reduced the ratio of flow
entering Dongting Lake to the remaining flow in the main stream,
and caused degradation of the Lower Jinjiang River.
Consequently, the scoured sediment deposits flowed downstream
from Luoshan to Wuhan City, capital of Hubei Province, and
raised the flood stage there. During the flood of 1931, 50.4 per
cent of the peak flow of 66 700 m3 s–1 was delivered to Dongting
Lake by the three connecting rivers, among them 28.4 per cent
(18 970 m3 s–1) by the Ouchi River. However, only 6 000 m3 s–1,
10 per cent of the total peak flow discharge, was delivered by the
Ouchi River in 1998. The flow volume annually delivered to
Dongting Lake was 146 billion m 3 from 1951 to 1958, but
decreased to 69.7 billion m3 from 1981 to 1994, which means that
the runoff through the Lower Jinjiang increased by 76.3 billion
m3, significantly aggravating flood disasters. Although the peak
flood of 61 500 m3 s–1 in 1998 was smaller than the 66 800 m3 s–1
in 1954, the flood stages at the stations along the middle reaches
were the highest on record.
Because of lake sedimentation and reclamation through
the occupation of the lake as farmland due to the pressure of population growth, the area and storage capacity of Dongting Lake
have been reduced significantly, as shown in Table 1.1. This
greatly weakens its regulatory role during floods of the Yangtze
River. Only about 10 billion m3 of water volume was diverted to
the detention areas during the 1998 floods, compared with 102.3
billion m3 in 1954. This is one important reason why the 1998
floods created a record high stage.
1.3.1.3 SAFETY OF TRAINING WORKS
Near bridges, groins, and other training works, flow velocity may
be larger than the upstream and downstream flow, due to the
reduction of flow width caused by the structures. Scouring of river
channels in the vicinity of structures is a common phenomenon,
and threatens the safety of the structures and training works. If the
estimated scour is wrong in the design stage, accidents may occur.
Jinjiang R.
Figure 1.2 — The middle reaches of the Yangtze River.
4
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
Table 1.1
Change in area and storage capacity of Dongting Lake
Year
Area (km2)
Storage capacity (109 m3)
1825
6000
1896
5400
1932
4700
1949
4350
29.3
1954
3910
26.8
1958
3141
22.8
1971
2820
18.8
1977
2740
17.8
1983
2691
17.4
1995
2625
16.7
1.3.1.4 SEDIMENT DEPOSITS BY FLOODS
Floods carry much higher sediment concentrations than normal
flows. Damage estimates of floods should therefore also include
environmental deterioration by sediment deposition and the high
cost of clearing up the deposition. Along the two banks of the
Yellow River, there are more than 40 alluvial fans formed by
floods. The sediment of the fans has a high content of fine particles which are easily blown away by wind. Those areas thus
become desertified. In August 1982, about 400 million m3 of flood
water was diverted to the Dongping detention area, and at the
same time about 5 million m3 of sediment (mostly sand) was
diverted. Consequently, 425 ha of farmland were lost because of
the sediment deposition.
1.3.1.5
1.3.2.2
IMPACTS OF RIVER CHANNEL SHIFTING ON ENVIRONMENT
AND ECOLOGY
In the Yellow River, about 1 billion tons of sediment enter the
delta region annually, most of which deposits in the delta coastal
area and near the sea, creating some new land (average of 20 to
30 km2 per year) and extending the river to the sea. Because of the
deposition, the river shifted its channel many times and created the
Grand North China Plain. Figures 1.3 and 1.4 show the modern
delta and the change of the river mouth channels since 1855.
Owing to the frequent channel shifting, the development of the
local economy was limited. China has made a great effort to stabilize the river mouth (Yang and Zhang, 1998).
VARIATION OF GROUNDWATER LEVEL AND SALINITY BY
RIVER SEDIMENTATION
Accumulated river sedimentation raises river water levels. Owing
to the recharge of river water to groundwater in the adjacent areas,
groundwater levels along river banks may rise and cause farmland
salinity or other environmental problems. In the Lower Yellow
River, the flow water is generally 3 to 5 m higher than the adjacent
ground surface. It is estimated that about 49 800 tons of salt is
recharged annually to the groundwater by lateral filtration of river
flow. The groundwater level at 0.5 km from the river channel
reaches 0.6 to 0.7 m, and serious salinity occurs along the river
areas.
1.3.2
1.3.2.1
1965), 233.3 billion m3 (8.04 billion m3 per year) of water and
3.865 billion tons (133 million tons per year) of sediment were
diverted into irrigation systems. From 1981 to 1990, the annual
values were 11.1 billion m3 and 120 million tons, respectively.
Among the 120 million tons of sediment, 33.22 per cent, 35.32 per
cent, 22.9 per cent, and 8.56 per cent were deposited respectively
in settling pools, irrigation systems, farmland and drainage
systems. This means that 77.1 per cent of sediment deposition, i.e.
92.52 million tons annually and about 3 billion m3 in total, must
be dredged or dealt with. In 1990, about 50 000 ha of settling pool
areas at heads of the irrigation systems were filled up with about
1 billion m3 of sediment. Moreover, the deposition in the canals
was dredged out and placed on a narrow belt along the two sides
of the canal. These depositions contain coarse sand, and form sand
hills or dunes. It is dry and windy in the winter and spring
seasons, which causes the local people to suffer disasters due to
serious desertification.
Environment of sediment-laden rivers
DEPOSITION IN IRRIGATION SYSTEMS AND
DESERTIFICATION AT IRRIGATION SYSTEM HEADS
The problem of sediment deposition in irrigation systems is
commonly encountered, especially in heavily sediment-laden
rivers. Dredging and clearing up the deposition in irrigation canals
is high-cost and labor-intensive work. There are many factors to
be taken into consideration to prevent, reduce and deal with the
sediment entering irrigation systems. Appropriate intake type;
settling pool at the head; reasonable design of canal, including
diverted discharge and sediment concentration; bed slope; side
slope; size of cross-section; material (roughness); operation; and
maintenance are some examples of such factors.
On the Lower Yellow River, a large amount of farmland
relies heavily on irrigation from the river. In total, there are 128
intakes and 1.86 million ha of irrigated land in Henan and
Shandong provinces. From 1958 to 1990 (stopped from 1962 to
1.4
RESERVOIR SEDIMENTATION AND
ENVIRONMENT
1.4.1
Loss of reservoir storage capacity
Reservoir sedimentation and the consequent loss of storage capacity affect reservoir benefits, such as flood control, water supply,
irrigation, navigation, power generation, fishing and recreation. In
arid and semi-arid regions, reservoir sedimentation problems
become most acute where the loss of storage capacity by reservoir
sedimentation is above 1 to 2 per cent per year and the lifetime of
most reservoirs is only 20 to 30 years. The Welbedacht Reservoir
in South Africa, completed in 1973 with a 152.2 million m 3
storage capacity, lost most of its storage capacity (66 per cent)
within the first 13 years of its existence (Rooseboom, 1992).
In India, measurements of reservoir sedimentation indicate that the average annual loss in storage capacity of nine
important reservoirs is between 0.34 and 1.79 per cent. Among 23
large reservoirs, the measured rate of storage loss was less than the
designed rate in only two reservoirs; in other reservoirs, it was
more than five times larger than the designed rate (Central Water
Commission, 1996).
In Italy, an analysis of 268 reservoirs distributed over the
country with a mean age of 50 years showed the following loss of
reservoir storage capacity: 1.5 per cent of the reservoirs were
completely filled by sediment, 4.5 per cent had lost 50 per cent of
their storage capacity, and 17.5 per cent had lost 20 per cent of
their storage. The Ichari Reservoir in India silted up to crest level
of the spillway in two years. The Austin Reservoir lost 41.5 per
cent of its total storage volume from 1893 to 1897, and the dam
gave way in 1900. The new Lake Austin of the Colorado River in
CHAPTER 1 — ECOLOGY AND ENVIRONMENT RELATED TO SEDIMENTATION
5
Figure 1.3 — The modern delta of the Yellow River.
Figure 1.4 — Change of the Yellow River mouth channels since 1855.
Texas lost 95.6 per cent of its capacity in 13 years, the Habra
Reservoir in Algeria 58 per cent in 22 years, and the Wuchieh
Reservoir in Taiwan 98.7 per cent in 35 years. The Indus River
carries about 74 billion m 3 of water and 300 million tons of
suspended sediment per year into the Tarbela Reservoir. In the six
years after its commissioning in 1974, it accumulated about 950
million m3 of sediment in the upper 30 km of the delta (Wu, et al.,
1996).
The loss of storage capacity in reservoirs in the United
States due to sedimentation accounts for an annual monetary loss
of US$ 100 million (Julien, 1994).
The average annual loss of storage capacity for 28
reservoirs in Taiwan, China (with original storage capacities
ranging from 0.65 to 708 million m3) is 1.45 per cent.
In China, the Yellow River is a heavily sedimentladen river with an annual sediment load of 1.6 billion tons.
As of 1989, the losses caused by reservoir sedimentation had
reached 10.9 billion m 3 , accounting for 21 per cent of the
total storage capacity of all reservoirs on the main stem as
well as tributaries. Among them, 2.9 billion m 3 were in the
reservoirs on the tributaries, accounting for 26 per cent of
the total.
6
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
1.4.2
Water pollution by reservoir sedimentation
In the initial stage of reservoir sedimentation, the deposition of
sediment can actually improve the water quality by absorbing
pollutants. According to observations carried out at Guanting
Reservoir, one ton of sediment can absorb 700 g of dissolved lead.
Mud deposited on the reservoir floor displays strong adsorption of
arsenic, of which the concentration on the floor is 10 to 100 times
higher than that in water. Similarly, the concentration of chromium
on reservoir floors is about 20 000 times higher than that in water.
Thus, deposited sediment as well as the layer of water near the
floor will be progressively polluted. Pollutants increasingly accumulate in the lower part of the reservoir. In time, they become so
concentrated that this part of the storage becomes in itself a source
of pollution. Phenol in the reservoir’s water has already slightly
polluted the groundwater of Beijing.
1.4.3
Rise of groundwater level and salinity by deposit
extension in reservoir backwater regions
Sediment deposition in reservoirs extends both downstream and
upstream. When sediment goes into a reservoir, it deposits in the
upper end of the backwater region first, due to slowing flow velocity. With the development of reservoir sedimentation, the
deposition may extend upward and cause a river bed higher than
the normal pool of the reservoir, which induces some environmental and ecological problems.
1.4.4
Problems of downstream reservoir
1.4.4.1 FLOOD PLAIN COLLAPSE
Since the impounding of Sanmenxia Reservoir on the Yellow
River of China from 1960 to 1964, the flow discharge and sediment transport rate downstream of the reservoir have been greatly
changed. Most of the sediment carried from upstream has been
stored in the reservoir, and the duration of medium floods (4 000
to 6 000 m3 s–1) has exceeded 20 days due to reservoir regulation.
Total scoured sediment has been as high as 2.31 billion tons, and
300 km2 of flood plains have been scoured away by floods, with a
loss of 47 000 ha of farmland on the flood plains.
The Danjiangkou Reservoir (DJK) is on the Hanjiang
River, the longest tributary of China’s Yangtze River. The river
downstream from the dam was originally a wide, shallow and
braided channel with a rapidly shifting thalweg and lots of welldeveloped unstable mid-channel bars, and was regarded as a
typical meandering braided river. Bank erosion was fast because
of quite high flood peaks, frequent and rapid channel shifting, and
low silt-clay content in the bank material. After the dam was
constructed in 1959, the river bottom was scoured down and bank
erosion slowed. However, after the bed scouring the bed material
was coarser and had higher resistance than before, which caused
the bank erosion to return. In the 130 km reach immediately
downstream from the dam in the 1968 to 1981 period, 16.35
million tons of sediment was annually scoured and supplied downstream by bank erosion, accounting for 42.8 per cent of the total
sediment load of the reach. Bank erosion became a major sediment contributor in the reach.
1.4.4.2 DOWNSTREAM NAVIGATION
When a reservoir is built on a river, much of the sediment is stored
in the reservoir. The flow released from the reservoir carries much
less sediment than the natural flow, which interrupts the sediment
balance and results in scouring in downstream reaches and a
lowering of the water level. For a navigable river, this may result
in insufficient water depth during the low flow seasons. Since the
Gezhouba Dam on the Yangtze River was built in 1981, the downstream river bed has been scoured and the water level during low
flow at Yichang has been lowered by 1.05 m, reducing the water
depth downstream, approaching the channel of the Nos. 2 and 3
Navigation Locks, to only about 3 m. The designed minimum
water depth for No. 2 Lock (for barge fleets of 10 000 tons) is
4.5 m. This affects navigation on the reach.
The Rhine River is the most important navigation
channel in Europe, due to its well-balanced discharge conditions.
A number of dams and navigation locks have been constructed in
the Upper Rhine above Iffezheim, Germany, to ensure a safe and
efficient navigation channel. Erosion is often observed due to a
deficit in bed-load transport caused by the dam impoundment and
trapping of the bed-load supply from upstream reaches and tributaries. Downstream from Iffezheim to the Dutch border, some
500 km long, is a freely flowing stream regulated by groins, guide
dikes and bank revetments, so the morphological changes can only
occur in the river bottom. A careful field measurement has indicated that on the 500 km of reaches there are nine reaches with
alternating aggradation and degradation, as shown in Table 1.2.
The total deficit of bed load in the Rhine is about 350 000 tons per
year, 50 000 tons per year in the Upper and Middle Rhine, and
300 000 tons per year in the Lower Rhine. The highest bed degradation rates of 8 to 9 mm per year have been observed between
Mannheim and Mainz, and 11 mm per year aggradation has also
been observed in the mining subsidence in the Karlsruhe and
Mannheim areas. Finally, 260 000 tons per year of bed load and
dredged material have to be artificially transported by barges and
dumped back to the river to compensate the bed-load deficit
(Dröge, 1992).
On the other hand, reservoir regulation greatly changes
the flow and sediment conditions in the reservoir downstream
reaches. After the construction of the Aswan High Dam (AHD),
the flood flows of the Nile River downstream were largely eliminated. During the winter closure (December to February), a
minimum flow discharge of about 700 m3 s–1 is released for navigation. With the small sediment supply and low flow velocity, the
thalweg of the low flow channel continuously shifts on the wide
and shallow channel and multiple thalweg channels are formed.
The water depths are only 1.75 m at Selwa Bahary, 1.4 to 1.65 m
Table 1.2
Mean bed-load in the Rhine River (1981–1990)
Section
Reach
(km)
Length
(km)
Balance
(103t)
Width Bed change
(m)
(mm per
year)
1
334.0–356.0
22.0
–73
170
–11
2
356.0–426.7
70.7
+207
180
+9
3
426.7–483.5
56.8
–201
220
–9
4
483.5–528.8
45.3
+103
400
+3
5
528.8–660.1
131.3
–129
200
–3
6
660.1–703.6
43.5
+109
240
+6
7
703.6–768.0
64.4
–281
260
-9
8
768.0–800.0
32.0
+178
280
+11
9
800.0–857.5
57.5
–251
300
–8
Sum/
334.0–857.5
average
523.5
–338
230
CHAPTER 1 — ECOLOGY AND ENVIRONMENT RELATED TO SEDIMENTATION
at Rozaiquat and 1.35 to 2.1 m at Armant between Aswan and
Luxor. The water depths are not sufficient for navigation
(Gaweesh and Ahmed, 1995).
1.4.5
Case studies
The construction of reservoirs, especially large reservoirs, greatly
changes the natural river conditions and causes a number of environmental and ecological problems related to sedimentation. On
the one hand, the sediment carried by flow largely deposits in the
reservoir because of the reduction of flow velocity, and diminishes
the benefits of the reservoir. On the other hand, the flow released
from the reservoir carries much less sediment than the natural flow
and scours the downstream river channel. It may cause water
supply and navigation problems. Engineers and planners should
pay close attention to these problems in the planning and design
stages and try to find available measures or operations to mitigate
the damaging effects of the reservoirs as much as possible. Some
case studies given below are expected to provide experience in this
area.
1.4.6
Guanting Reservoir in China
Guanting Reservoir on the Yongding River in northern China
(Figure 1.5) has a storage capacity of 2.27 billion m3, consisting
of two parts. One is on the Yongding River with a capacity of
about 0.91 billion m3 (40 per cent of the total); the other is on the
Guishui River, a tributary of the Yongding, with 1.36 billion m3
(60 per cent of the total). Almost all runoff and sediment load
comes from the Yongding River. By 1998, the total sedimentation
in the reservoir had reached 0.646 billion m3, with only about
52 million m3 (9 per cent of the total deposit) in the Guishui and
more than 90 per cent in the Yongding. The reservoir sedimentation greatly reduces the functions of the reservoir for flood control
and water supply. Moreover, as the deposition delta in the
Yongding River progressed forward to the dam, a mouth bar at the
Guishui River mouth formed and rose to an elevation of 474.4 m
in 1997, making the storage capacity of 0.254 billion m3 in the
Guishui River useless.
On the other hand, the deposition has extended upward
to a point 36 km from the dam where the bed elevation reached
507 m, 29 m above the normal pool of the reservoir. It caused the
river, at the confluence of two upstream tributaries, the Sanggan
and Yang Rivers, to rise by 4.3 m, which is 1.6 m higher than the
ground levels outside of the levees. A rise of the water level in the
7
backwater region due to sediment deposition there led to a general
rise in the groundwater table in the riparian region. Contours of
equal rise in the groundwater table occurred in the triangular area
between the Sanggan and the Yang Rivers. A major part of the area
had a rise in groundwater table of 3 to 4 m, coming to within
about 1.5 m below the ground surface. This caused extensive land
salinization. In the past, the area subjected to salinization was only
533 ha, but it has increased about 14-fold, to 7 333 ha. The annual
loss in food production due to waterlogging in the reservoir region
has been estimated at 25 000 tons. With the deposition in the backwater region progressing upstream to an extent greater than
anticipated, some relocated people were again affected by the rise
in the groundwater table subsequent to the rise in river level,
resulting in waterlogging, the collapse of numerous houses and
even the formation of some marshes. The total area affected is
over 20 000 ha. Rehabilitation involves both economic and sociological problems (Zhang, Jiang and Lin, 1986).
1.4.7
Aswan High Dam
The Nile River in Africa is the second largest river in the world,
with a total river basin of 2.9 million km 2 and a length of
6 825 km. The Nile flows through nine countries: the Republic of
Tanzania, Burundi, the Democratic Republic of the Congo,
Rwanda, Kenya, Uganda, Ethiopia, Sudan and Egypt. It has
1 400 km in Egypt, where it empties into the Mediterranean Sea.
About 96 per cent of the territory of Egypt is desert, with an
annual precipitation of only a few centimetres. The population is
concentrated along the Nile and the river delta. The annual runoff
at the dam site is 84.0 billion m3, with a yearly fluctuation of 41.3
to 134 billion m3. If the yearly runoff is more than 130 to 140
billion m3, a food disaster occurs. However if it is less than 40 to
50 billion m3, it causes droughts. The annual sediment load is 316
million tons, with a sediment concentration of 3.764 g/l. Floods
like the one in 1878, with a maximum daily runoff of 1.14 billion
m3, and droughts like the one lasting nine years (1979 to 1988)
can create disastrous situations for the Egyptian people.
The Aswan High Dam (AHD) is on the Lower Nile River
in southern Egypt. The reservoir is called Lake Nasser, with a total
capacity of 168 billion m 3 . The construction of the AHD has
provided Egypt with comprehensive benefits. The water discharge
in a year ranged from 1 000 to 10 000 m3 s–1 before the dam was
constructed. After the dam was completed, the maximum water
discharge was limited to 2 500 m 3 s –1 and the sediment
Figure 1.5 — Guanting Reservoir.
8
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
concentration was reduced to between 0.03 and 0.1 g/l. The huge
storage capacity of the reservoir successfully controlled the floods
in 1968, 1975 and 1988, and met the irrigation requirements for the
nine-year drought from 1979 to 1988. Before the AHD, the river
supplied 4 billion m3 and 48 billion m3 of water to Sudan and Egypt,
respectively. The water benefit from reservoir regulation is 22 billion
m3, shared by Egypt and Sudan according to the 1959 Nile Waters
Agreement. Now, 14.5 billion m3 and 7.5 billion m3 of additional
water annually goes to Sudan and Egypt, respectively, or 18.5 billion
m3 and 55.5 billion m3 in total. Egypt’s agricultural production
increased 20-fold from 1960 to 1987, and the wheat yield rose from
1.1 million tons in 1952 to 4.5 million tons in 1991. The AHD has
also created other benefits for both Sudan and Egypt, such as power
generation of about 10 billion kWh per year (53 per cent of the total
electric power of Egypt in 1977), improvement in the navigation
conditions upstream and downstream, development of tourism, and
10 000 tons of fish annually.
The AHD also creates some ecological and environmental problems for both the upper and lower reaches. Some of them
are related to reservoir sedimentation (Said Rushdi, 1993):
(1) Residents moving. The homes of close to 400 000 Nubians
and an array of temples, tombs and fortresses were inundated
forever by the reservoir.
(2) Water loss. Water losses during the 1970 to 1986 period were
190.4 billion m3, with an annual loss of 11.2 billion m3 due
to evaporation and seepage of Lake Nasser. Evaporation in
the lake results in a 10 to 15 per cent increase in the total
dissolved solids of water, and affects the water quality.
(3) Salinity of irrigated land. The salt content in the water of the
Nile is 0.02 per cent and 0.035 per cent, at the dam site and
mouth, respectively. The annual irrigation water from the
river is about 40 billion m3, meaning about 12 million tons of
salt is added to soil and groundwater by filtration. It is estimated that about 96 kg of salt are deposited on each feddan
(about 0.42 ha) per year. Therefore, an appropriate drainage
system had to be established. During the initial irrigation
period, drainage was not given appropriate consideration. The
groundwater table rose and land salinity occurred. Since the
1970s, the Government has paid much attention to drainage,
and has taken rational measures to control salinity. By July
1992, 87 per cent of the drainage system was completed. The
salt content in soil has been controlled effectively. The situation has improved greatly, and agricultural production has
increased by 15 to 30 per cent.
(4) Decline of land fertility. In the past, the silt left by floods
provided inundated farmland with a large amount of natural
and organic fertilizer. However, the clear water released from
the reservoir is lacking in such fertilizer, and therefore the
fertility of the land has deteriorated.
(5) Degradation and channel shift of the downstream reach of
the dam. The annual sediment load at the dam site is about
134 million tons (ranging from 60 million to 180 million
tons). After the reservoir was put into operation, most of the
sediment load deposited in the reservoir. The maximum sediment concentration of the Nile at the dam site before the
reservoir construction was about 3 764 mg/l. However, after
the dam was completed it was reduced to 30 to 100 mg/l.
Scouring has occurred in the downstream reaches. At the
beginning, the scouring rate was fast, ranging from 2.2 to
3 cm per year in a 478 km-long reach downstream from the
dam, as the riverbed became coarse and the flow conditions
much more uniform than before. The river bed in most
downstream reaches was scoured by 42 to 66 cm on average
until the 1980s (Figure 1.6). The maximum local value was 2
m. The water level was lowered too, which reduced the water
head difference between upstream and downstream of some
weirs, and therefore produced safety problems. The water
depth in some reaches was not enough for navigation due to
the water level falling. Another problem caused by the scouring of clear water was shifts of the river channel and the
collapse of river banks, as shown in Figure 1.7. In the mid1980s, stable river channel conditions reached downstream
and the rates of scouring almost stopped. After river channel
protection works were constructed, the lateral shift of the
river channel was limited. On the other hand, about 300 000
feddans of old farmland deteriorated because the topsoil was
used as raw material for brick making. Before the AHD,
large amounts of sediment provided by the annual floods
were the raw material source for brick-making.
(6) Erosion of coastline. Erosion at the river mouth is found and
the delta area is threatened because the reduced incoming
sediment cannot fully supply the amount carried away by
tidal flow. The coast line at the mouth of the Rosetta draws
back about 150 m per year. Sand losses are in the order of
200 000 tons per year west of the Rosetta mouth and
400 000 tons per year west of the Damietta mouth. The
aquifer beneath the northern reach of the delta 15 to 35 km
inland from the sea has the same salinity as the sea.
(7) About 82 per cent of irrigation and drainage canal systems
are overgrown with weeds and grass, which increases the
roughness of the canal system, reduces the flow conveyance
capacity of the system, impacts the navigation conditions,
and increases the loss due to evaporation. Moreover, weeds
and grass provide a habitat for some vehicles of diseases.
The Government and people of Egypt have made great
efforts to control and eliminate the negative impacts of the AHD
on the ecology and environment, and the benefits of the AHD have
made the reservoir shine with the great splendour of Egypt.
1.5
UTILIZATION OF SEDIMENT RESOURCES
River sediment brings many problems, as described above.
However, it does not always cause trouble and can sometimes
even be utilized as a precious resource. Sediment eroded from
upstream basins normally contains organic manure, fertilizers and
Figure 1.6 — Change in water level below the AHD.
CHAPTER 1 — ECOLOGY AND ENVIRONMENT RELATED TO SEDIMENTATION
9
Figure 1.7 — Change of the Nile River channel between 364 and 381.75 km below the AHD.
other matter. Farmland irrigated by water with sediment may have
higher production levels because of fertility in the sediment.
Sediment may also be diverted to warp and improve lowlands. On
the Lower Yellow River, by 1990 230 000 ha of lowland had been
developed into highly productive farmland by warping, including
120 000 ha as paddy fields.
The sediment may also be used as construction material
for earth embankments and dikes for flood control. It is a good
local material, with the advantages of low costs, short
transportation, and convenience. In some developing countries, the
sediment dredged from rivers, lakes or reservoirs is used to make
bricks.
REFERENCES
Central Water Commission, 1996: Experience in sedimentation of
Indian reservoirs and current scenario. Proceedings of the
International Conference on Reservoir Sedimentation, 1996,
Volume 1, pp. 53-72.
Dröge, B., 1992: Changes of river morphology by controlled
erosion and deposition-bed load budget of the River Rhine.
Proceedings of the Fifth International Symposium on River
Sedimentation, Karlsruhe.
Egyptian Committee on Large Dams, 1993: Aswan High Dam: A
vital achievement fully controlled, Volume 11, Cairo.
Gaweesh, M.T.K., and A.F. Ahmed, 1995: Navigation difficulties
under controlled flow conditions on the Nile River. The
Hydraulics of Water Resources and their Development,
HYDRA 2000, Twenty-sixth IAHR Congress, Volume 4
pp. 30-35.
Hu, Yisan (ed.) 1996: Flood Control of the Yellow River. Yellow
River Press (in Chinese).
Julien, P.Y., 1994: Erosion and Sedimentation. Cambridge
University Press.
Rooseboom, A., 1992: River sediment problems in South Africa.
Proceedings of the Fifth International Symposium on River
Sedimentation, Karlsruhe.
Sabo Technical Centre (STC), 1995: Sabo in Indonesia. Ministry
of Public Works, Indonesia, JICA.
Said Rushdi, 1993: The River Nile: Geology, Hydrology and
Utilization. Pergamon Press.
Wu, Chian Min, et al., 1996: International Handbook on Reservoir
Sedimentation. Proceedings of the International Conference
on Reservoir River Sedimentation, pp. 571-612.
Yang Xiaoqing and Zhang Shiqi, 1998: Fluvial Processes of the
Yellow River Delta. International Workshop on Aspects and
Impacts of a Changing Sediment Regime, Bangkok,
Thailand, 16-20 November 1998, p. 149.
Zhang Qishun, Jiang Naisen and Lin Bingnan, 1986:
Environmental problems associated with sediment deposition in Guanting Reservoir. International Journal of
Sediment Research, Volume 1, Number 1, August 1986,
pp. 67–78.
Zhu Zhenda, 1992: Desertification Disasters Prevention and
Control Methods in China. Hubei Kexue Press (in Chinese).
CHAPTER 2
SOIL EROSION
2.1
INTRODUCTION
Erosion is a process in which earth or rock material is loosened or
dissolved and removed from any part of the Earth’s surface, and is
often differentiated according to the eroding agent (wind, water,
rain-splash) and the source (short, gully, rill, etc.). Soil loss is
defined as the quantity of soil actually removed by erosion from a
small area (Piest and Miller, 1975). Whereas weathering involves
only the breakdown of rock, erosion additionally entails the
detachment and transport of weathered material from one location
to another, denuding the Earth’s surface and delivering sediment
to the fluvial system by exogenous and geological forces.
Exogenous forces include solar radiation, rain and micro-organic
activities, especially the aspects of water, ice and wind, and with
humans as a significant, anthropogenic factor. Geological force is
a reference to the Earth’s crustal movement caused by geological
tectonic movement. According to the agents causing and affecting
the erosion process, erosion can be classified into two major types,
natural (normal) erosion and accelerated (abnormal) erosion.
2.2
NATURAL EROSION
In its broadest sense, natural erosion is a process which refers to
erosion that occurs under normal conditions and which leads to
the formation of a normal soil profile of the Earth in its natural
environment without human interference. The rate of this erosion
is less than that of genetic soil forming. It is caused mainly by
natural exogenous forces such as water, gravity, wind, temperature
variation and glaciers. Natural erosion includes geological erosion.
Erosion caused by geological factors is defined as the
erosion of the Earth’s surface under natural or undisturbed conditions (Gottschal, 1975). This process has been occurring since
continents emerged from the sea, and includes soil formation as
well as erosion processes. The rate of this erosion, combined with
the complex processes of soil formation, largely determines the
type and distribution of soil on the Earth’s surface.
2.2.1
Water
Erosion can be caused by the kinetic energy of raindrops impinging on the soil surface and by the mechanical force of surface
runoff. Surface runoff is caused by heavy rainfall and snow water
from spring thaw in the natural or artificial hydrographic network.
Erosion caused by water is the most common, widespread and
harmful type of soil erosion in the world. The main categories of
this type of erosion are surface erosion and channel (or gully)
erosion.
(1) Surface erosion. Surface erosion is caused by
precipitation and surface runoff. Soil particles are first detached by
raindrops, then carried down a sloping surface. Surface erosion is
a feature of splash erosion, sheet erosion and rill erosion. The rate
of this type of erosion is determined by slope gradient, kinetic
energy of raindrops, direction of splash, shear stress among soil
particles and soil structure.
Splash erosion: Splash erosion refers to the destruction of
the Earth’s surface by raindrops. Soil particles which are detached
and displaced from the soil surface by raindrops are carried and
gathered up by runoff to form a thin mud flow on the land surface,
moving from upper parts to lower parts of slopes. This leads to soil
erosion during the process of rainfall. Splash erosion destroys soil
structure and blocks the porosity of soil; as a result, it creates the
conditions to form runoff on slopes, since rain water cannot permeate the soil. Experiments have shown that on moderate slopes, 90
per cent of the erosion is caused by splash. Runoff scouring can
play a key role only when the land slope is 9°.
Sheet erosion: Sheet erosion is the weathering away of a
thin layer of land surface, and is caused by runoff, which is
distributed over the land surface with relatively lower velocities.
Sheet erosion generally occurs on gentle slopes close to mountain
ridges. Sheet erosion more or less removes a thin layer or sheet of
soil from a gentle sloping land or watershed. It is a rather inconspicuous type of erosion because the total amount removed in a
storm is usually small. However, over a period of years, the
amount of eroded sediment can become significant. Sheet erosion
involves two processes. First, soil particles are detached from the
body of the soil by raindrops. Second, the particles are transported
from their original location by surface runoff, which is formed
when the rate of rainfall exceeds the infiltration rate of water into
soil and water starts to flow over the surface of sloping land. At
this point, the second erosion transport process takes place. The
flowing water picks up the raindrop-detached particles and carries
them along. The action of sheet erosion causes the soil mantle to
thin, and finally the underlying rock and mineral substrata are laid
bare over a large area.
Rill erosion: Rill erosion is the process of a thin layer of
surface flow accumulating and concentrating in depressions to
form rills. In rill erosion, detachment is caused primarily by the
energy of flowing water. According to field measurements of a rill,
when the land slope is 5.7 to 40 per cent and the rain intensity is
32 to 117 mm per hour, the water depth and runoff velocity are
0.28 to 0.99 mm and 5.4 to 32 cm s–1, respectively, and the rill
width is less than 20 cm. The rill depth is over the cultivation layer
and the rill is easily removed by normal tillage operations. There
is no sharp break marking the end of sheet erosion and the beginning of rill erosion. Rills form as soon as surface flow begins. The
number of rills that develop in a given area can vary widely,
depending mainly on the irregularity of the soil surface and the
amount and velocity of runoff. Detachment and transport of soil
particles are greater in rill erosion than in sheet erosion. This is
due to acceleration of the water velocity as it concentrates and
moves in rills.
(2) Channel erosion. Channel erosion cuts deeply into
the soil when ordinary tillage tools cannot smooth the ground. It
often follows sheet and rill erosion. It occurs on the steeper sloping
land, either where runoff from a slope increases sufficiently in
volume or velocity to cut deep incisions, or where the concentrated
water flows long enough in the same channel. Gullies may develop
from rills which are allowed to go unchecked. Often, they develop
in natural depressions of the land surface where runoff water accumulates. The rate and extent of gully development is closely related
to the amount and velocity of runoff water. Gully depth ranges
CHAPTER 2 — SOIL EROSION
between 30 cm and 2 to 3 m in general, and may sometimes even
reach several dozen metres. Gullies have large dimensions, and
their development is more complicated. The forms of erosion
include retrograde or backward erosion, vertical erosion and lateral
erosion, together with accompanying landslides and mudflow, etc.
Gullies may grow into gorges and canyons, which are usually
moulded by watercourse erosion.
Channel erosion can be subdivided into shallow gully
erosion, gully erosion, gulch or canyon erosion and watercourse
erosion.
Shallow gully erosion: Shallow gully erosion mainly
occurs on relatively steep slopes, and is the result of the further
development of many rills with concentrations of sufficient runoff.
The depth of gullies is generally between 0.5 and 1.0 m and the
width is in excess of their depth, forming shallow cross-sections.
Shallow erosion develops to form gully heads and drops, which
are the main features of gully erosion.
Gully erosion: Through the accumulation of large quantities of runoff coming from rills and shallow gullies, or through
the gradual deepening of rills, gully erosion of various sizes and
forms comes into being. The first form includes any gully with a
depth of between 30 cm and 2 to 3 m. In this form, typical wash
prevails over marked backward or retrogressive erosion and vertical or depth erosion, the erosion curve being compensated by
waterfall erosion. Besides retrogressive and vertical erosion,
lateral erosion also appears here, together with accessory landslides, soil flow and other phenomena. According to the forms of
erosion gullies viewed in cross-section, flat, narrow, broad and
round gullies are distinguishable. Flat forms occur mostly on
shallow soil, or in connection with a specific lithic structure of
slope. In this form, characterized by a V-shaped cross-section,
lateral erosion prevails over vertical erosion. Narrow acute forms
are created with a narrow V-section, the breadth of the gully
usually being equal to or smaller than its depth.
Gulch erosion: Gulch gullies have a wide bottom and are
U-shaped. Here, lateral erosion prevails over depth erosion; active
gullies maintain steep or even perpendicular sides (Zachar, 1982).
With concentrated runoff cutting the gully bed, retrogressive or
headward erosion, gully bed erosion and lateral erosion are active.
As runoff discharge increases, gully erosion develops rapidly by
vertical erosion, retrogressive and lateral erosion to make a
U-shaped cross-section. The slope of the gulch bed is distinguished from the original land surface. Its slope upstream of the
gully bed is steeper than downstream. Vertical erosion decreases,
and retrogressive erosion and lateral erosion collapse are active.
Gulch development depends on large quantities of water to supply
energy for both detaching and transporting the soil. Drops can be
found at the gully heads, where the retrogressive erosion will start
with the next rainfall. The retrogressive erosion causes the drop
head gradually to increase; as a result, collapse of the lateral slope
takes place due to vertical erosion of the gully bed.
Watercourse erosion: Watercourse or river erosion
occurs where there is a permanent water flow, and usually shows a
varying intensity as the flow varies. The smaller the catchment
area of the watercourse, and the less favourable the conditions of
discharge, the greater the fluctuation of erosion intensity. The
uppermost branches resemble gullies and therefore constitute a
transition between river and gullies. The boundary line between
the hydrographic network and gullies remains arbitrary, especially
in semi-arid and arid regions. According to the prevailing direction
11
of influence, a distinction can be made between vertical or bottom
erosion, which deepens on the profile and compensates the erosion
curve; lateral erosion, which broadens the river bed and may cause
a change in the flow direction; and retrogressive or retrograde
erosion. From this point of view, gully and river erosion are
similar, but river erosion changes the surface of the watercourse
only to a small extent and damages only soil. In general, by lateral
movement of the river course as it meanders, the area covered by
gullies may considerably increase at the expense of agricultural
land. In gully erosion, the typical action is retrogressive erosion; in
river erosion, it is lateral erosion. In this connection, it is possible
to speak of river erosion of the soil occurring along banks and
during flood conditions. Under the influence of this process,
various kinds of undermining action may occur together with slips
and rifts of banks and slopes. During floods, surface wash, gullies,
hollows and other forms may also occur (Zachar, 1982).
(3) Gravitational erosion. Gravitational erosion is, as
the name imples, caused mainly by gravitational agents. Its main
characteristic is the transport of surface materials as part of a
joint action with other exogenous agents, especially water
erosion and infiltrated water. The stability of the earth on the
steep slope is maintained by internal soil friction and cohesion,
as well as protection of vegetation. This internal friction and
cohesive force is decreased when influenced by exogenous
agents such as vegetation depletion or raindrop splashing.
Consequently, under the influence of gravity, soil and parent
materials begin to move. Gravitational erosion includes
avalanches, landslides, debris slides, cave and hole erosion and
various kinds of mudflows and debris flows.
Avalanches: Avalanches are a phenomenon of the sudden
collapsing, rolling and dropping of rock and earth when they are
separated by cracks. Avalanches usually occur in high mountainous areas with steep side slopes, especially in areas of severe river
erosion.
Landslides: Landslides are primarily caused by gravitational forces, the result of shear failures along the boundary of the
moving mass of soil or rock. However, owing to progressive
failure, landslides can occur at an average shear stress considerably less than the peak strength of the soil or rock. Landslides
generally occur on slopes of 12 to 32°. Within this range, the
larger the slope gradient, the higher the possibility that gravitational force exceeds resistance to movement. Landslides usually
occur in strongly weathered rock, and have close relationships
with faults or shattered zones. Abnormally high water tables along
a fault often cause landslides. A small-scale shattered zone around
intrusive rock, which forms a good conduit of groundwater, can
also trigger them.
Debris slides: Debris slides are a phenomenon in which
crushed materials, weathered from rocks and earth on steep slopes
and cliffs, slide downward along the slope under the pull of
gravity. On steep slopes, soil and rocks are affected by cold, heat,
dryness and humidity. The alternate action of freezing and thawing
will thus cause a decrease of cohesive force and loosening of soil
and rock surfaces. Unstable crushed materials with parent rocks
will be formed. These crushed materials will go downward under
the action of gravitation during the rainy season.
Cave and hole erosion: Sinkholes, loess caves and
natural loess bridge erosion are forms of erosion in the loess
regions. Loess soil is typically loose, porous, homogeneous and
easy to cultivate, and causes erosion. Surface runoff permeates
12
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
underground along vertical cracks of the loess soil, and thus
results in underground dissolving and washing, and the binding of
dissolved elements and small particles to deep layers. With consequent weakening of the ceiling, the stability of the overlying
layers is impaired. The final stage creates corridors and caves or
sinkholes.
Debris flows: Debris flows are mixed flows of rock, soil,
water and air between sediment-laden water flows and landslides.
The occurrence of debris flows is mainly connected to the
geomorphologic conditions of a certain gradient in mountain areas
where, under certain water content conditions, large quantities of
unstable, loose rocks cause rock, soil, water and air materials to
start, collect, mingle and move. Debris flows, the product of
degraded mountain environment, are one of the main hazards with
the most unexpected consequences among the numerous natural
disasters in mountainous areas. They are closely related to a
certain topography, geomorphology, geological structure and geotectonic movement, as well as hydrological and climatic
conditions. Usually, areas of debris flow development have the
characteristics of complicated geological structures, soft and loose
rock formations, developed joints and fissures, and active
collapses and landslides. The entrapment of heavy rains, glacial
and snow melt-waters or rivers or lakes all can trigger the occurrence of debris flows.
2.2.2
Wind
Erosion caused by wind is a process of detachment, transport and
deposition of soil and sand particles due to air current. This type
of erosion occurs mainly in those areas where there is a lack of
precipitation together with predominantly high temperatures, i.e.
arid regions. Wind affects the soil by desiccating the surface layers
and drying up and removing soil particles by deflation. The
stronger the wind, the greater its influence on soil.
2.2.3
Freeze-thaw
Erosion caused by freeze-thaw is a process of mechanical abrasion
of soil caused by temperature changes, which occurs predominantly in cold regions where the average temperature is below 0°C.
2.2.4
Living organisms
Soil erosion can be caused by living organisms, through phylogenetic and zoological processes. Phylogenetic processes include
soil destruction by roots. Zoological processes occur when
animals destroy the soil when searching for food, moving or excavating their hiding places on the surface and under the ground.
2.3
ACCELERATED EROSION
Accelerated erosion is defined as the increased rate of erosion over
the normal or geologic erosion, brought about by human activities,
such as deforestation, indiscreet reclamation cultivation, overgrazing for food fibre and meat, and development of industries. This
accelerates normal soil erosion rates caused by water, wind,
temperature and gravitational force, etc. The accelerated erosion is
in excess of the natural erosion which has brought changes in
natural cover and soil conditions. The accelerated erosion rate is
higher than the rate of soil formation, which causes a restructuring
of the Earth’s surface by the wash of soil particles and nutrients
which can no longer be resupplied by the soil formation process.
The unfavourable consequences of industrialization and urbanization processes pose a threat not only to soil, but also to water.
2.4
FACTORS AFFECTING SOIL EROSION
Erosion is initiated by natural forces and can be intensified by
human activities. The erosion process is controlled by the action
and interaction of many factors. The factors affecting soil erosion
may be grouped into two categories: natural factors and human
activities. The most prominent natural factors include meteorology, geology, topography, composition of earth surface and
vegetation cover. Human activities play both positive and negative
roles in soil erosion, and are the major factors causing modern
acceleration erosion. The positive ones include various measures
of soil erosion control and proper comprehensive watershed
management. The negative ones include poor or improper land
use, reclamation, construction and urbanization.
2.4.1
Meteorology
Meteorological factors affecting soil erosion are precipitation,
wind and snowmelt.
Precipitation: Precipitation includes rainfall, snow, hail
and many other types. Rainfall, especially rainstorms, is the main
factor affecting soil erosion. Main elements of rainfall include
amount, intensity, duration, spectrum of raindrop and falling
velocity. The most significant characteristic value of rainfall is the
kinetic energy of raindrops impacting the soil surface.
Amount of rainfall: In general, soil erosion will increase
to a point with an increase in the amount of rainfall. However, this
is not the only factor. Rainfall intensity and the spectrum of raindrops, etc. also determine the amount of soil erosion a storm
causes. A storm with an intensity of less than 10 mm/h, the erosive
threshold value, will not result in soil erosion.
Raindrops: Raindrop characteristics include form, size,
velocity of falling drops and terminal velocity. In general, small
drops are in the shape of a circle, and larger drops are oblate. The
diameter of drops ranges from 0.2 to 7 mm.
Mutchler and Young (1975) studied the process of soil
splash erosion by raindrops and found that when the water layer
on the land surface was thinner than one fifth the diameter of a
raindrop, the raindrop had strong erodibility. However, it was also
determined that when the water layer exceeded three times the
diameter of the raindrop, the erodibility was greatly weakened
(Jansson, 1982). The relationship between rain intensity, kinetic
energy and erosive force of rain is of most importance for rain
erosion. Low intensity rain is mainly composed of small drops,
while high intensity rain has at least some much larger drops. The
formulae to calculate the medium diameter of raindrops are as
follows (Zhu, 1992):
Laws and Parson’s equation:
d50 = 2.23I 0.182
(2.1)
Hudson (1981):
d50 = 1.63 + 1.33I – 0.33I2 + 0.02I3
(2.2)
where d50 is the medium diameter of raindrop in mm, and I is the
rain intensity in mm/h.
The relationships between rain intensity and kinetic
energy are:
Zhong’s equation:
E = 23.49I0.29
(2.3)
CHAPTER 2 — SOIL EROSION
Wischmeier and Smith’s equation:
E = 210 + 89 log I
(2.4)
Kinnell’s equation:
E = 29.82 [1 – eρ (0.044π0.214)]
(2.5)
Studies have shown that EI30 or EI15, the product of
kinetic energy of rainfall and the maximum rain intensity in 30 or
15 minutes, is an appropriate parameter to estimate soil loss.
Snow and glacier: A solid form of precipitation significant for erosion is snowfall, because in the spring when snow
thaws it may cause surface runoff and soil erosion. The runoff
from snowfall is dependent on the physical properties and depth of
the snow distribution of the snow cover and on thaw processes.
A glacier is a mass of ice predominant in cold regions
where the annual mean temperature is below 0°C. A specific
feature of glacial erosion is the action of a large mass of ice
moving slowly. Furrowing, cutting, ploughing and scouring are the
most pronounced forms of glacial erosion.
13
depth of runoff, erodibility of soil and roughness of ground
surface, and α is the slope angle. As the slope angle increases to
56.8°, the value of (sin α / tan 0.3 α ) reaches the maximum of
0.737 (Jansson, 1982).
According to the analysis of observation data by the
runoff plots at the Suide and Lishi soil conservation experimental
stations, in gully hilly loess areas the turning gradient is generally
25 to 28° (Chen, et al., 1988).
(2) Slope length. There are different views on the
impacts of slope length on soil erosion. Rose suggested that soil
erosion decreases with an increase in slope length because long
slopes increase the sediment concentration of flow and therefore
more energy is consumed in sediment transport and less soil is
eroded. For flat slopes, erosion is not closely related to slope
length; for steep slopes, erosion is in proportion to slope length.
Some formulae expressing the relationships between soil
erosion and slope length follow (Zhu, 1992):
Zingg’s formula:
E = AL1.6
(2.7)
Kernev’s formula:
2.4.2
Geology
The character of bedrock and tectonic movement has significant
effects on soil erosion.
Rocks susceptible to weathering often suffer strong
erosion. Soils weathered on limestone and dolomite formations
are relatively resistant; those on igneous rocks are less so; and
those on various sediments such as sandstone, loam, clay, chalk,
flysch formations and loess sediments are least resistant.
New tectonic movement is the most significant cause of
erosion changes, affecting the degree of erosion as well as the
speed of gully development. Earthquakes are quick tectonic movements that loosen surface materials and produce landslides or
collapses, therefore greatly increasing soil erosion.
Volcanos are geological actions releasing large amounts
of loose volcanic material. They not only change topography, but
also plug mountain areas and cause serious debris flows during
high-intensity rainstorms, increasing erosion rates greatly.
Volcanic rock is easily eroded.
2.4.3
Topography
Topography is the basic factor constituting the natural environment. Erosion is closely related to the types and characteristics of
topography. Topographical characteristics include the gradient,
length and direction of slope, which affect erosion through the
intensity of runoff formed on it.
(1) Slope gradient. The relationship between gradient
and erosive intensity are shown in Table 2.1.
Soil erosion increases first with an increase of gradient,
but when the gradient reaches a certain value, erosion no longer
increases with the increase of gradient. The turning value of the
gradient is called the critical gradient.
L < 50 m
L = 50–200 m
e=K
–
xc0.6)
sin α / tan
0.3
α
(2.6)
where e is the depth of eroded soil per unit of time, x the distance
from the slope top, xc is the critical distance from the slope top
where no erosion occurs, K is the coefficient depending on the
(kg s–1)
(kg s–1)
(2.8)
(2.9)
where E is the mass of eroded soil (t), Rck is the erosion rate in
kg s–1, L is the slope length in m, i or θ is the slope angle, and M
is the rain intensity in mm min–1.
(3) Slope shape and direction. Slope shapes can be
divided into straight, convex, concave and compound types. The
straight slope has an approximately constant slope gradient
throughout; the maximum runoff at highest velocity is concentrated on the lower part, and erosion intensity is higher on the
upper part. The gradient of convex slopes increases along the
slope length and the flow disperses down the slope. Convex slopes
have the highest intensity of soil erosion. Concave slopes flatten
out toward the bottom of the slope and sediment carried in runoff
water settles as flow velocity decreases. Compound slopes have
combinations of different slopes.
Table 2.1
Relationship between gradient and soil erosion
Authors
Formulae
Musgrave, 1947
Hudson and Juckon, 1971
1.49
E ∝ Sa
Kilinc and Richardson, 1973
2.0
1.66
Smith and White, 1947
E ∝ b + cSa
Meyer and Monke, 1965
E ∝ (S –
Sc)a
Wischimeir, et al., 1958
E ∝ (0.43 + 0.3S + 0.043S2)
Liu (from Chen, 1988)
d = 0.012S1.4 + 0.56
Chen, et al., 1988
Coefficient a
1.35
Zingg, 1940
Horton’s formula:
(x0.6
Rck = Ai0.75 M1.5 L1.5
Rck = Ai0.75 M1.5 L1.5
1.33
2.0–2.5
h = 3.47 × 10–3 I 2.16 + 0.57
h = 3.98 × 10–4 I 2.44 + 0.2
h = 3.16 × 10–7 I 5.35 + 10.5
h = 3.02 × I 3.18 + 0.55
NOTE: E, W – eroded soil amount (t km–2); h – scouring depth (mm); I, S – gradient
(degree).
14
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
Slope direction in the sunlight or in the shade influences
the soil moisture and temperature. Slope exposure to solar radiation on southern and western slopes causes the rapid thaw of snow
resulting from differences in day and night temperatures.
Consequently, this results in higher surface runoff from the snow
thaw, increasing the intensity of soil erosion.
2.4.4
Soil characteristics
Erosion is affected by soil characteristics such as infiltration,
detachability by raindrops and runoff and susceptibility to rill,
gully and channel erosion.
(1) Texture. Erosion due to raindrops is affected by soil
texture. Ekern obtained the relationship between soil particle size
and relative amount of splashed soil material, as shown in
Table 2.2. Jansson demonstrated that filtration is a function of soil
texture, as shown in Table 2.3.
(2) Structure. The contents of clay, organic matter,
calcium, magnesium, and free iron oxide contribute to soil aggregation. Aggregation increases the number of large pores and thus
increases the soil infiltration rate and reduces runoff. But disintegration of soil structure may be caused by both mechanical and
natural agents. Dryness, wetness and freezing-thawing are important in soil disintegration processes.
Soil profile: In areas where the soil consists of layers of
different textures, the water erosion resistance of the areas is
affected by stratification of the layers. Where a permeable layer
rests on an impermeable layer, it may become oversaturated with
water which the lower layer is unable to absorb. This leads to an
intensive wash of the permeable layer.
Content of soil moisture: Water always moves under
tension in well-drained soil, and an increase of soil moisture
makes soil less porous. This is the reason that soil moisture affects
surface runoff.
Soil porosity: Water moves more readily through porous
soil than dense soil. A dense layer near the surface slows water
movement because of low porosity. The moisture content of the
overlying layer soon begins to increase.
Aggregation and surface sealing: Aggregate formation is
dependent on organic matter and the types of bases in the soil.
Black soil with 40 per cent aggregates has two to four times the
Table 2.2
Relationship between particle size and relative amount of
splashed material
Grain size (mm)
0.84–0.59
0.42–0.25
0.25–0.175
0.10–0.05
0.05–0.002
Relative amount of splashed
material during 5 min (%)
30.0
77.2
100.0
61.0
21.0
infiltration rate of loose soil with less than 5 per cent aggregates,
according to an investigation conducted in the Loess Plateau
region.
Topsoil depth: Topsoil depth affects soil erodibility. Its
primary effect is on infiltration. Topsoil allows water infiltration to
proceed unrestricted for a time until layers of different porosity
are reached. Its second effect is on the organic matter content of
the surface. If the topsoil is thin and subsoil is ploughed up and
mixed with it, the organic matter content is lowered. This results
in lower aggregate stability and higher erosion. The third effect is
on the general fertility of the soil. The deeper the topsoil, the
greater the nitrogen release and, as a consequence, the greater the
vegetative cover produced. Erosion losses are less than those from
an area of shallow topsoil.
Water-holding capacity: Soil texture largely determines
water-holding capacity. Various textured soils erode differently
because of differences in infiltration, percolation and detachability. Clay, compared with sand, can hold a great deal more water,
and a high percentage of available pore space can be filled.
Water-holding capacity affects soil erosion through its influence
on detachability of soil by runoff during heavy rains. Sand is
easily detached and washes away readily under a high velocity
of runoff, and clays may seal over and be virtually impossible to
detach.
(3) Soil erodibility. Bouyoucos suggested that soil
erodibility equals (per cent of sand + per cent of silt) / (per cent of
clay). Wischmeier, et al., defined it as (per cent of silt + per cent
of very fine sand) × (100 – per cent of clay). There are many
erodibility indices in the literatures. Some are expressed in terms
of soil texture, some in soil structure, and some in water transmission and aggregation stability or dispersion (Jansson, 1982).
Dispersion rate: This is the weight ratio of sand and clay
particles dispersed during the experiment time period to the total
sand and clay.
Rate of surface aggregation: This is the ratio of surface
areas of the sediment particles larger than 0.5 mm (cm2 g–1) to the
total surface areas of aggregate silt and clay particles.
Factor K in the Universal Soil Loss Equation (USLE):
(Wischmeier, et al., 1971): The nomogram has been drawn using
five parameters, i.e. percentage of silt (0.002 to 0.05 mm) and fine
sand (0.05 to 0.10 mm) in the total, percentage of sand (0.1 to
2.0 mm) in the total, content of organic matter, structure and infiltration. By artificial rainfall, Dumas determined the K value in
USLE as:
1g 1 000 K = 3.4623 – 0.0282 X1 – 0.1695 X2 – 0.0212 X3 (2.10)
where X1 and X2 are the percentages of gravel and organic matter,
respectively, and X3 is the equivalent weight of soil moisture
retention.
Melton’s formula:
Table 2.3
Variation of soil infiltration
Soil texture
Clay loam
Silt loam
Loam
Loamy sand
Infiltration (mm/h)
2.5–5.1
7.6–12.5
12.7–25.4
25.4–50.8
E=
d z Me
(2.11)
e
where E is the anti-erosion rate of soil, dz is the soil dispersion
rate, M e is the equivalent weight of soil moisture retention,
and e is the content of soil colloid. E < 10 means high antierosion properties, and 12 < E < 115 means low anti-erosio
properties.
CHAPTER 2 — SOIL EROSION
2.4.5
Vegetation cover
Vegetation cover protects the soil surface from the direct impact
of raindrops and from the effects of wind. It enhances the
infiltration of rainfall into the soil and slows surface runoff,
thereby improving the physical, chemical and biological
properties of the soil.
2.4.6
Human activities
Soil erosion is the result of exogenic forces exceeding soil
erodibility thresholds. Natural factors are potential effects while
human activities are main factors that positively or negatively affect
erosion intensity.
(1) Destruction of vegetation cover. Population
increase has brought about and continues to bring about
extensive changes in land use. Operations that reduce vegetation
cover may induce accelerated erosion. These include cutting
trees and forest fires, etc.
(2) Cultivation. Cultivation on steep slopes may destroy
vegetation and loosen the soil, thus causing serious soil erosion.
Different crops provide different degrees of vegetation cover. As
an example, the relative erosion, C, on crop plots and bare soil in
West Africa is compared, as shown in Table 2.4. Cultivation
approaches are of great significance in erosion. Contour ploughing, strip cropping and terracing reduce erosion significantly
(Jansson, 1982).
(3) Overgrazing and burning. Overgrazing and burning
are land use practices that leave the soil unprotected. In semi-arid
and arid marginal lands, where recovery of vegetation is slow,
overgrazing causes low vegetation coverage and major erosion.
Burning of grass, bushes and trees is a practice in remote mountainous areas, where people live simple lives. Burning before
intensive rain also increases erosion tremendously.
(4) Mining, road and dam construction, urbanization.
Spoil banks resulting from strip mining, particularly in coal mines,
are often steep-sided and devoid of vegetation. Construction and
exploitation activities may produce a great amount of waste soil,
rocks and coal, which may be washed into rivers, accelerating soil
erosion. Dam construction may cause sedimentation in upstream
reaches and scouring in downstream reaches. Water withdrawal
from wells may lower the groundwater table and thus increase
gully erosion. Urban expansion involves the construction of roads,
pipes, buildings and ground paving. During landscaping and the
Table 2.4
Soil erosion on crop plots and bare soil
Type
Relative erosion C (%)
Bare soil
100
Dense forest or thick straw mulch
0.1
Savannah and grassland, no grazed crops
Late planted with slow development: 1st year
10
10
Maize, sorghum, millet
30–90
Intensive rice (second cycle)
10–20
Ground nuts
Cassava (first year)
Palms, coffee, cocoa with crops
NOTE: C – factor in USLE.
2.5
DEGREE AND INTENSITY OF SOIL EROSION
2.5.1
Soil loss tolerance
An evaluation of the seriousness of soil erosion needs to take into
account how much soil a given specific site is losing currently and
the maximum soil loss tolerable by natural resources. Soil loss
tolerance is defined as ‘the maximum rate of annual soil erosion
that may occur and still permit a high level of crop productivity to
be obtained economically and indefinitely’ (Schertz, 1983). Some
scientists have suggested that soil loss tolerance is in the range of
two to six tons per acre for various types of soil.
The soil formation rate is an important factor in determining soil loss tolerance. Under natural conditions, the formation
of one inch of soil takes 100 to 300 years, while it takes about 100
years under farming conditions. An estimate puts the renewal rate
at 0.5 tons per acre per year for unconsolidated parent material,
and much less for consolidated material. The formation of the
weathering surface layer on a base rock of granite requires 10 000
to 100 000 thousand years, while a base rock of non-granite needs
much more time (Margan, 1980).
2.5.2
Soil erosion intensity
Soil erosion intensity means that under the action of natural agents
and human activities, the soil eroded due to denudation and
displacement per unit area and unit time is expressed by the soil
erosion modulus. According to the Chinese Standard, erosion
intensity is classified as shown in Table 2.5 (Guo, 1998).
2.6
SEDIMENT YIELD IN A BASIN
Water erosion is the most important type of erosion because runoff
is essential to transport the eroded sediment. In the entire process
of erosion and transport, soil erosion, soil loss and sediment yield
in a basin are three different but closely related concepts.
Sediment yield is defined as the total sediment outflow
from a watershed or drainage basin, measurable at a cross-section
of reference in a specified period of time (Piest and Miller, 1975).
In the comprehensive planning of a medium or small watershed, if
the gross erosion and sediment delivery ratio are known, the sediment yield can be predicted.
1
Crops with rapid development
Cotton, tobacco (second cycle)
construction of urban areas, sediment yield reaches a high peak,
then declines as the land ‘heals’, and finally reaches a low, stable
value.
(5) Land use and tillage. Land use and tillage are typical
anthropogenic factors which affect erosion intensity. The intensity
of soil erosion in agricultural soil is significantly affected by the
position and shape of the plot. Observations have shown that
erosion intensity in contour farming is considerably less than that
in plots tilled downslope in straight lines.
Table 2.5
Classification standards of soil erosion intensity
30–80
2nd year
50
40–80
20–80
10–30
15
Degree
Mean annual erosion
modulus (t km–2.a)
Mean lost
thickness (mm/a)
Slight
< 200, 500, 1 000
< 0.15, 0.37, 0.74
Light
200, 500, 1 000–2 500
0.15, 0.37, 0.74–1.9
Moderate
2 500–5 000
1.9–3.7
Intensive
5 000–8 000
3.7–5.9
Utterly intensive
8 000–15 000
5.9–11.1
>15 000
>11.1
Severe
16
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
In estimating the gross erosion in a basin, if sheet
erosion plays the main role, the Musgrave Equation or USLE can
be used. But if gully erosion is the main type, it should be the sum
of slope, gully and river channel erosion. The sediment delivery
ratio on the first sub-area of the hilly loess gully of China’s
Middle Yellow River can be close to 1.0, while in the Yangtze
River Basin it is only about 0.25. The value for 12 small watersheds with areas of 0.98 to 129 km2 in the Pigeon Roost Creek
region of the United States is in the range of 0.28 to 0.76.
Factors affecting sediment delivery ratio include land
use, vegetation cover, runoff discharge, sediment size, density of
channel network, relief and catchment area, etc.
2.7
MONITORING OF SOIL EROSION AND
SEDIMENT YIELD IN A BASIN
The measurement should be conducted on three levels, i.e. developing a network of measurement stations in a basin, establishing
an observation base in representative and experimental watersheds, and establishing groups of runoff plots.
2.7.1
Runoff plots and experiments in the laboratory
A runoff plot is an isolated slope used to study the mechanics of
runoff and sediment yields and the effects of soil conservation
measures. Experiments in the laboratory can be conducted by artificial rainfall to simulate natural phenomena, including debris
flows and landslides (Meng, 1996).
(1) Plots under natural conditions.
Plot size: The smallest plot is only about 1 to 2 m2, and
easy to manufacture and install, especially for the preliminary
experiments which require large numbers of plots for studying the
relative erodibility of various types of soil. In measuring runoff, a
longer plot is essential. In the United States, a runoff plot of
2 × 22 m is normally used for the study of cultivation and crop
rotation.
Flume: A flume is installed at the outlet of the plot to
measure the runoff discharge. One type of measuring flume is an
H-shaped flume designed and produced by the United States. Its
range is 0.0028 to 3.08 m3 s–1. The Parshall flume is another
popular tool for measuring the flow discharge on a plot (Hudson,
1981).
Tank and divisor: On a small plot, the runoff is led first
into a collecting tank. But on a large one, a divisor is used to
divide the runoff accurately, so as to reduce the tank size. The
Geib divisor with a number of similar rectangular slots is widely
used in the United States. Only the flow passing through the
middle slot is collected and measured. The total flow discharge
can then be calculated by the calibrated ratio. A highly successful
moving sampler is the Coshocton revolving wheel sampler. This is
installed under the discharge from a flume, such as an ‘H’ flume,
with the water force turning a wheel mounted on a vertical axis. A
narrow slot in the wheel passes the flow on each revolution, and a
sample is taken.
In China, more than 500 runoff plots of various sizes
have been established since the 1950s. Mini-plots are generally
one to several square metres in size. They are used to study the
basic rules of runoff and soil erosion, such as splash erosion,
stability of soil aggregate, the processes of topsoil becoming
crust, soil erosion durability and so on. Common plots are generally 5 × 20 m. They can be used to study the whole progress of
rill and inter-rill erosion. Normal cultivation and relevant
measurements can be carried out on them. The natural large size
plot is a small natural catchment with an area of several hectares
and includes rills, shallow gullies and even cutting gullies with
farmland, wasteland and forest. They can be applied to study the
transportation of runoff and sediment and the equilibrium of
sedimentation.
(2) Rainfall simulator. The rainfall simulator has two
important advantages. It is not restricted by the existence of
natural rainfall and it can repeat heavy rainstorms to obtain
desired results. In 1950, the application of a rainfall simulator
found that erodibility is linked to the kinetic energy of raindrops.
There are different rainfall simulators such as the non-pressurized
dropper and spray simulator. China has developed an experimental
rainfall simulation device system in the laboratory, and a large
slope surface simulator device in the field, as well as a portable
small rainfall simulator.
2.7.2
Measurements of soil and water losses on pilot
watersheds
The measurement of soil and water losses on representative or
pilot watersheds started in the 1930s, and there were nearly 1 000
such watersheds in the world by 1974.
Representative watershed: This should be a natural
watershed with an area of 10 to 250 km2 in general. Its purpose is
to study the mechanics of runoff and sediment yields, to explain
the essences of physical processes of various factors and to
develop a mathematical model of sediment yield in a basin.
Experimental watersheds: These are usually coupled
watersheds used for comparison tests. Two watersheds should
have similar topography, relief, soil and vegetation and the area
should not exceed 4 km 2 in general. A calibration period is
required before any soil conservation works are done.
Subsequently, soil conservation measures are conducted on one
watershed while the other is kept under natural conditions (Zhu,
1992).
Runoff plots are not able to reflect the runoff and soil
loss of the whole watershed. It is necessary to establish some
monitoring stations and conduct measurements simultaneously.
These measurements include soil moisture, groundwater table,
scour and deposit in river channels, sedimentation in small reservoirs, pools and check dams, as well as the discharge and
sediment concentration at the outlet station of the watershed.
The relevant samplers used for the measurements in
watersheds are as follows.
Pumpable automatic sampler: This can take samples
intermittently and put them in order into bottles. This sampler is
quite useful for monitoring sediment delivery in small rivers. It is
controlled by a transducer of water level (Walling, 1981).
C-type wheel sampler: This was developed in 1947 by
Pomerence and applied to the watersheds near Coshocton, Ohio,
United States. Parsons later calibrated and improved it. This
sampler has been used in combination with the ‘H’ flume to take
an equal volume sample intermittently.
2.7.3
Measurement method with Cs-137
The spatial distribution of soil erosion is essential for the study of
long-term soil erosion. The Cs-137 method is useful for this. The
radioactive micro-particles are the result of a nuclear test in 1954.
Cs-137 has a long half-life so there are many Cs-137 samples still
preserved in the topsoil layer. Cs-137 can be absorbed intensively
CHAPTER 2 — SOIL EROSION
by topsoil and transported, with some soil loss. If the accumulated
Cs-137 in a region is measured, then the amount of soil loss or
deposition can be estimated.
The Cs-137 method has been extensively used to
measure the age of depositions by taking rock-core samples in
lakes. It is also a perfect tool to determine the location, size and
duration of a deposition belt in a river system. For example, in
Maluna Creek, New South Wales, Australia, the deposition rate of
the alluvial fan was 4 cm per year as determined by the Cs-137
method (Walling, 1981).
2.7.4
Dynamic monitoring by remote sensing and GIS
(1) Dynamic monitoring by aerial photography. Aerial
photography is a proper means to illustrate and evaluate the topography, soil, climate and influence of best management practices
(BMPs) on erosion. Aerial photography is easy to obtain. Using a
35 mm camera with a 135 mm lens at about 300 m above the
slope, photographs can be enlarged to 1:2000-scale negatives. This
scale can provide a clear picture of rill erosion. Erosion rates can
be measured accurately using a sequence of time-lapse, low-altitude aerial photographs and photogrammetric procedures (Frazier,
et al., 1983).
(2) Investigation of soil erosion by space remote
sensing. Satellite remote sensing imaging can provide information
on various factors affecting soil erosion on the ground surface,
such as the relief, topography, constitutive material on the ground,
vegetation cover and land use. Based on the image vein of the
structure, tone, picture, geometry and topography obtained by the
satellite image, the relationship between the main factors can be
analysed and calibrated in the sample plot. Then the pattern, intensity and level of soil erosion can be obtained (Yellow River
Conservancy Commission (YRCC), 1991).
(3) A geographic information system (GIS). By using a
GIS, planners can establish the correlation of land cover and
topography with runoff, drainage area and terrain configurations
obtained in various environmental conditions. This approach
enables water quality data from various sources to be integrated
into a comprehensive system capable of combining and referencing such diverse data elements as conventional map information,
Landsat imagery and tabular data obtained on the ground.
Technologies of the 1980s, including remote sensing and
GIS, are attractive because of their capabilities for analysing data
of large and small areas, integrating numerous variables into the
evaluation processes, and easily updating databases (Walsh, 1985).
2.8
PREDICTION OF SOIL EROSION AND
SEDIMENT YIELD
2.8.1
Prediction of soil erosion
Scientific planning and land treatment for soil and water conservation require relationships between erosion-causing factors and
those that help to reduce soil loss. USLE, usually with some modifications, is the frequent basis for determining the quantity of soil
that detaches from each small area of a watershed (Foster and
Wischmeier, 1973).
2.8.2
Prediction of sediment yield
Sedimentation is the consequence of a complex natural process
involving soil detachment, entrainment, transport and deposition.
Sediment yield is the amount of sediment transported from a
drainage basin. It is a portion of gross erosion (the sum of all
17
erosion in the watershed). Sediment sources include upland sheetrill erosion, gullies, river banks, channels, construction sites, spoil
banks and roadsides. Sediment yield from upland sheet-rill
erosion sources is usually greater than that from other sources
(American Society of Agricultural Engineers (ASAE), 1977).
Sediment yield prediction is needed for many specific
purposes. Simulations are used to extend short-term sampling
programmes to compose adequate databases. This is frequently
done to predict sediment storage requirements for the design of
flood control structures. Models are used to predict the future watershed response to various land-use alternatives. This is an integral
part of evaluation of the effectiveness of alternative plans in a basin.
Another concern is related to research, because modelling is an
ordered sequence of steps in time and space, presenting a complex
process, and information gaps can be identified (ASAE, 1977).
The specific needs for sediment yield prediction are so
varied that no single model could meet them without a great loss
of efficiency. The needs generally fall into the categories of length
of model event time, area to be simulated, and sediment sources.
Event time: In selecting or designing a model, the length
of event time should be determined. In situations where animals
and plants are affected by high concentrations of sediment and
chemicals in public waters, the storm model of sediment concentration is required. A single storm simulation is required when
information on sediment concentration throughout a storm is
needed. Longer simulation periods may be more useful in considering other problems. For example, estimating quarterly or
monthly sediment yield is desirable for determining seasonal variations of sediment yield. These determinations are required for the
selection of land use and management techniques to control sediment yield and runoff. The estimation of average annual sediment
yield is sufficient for the design of reservoirs and conservation
structures and for other concerns with sediment deposition over a
long period. Prediction of long-term sediment yield trends is
required for the planning and maintenance of channels. Channel
stability depends greatly on sediment yield from upland watersheds. These types of problems require long-term estimation.
Watershed size: Large watersheds usually need less
modelling details than small ones. Therefore, in developing
models, different sediment yield predictors are needed for different watershed sizes. The contribution of groundwater to runoff is
usually higher for a larger watershed than a small one. Sediment
sources are more variable in a large watershed (ASAE, 1977).
Sediment sources: Gullies in a watershed contribute
quite a large amount of sediment per unit area to gross erosion.
Some gullies are sand sources that contribute to the bed material
in channels. This requires bed load transport to be included in the
model.
Urban areas in watersheds present special problems
because of their high runoff rates and pollution potentials.
Conservation structures create problems in predicting sediment
yield from their drainage areas. The contributions of roadsides and
ditches often may be ignored. Channels, especially in large watersheds, may contribute significantly to sediment yield. Sand is
transported differently than fine sediment (silt and clay) and
presents special computational needs.
Sand in sediment yield often merits special attention
because its deposition causes the most damage. Gully and channel
sediment sources are especially important for downstream damage
if they contain a large percentage of sand.
18
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
Usually, the texture of surface soil is different from that
of sub-soils. This is particularly significant when chemical transport is considered. Large amounts of chemicals attach to clay,
while small amounts of chemicals attach to sand.
Methods of sediment yield prediction: At present, many
sediment yield prediction methods are available and have been
used for various purposes. In general, such methods can be
grouped into two categories: those derived from statistical analysis
(statistical equations); and deterministic models which include
empirical parametric approaches, using time variant interactions
of physical processes.
(1) Statistical equations. These are usually the equations
relating sediment yield to one or more factors on watersheds or
climate. The methods are commonly used for measuring average
sediment yields over long time periods. Long-term sediment yield
can be adequately estimated for a particular watershed, but results
cannot be extrapolated to other watersheds.
(2) Deterministic models. These models introduce
numeric values (parameters) to quantify the factors affecting
erosion, transport and deposition. These parameters can be derived
empirically or calibrated using fitting techniques. One example of
the parametric model is the Wischmeier-Smith soil loss equation.
Many sediment yield models use the equation as a basis because
of the widespread availability of the parameters for conditions in
the United States. A modified version of the Wischmeier-Smith
equation incorporates the soil detachment-soil transport concept of
Mayer and Wischmeier.
Some models use time variant interactions of physical
processes. These models are developed using theoretical dynamic
equations. They have a structure of hydrologic or hydraulic
processes, depending on the model objectives. These processes are
defined and linked mathematically using sound theoretical
approaches. In general, the basic equations are for conservation of
mass, momentum and energy. Certain flow and sediment properties must be evaluated. Among them are time distributions of flow
rate, sediment concentration, flow depth, and rates of rill and
inter-rill erosion (ASAE, 1977).
A theory of variable and non-point-source areas has
been developed. This theory can be used to explain facts such
as the hydrological response on runoff formation. This viewpoint has provided the basis to establish the idea of partial
area, that is that the runoff yield area is much smaller than the
total area.
Similarly, for most rivers, most sediment discharge
comes from only a relatively small area in the watersheds. In the
Nile River basin, the area suffering soil erosion is only 10 to
15 per cent of the total area of 2.9 million km2. In the Alberta
River basin, more than 90 per cent of sediment yield comes from
an area of less than 10 per cent of the total of 430 000 km2. In the
Yellow River basin of China, the total area of which is
750 000 km 2, 90 per cent of sediment discharge comes from
40 per cent of the total, while 75 per cent of coarse sediment
(particle size greater than 0.05 mm) comes from an area of less
than 15 per cent of the total.
In 1987, a new model Revised USLE (RUSLE) was
developed by the United States Department of Agriculture
(USDA). It has several distinguishing features: (1) Data are
processed by computer; (2) A new erosive factor R, rainfallrunoff, is introduced and its seasonal distribution is related to the
crop rotation system; (3) The factor of soil erodibility varying with
seasons, K, is introduced; (4) The factors of vegetation cover, C, is
calculated using subfactors of land use (PLU), canopy density
(CC), ground cover (SC) and ground surface roughness (SR); (5)
A factor of gradient and length of slope, LS, representing the rate
of rill erosion to inter-rill erosion and different shape of slope, is
introduced; (6) The P value of soil erosion presents the rotation
system of grass and crop land, contour ploughing and the drainage
of the soil subsurface layer.
Recently, with the development of the new model of
water erosion prediction project (WEPP) for the purpose of replacing USLE, USDA has modified the erosion predicting models
based on erosive processes. There are three types of WEPP
models, i.e. cross-section model, watershed model and net and
grid model (Chen and Fei, 1996).
In the 1980s and 1990s, China developed several parametric or conceptual models of sediment yield for sediment-laden
rivers and high sediment-yield regions, especially for the Yellow
and Yangtze Rivers.
2.8.3
USLE and RUSLE
(1) Historical review. Zingg’s equation:
X = CS1.4L1.6
or
A = CS1.4L0.6
(2.12)
where X is the total soil loss, A is the average soil loss per unit
area, C is the constant, S is the degree of slope, and L is the slope
length.
In the early 1950s, Van Doren and Bartelli proposed the
erosion equation A = f (T, S, L, P, K, I, E, R, M), where A is the
annual estimate of soil erosion, T is the measured soil loss, S is
the slope, L is the length of slope, P is the practice effectiveness,
K is the soil erodibility, I is the intensity and frequency of 30minutes rainfall, E is the previous erosion, and M is the
management level (Mayer, 1984).
By 1956, precipitation, soil loss, and related data of
more than 7 000 plot-years and 500 watershed-years had been
assembled at the National Runoff and Soil Loss Data Center in the
United States. Between 1956 and 1970, additional data of several
thousand plot-years and watershed-years were added to the
databank. The resulting USLE was introduced at a series of
regional soil loss prediction workshops from 1959 to 1962. A
complete presentation of USLE is in USDA Agricultural
Handbook 282, which was revised in 1978 (Mayer, 1984).
(2) USLE and RUSLE. USLE is a comprehensive
technique to estimate cropland erosion. It considers six major
factors affecting upland soil erosion, i.e. rainfall erosion, soil
erodibility, slope and slope length, cropping, management techniques, and measures of soil conservation. Wischmeier clarified
the term as follows: The name ‘universal soil-loss equation’ originated as a means of distinguishing this prediction model from
the highly regionalized models. However, its application is
limited to states and countries where information is available for
local evaluation of the equation’s individual factors. The uses of
USLE are tremendous. It has become a major tool for estimating
soil erosion in the United States and many other countries. As is
true for any tool, however, its use is limited to certain purposes,
and it can always be improved. The result of one such improvement is RUSLE. In RUSLE, the major factors have been
extended, and it is also used to measure the conditions of forest
land and roads, etc.
CHAPTER 2 — SOIL EROSION
Wischmeier-Smith’s equation:
A = RKLSCP
(2.13)
where A is the amount of soil loss per unit area in a specific field
in t/a, R is the factor of rainfall erodibility, K is the factor of soil
erodibility defined as the amount of soil loss under the single
number of index EI30, E is the rainfall kinetic energy, I is the 30minute maximum rainfall intensity, and LS is the factor of slope
degree and slope length:
λ 0.3 S 1.3
S > 9%
(2.14)
LS = (
) ( )
72.6
9
λ
S ≤ 9% LS = (
)
0.3
0.43 + 0.30 S + 0.043S
72.6
Ej =
Ej =
185.58
Wj ( KCPS ) j
31
Kc = 1 + 0.69 cos [(t – 2.2) 2π / 12]
m = 1.2 (sin θ)1/3
(2.15)
S = 65.41 sin2 θ + 4.56 sin θ + 0.065
Factor S,
1.5
(x j
−
1.5
x j −1 )
(lb ft–1)
(2.16)
–
x j −1 )
1.5
(kg m–1)
(2.17)
Wj = 0.5RST + 15Qjq pj1/3
(English system) (2.18)
Wj = aRST + 0.22 (1 – a) Qjq pj1/3
(metric system)
(2.19)
where Wj is the energy factor representing the combination of
rainfall energy and runoff energy, a is the coefficient (0–1), Rst is
the factor of rainstorm (EI unit of USLE), Qj is the volume of rainstorm at segment j (in or m3), and qpj is the peak value of the rate
of rainstorm runoff on slope segment j (in/h or m3/h).
The accumulated soil detachments on the whole slope
are the sum of all slope segments.
Wj ( KSCP )
185.58
1.5
xj
(lb ft–1)
(2.20)
2.8.4
Empirical regression statistical model
(1)
Slope sediment yield models
The YRCC:
MS =
51.1
C
0.15
1.2 1.5 0.26 0.48
i J
Pa
P
Wj ( KSCP )
31
1.5
xj
(kg m–1)
(2.21)
where Txj is the soil transportation per unit width on Xj.
If Txj > Ej, soil erosion will occur on the segment; if Txj
< Ej, soil deposition will occur on the segment.
The empirical modification of USLE was done by
Mutchler and Murphree. Factor Re was derived by McGregorMutchler as:
Re = 0.273 + 0.217 exp (–0.048i) – 0.413 exp (–0.072i) (2.22)
(2.25)
(2.26)
(2.27)
where M s is the modulus of slope erosion (t km –2), C is the
percentage of vegetation cover in per cent, P is the rainfall in a
rainstorm in mm, i is the average rain intensity in a rainstorm in
mm min–1, J is the gradient (5), and Pa is the percentage of soil
moisture content before the rainfall in per cent.
North-West Institute of Soil and Water Conservation,
China Academy:
Ms = 3.27 × 10–5 (EI30)1.57 J1.06
(2.28)
where E is the kinetic energy of rainfall in kg.m m–2, I30 is the
maximum 30-minute rain intensity in mm min –1 , and J is the
gradient.
Kolnev (Russian Federation):
R = ainCTI0.75L1.50i1.50
Txj =
(2.24)
For factor C, Mutchler has recommended a set of
subfactors, i.e. C1 residual products of land use; C2 the combination of residual stubble; C3 the ploughing intensity; C4 the large
roughness; C5 the influence of vegetation cover.
The recommended RUSLE is:
A = RRcKKcLSC1C2C3C4C5P
1.5
(x j
(2.23)
where Kc is double the increasing rate of average K values during
the different periods.
For factor L = (λ/22.13)m, Mutchler and Greer have
obtained m = 0.15 for a slope of 0.2 per cent based on simulated
rainfall data. For steep slopes, Wischmeier derived:
6.613
where Ej is the ability of soil detachment on slope segment j, and
Xj is the slope length of slope segment j.
Txj =
For factor Kc, recent studies show that the soil erodibility in an entire year should be a variable. According to the data
from Holly Springs, Mississippi, United States, it is:
2
where λ is the real slope length in ft, the slope length of a standard
plot being 72.6 feet, S is the real slope degree in per cent, the standard value being 9 per cent, C is the factor of vegetation cover and
management, and P is the factor of soil conservation measures.
Onstad and Foster (1975): A slope is divided into several
slope segments and the soil detachment on each segment is:
Wj ( KCPS ) j
19
(2.29)
where R is the soil loss on slope surface per 1 m width, I is the
gradient, L is the length of slope, i the rain intensity, nCT is the
runoff coefficient, and ai is the coefficient.
Caroni (Italy):
W = aImbDcHd
(2.30)
where W is the soil erosion at the outlet of the plot, Im is the
maximum rain intensity, D is the duration of rain intensity
> 10 mm/h (Dc) or total rain duration (Dt), H is the total runoff,
and a, b, c, and d are coefficients.
20
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
(2)
Prediction models on gully erosion
Beer-Johenson’s equation (data source: loess area of
serious gully erosion in the western part of Idaho, United States):
x1 = 0.01x40.982 x6–0.044 x80.7954 x14–0.2473 e–0.036x3
(2.31)
where Y is the mean annual soil erosion in the watershed in t, L is
the length of the watershed (100 m), S is the soil sort (for loess
S = 1, for brown soil in forestry land S = 0), B is the mean width
of watershed in 100 m, T is the ratio terrace area to the total, and F
is the ratio of forestry area to the total (Zhu, 1992).
Anderson’s equation:
where X1 is the increase of gully surface area during a given time
duration, X3 is the difference between the annual and normal
precipitation, X4 is the index of runoff in inches, X6 is the terrace
area basin in acres, X8 is the gully length at the beginning in ft, and
X14 is the length between the gully end and the basin divide in ft.
Tompson’s equation:
R = 0.15A0.49 S0.14 P0.74 E1.00
(2.32)
where R is the annual rate of gully head forward (ft), A is the basin
area in acres, S is the gradient of drainage channel in per cent, P is
the annual accumulated rainfall of intensity larger than 0.5 in/24hr in inches, and E is the percentage of clay content in weight in
the profile of erosive soil.
Estimation of length of erosive gully in the European
part of the former USSR:
L=
0.667
HQ0
0.28 2.67 2 0.67
Vp n A
log SS = –3.721 + 0.116A + 1.637FQp + 1.244MAq
+ 0.401S + 0.0486Sc + 0.482S/A + 0.028Bc
(2.37)
– 0.0036Oc + 0.942R + 0.00086Rc
where SS is the mean annual sediment flux in the watershed, A the
watershed area, MAq is the amount of mean annual runoff, FQp is
the peak degree of discharge, S is the gradient of the river, Sc is the
content of silt and clay, S/A is the rate of surface aggregate, Bc is
the cultivated area where there is a lack of crop cover effect, Oc is
the other cultivated area, R are the roads, and Rc is the recently
cut-off forest area.
Flaxman (data source: 27 watersheds with areas ranging
from 12 to 54 miles2 in the western United States):
log (Y + 100) = 524.2 – 270.71g (x1 + 100)
+ 6.41g (x2 +100) –1.71g (x3 +100) + 4.01g (x4 +100)
+ 1g (x5 + 100)
(2.33)
where H is the depth of the erosive benchmark in m, Qo is the
cross-sectional discharge of the gully in m3 s–1, Vp is the flow
velocity of erosive rock-soil in m s–1, n is the roughness, 0.05, and
A is the coefficient between 5 and 10.
(2.38)
where Y is the modulus of mean annual sediment yield in t mile–2,
x1 the ratio of mean annual precipitation in in to the mean annual
temperature in °F, x2 is the mean watershed gradient in per cent, x3
is the portion of soil particles larger than 1.0 mm in per cent, x4 is
the index of soil aggregate in per cent, and x5 is the flood discharge
stored by soil (1 csm = 0.011 m3 s–1 km–2) (ASAE, 1977).
(3)
Sediment yield model in small watershed
YRCC Formula (data source: the loess hilly gully region
in northern Shaanxi Province, China): There are 14 small gully
watersheds with an area range of 0.1 to 187 km2, a gully length
range of 0.5 to 24.1 km and a gradient range of 0.017 to 0.27.
1.38
Ws = 1.16
WT
(2.34)
0.92
L
where Ws is the total sediment yield from a single storm flood in t,
WT is the total amount of flood in a single storm in m3, and L is
the length of major gully channel in km.
North-west Institute of Soil and Water Conservation,
China Academy (data source: the small watersheds in loess hilly
gully regions):
Ms = 0.37M1.15 JKP
(2.35)
where Ms is the modulus of sediment yield in a single storm in
t km–2, M is the modulus of flood volume in m3 km–2, J is the
mean gradient of the watershed, K is the factor of soil erodibility
presenting the ratio of the amount of sand and clay to the total,
and P is the vegetation coefficient related to the canopy density in
the watershed.
Wang (1997) (data source: the small watersheds of the
Nanchuan River, western Shanxi Province, China):
Bali and Karale (India):
SI =
∑
Ei Aie D
× 10
7
AW
(2.39)
where SI is the index of sediment yield, Ei is the weighted value of
erosion intensity unit, Aie is the area of watershed erosive intensity
unit in hm2, D is the sediment delivery ratio in per cent, and AW is
the total watershed area in hm2 (Zhu, 1992).
(4)
Sediment yield models in large and middle catchments
Bivariate regression on sediment yield. Concerning
precipitation as an independent variable, Langbein and Schumm
derived two equations by regression theory.
S=
20.57 ⋅ 10
−4
1 + 1.47 ⋅ 10
P
−8
2.3
(2.40)
P
3.33
and
S=
4.14 ⋅ 10
−4
1 + 1.47 ⋅ 10
P
−8
2.3
(2.41)
P
3.33
where S is the sediment yield in t km–2.a, and P is the effective
precipitation in mm (Jansson, 1982).
Fournier found the following relationships:
0.00218L2
8.414L0.5
LnY = –2.650 + 0.962S +
+
– 4.162L2/3 + 3.252B0.5 – 1.459B2/3 – 2.227T
+ 2.456T2 – 1.392F2
2
(2.36)
Y = −49.78 + 6.14
Pm
P
(2.42)
CHAPTER 2 — SOIL EROSION
2
Y = −475.4 + 27.12
Pm
P
(2.43)
21
of the monthly sediment concentration (ρ) to the monthly runoff
(W) and maximum sediment concentration (ρm) is:
ρ = ρm [1 – e–K (w0 – Ws)]
2
Y = −513.21 + 52.49
Pm
P
Y = −737.62 + 91.78
2
Pm
(2.44)
(2.45)
P
where Y is the suspended sediment yield in t km–2.a, Pm is the
mean precipitation in the month with maximum rainfall, and P is
the mean annual rainfall. Equation 2.42 is for a region with low
relief and a precipitation regime of Pm2/P<20 and Pm2/P>8.1;
Equation 2.43 for a region with low relief and a precipitation
regime of P m 2 /P>20; Equation 2.44 is for a region with
pronounced relief in all climates except semi-arid, and Equation
2.45 is for a region with pronounced relief in a semi-arid climate.
The basic data are from 96 drainage basins, each larger than
2 000 km2 (Jansson, 1982).
Multiple regression on sediment yield. The variables
include five factors: climate, relief, soil, vegetation and land use.
Hindall, for a northern plateau:
Qs =
51.1St–0.72
Qs = 2.82 • 1010 Qa1.43 Q250.43 L–3.29 S–2.26 I–1.52
Ws = ρW0
Sa = CPad
Sa = aPa1β
Pa1 = P1 + (
∑(S
S fi
P30
L–2.52 Rc1.31 Si–6.33 Fd8.26
(2.48)
For a region without ice-laid drift:
Qs = 197 • 1A–2.38 Qa–3.14 Si0.19 L4.14 Fd–4.48 D–1.43 S2.01
)+(
30
Pf
122
) / n; β m =
ai
(2.53)
) + Pa
∑( S
S30 i
) / n; β f =
ai
(2.54)
∑(S
S fi
) / n;,
ai
where P 1 and P 30 are the maximum 1-day and 30-day rains,
respectively, and Pf and Pa are rain in the flood season (June to
September) and annual rains, respectively.
The relationship of the annual sediment discharge and
effective annual precipitation (Pa2) is:
Sa = αPa2β
For a northern ridge of a mountain and low-lying land:
Qs = 4.37 • 10–12 Qa2.63 Q2–5.83 Q255.92
(2.52)
The relationship of the annual sediment discharge and
the effective annual precipitation (Pa1) is:
(2.46)
(2.47)
(2.51)
where W0 is the base flow without sediment yield, K is the parameter, and Ws is the monthly sediment yield.
(2)
Statistical model on annual rain-sediment yield: The
relationship of the annual sediment discharge (Sa) to the annual
precipitation (Pa) is:
where =
For a central plain and western part of a glacial erosion
mountain area:
(2.50)
(2.55)
Pa2β1P1 + (βm – β1) (P30 – P1) + (βf – βm) (Pf – P30) +
(2.56)
(1 – βf) (Pa – Pf)
S1 and S30 are maximum one-day and 30-day sediment discharges
respectively, and Sf is the sediment discharge during the flood
season (Zhang, et al., 1998).
(2.49)
(5)
where Q s is the modulus of mean annual sediment yield in
t mile–2.a, Qa is the mean discharge in ft3 s–1, Q25 is the flood
discharge with a recurrence interval of 25 years, Q2 is the flood
discharge with recurrence interval of two years, L is the length of
the major river, in miles, St is the area of lake or marsh, S is the
exponent that refers to the soil seepage ability, Si the gradient of
the main channel in ft mile–1, I is the exponent of rain intensity
(24-hour rain with recurrence interval of two years) in inches, Ro
is the modulus of flood volume in ft3 s–1.mile2, Fd is the mean
freeze depth on 28 February in inches, and D is the duration coefficient of river channel discharge (10 per cent discharge divided by
90 per cent discharge using duration curve).
To estimate the sediment yield reduction on the major
tributaries of the Middle Yellow River located in the high and
coarse sediment-yield regions due to soil and water conservation works, YRCC developed the sediment yield model for
tributaries.
(1)
Statistical model on monthly effective rain-sediment
yield: From the 17 tributaries in the middle reaches, the relationship
Remote sensing information model for water erosion
E = Co (
I − Io
Io
c
) 1 h(
ST
)
c2
e
− c3 v
(2.57)
d
where E is the depth of soil erosion in the basin in mm; I is the
rain intensity in mm min–1, I0 is the threshold value of rain intensity for erosion in mm min–1, h is the depth of land surface runoff
in mm, ST is the effective depth of soil layer in mm, d is the mean
diameter of soil particles, v is the degree of vegetation cover in per
cent, and C0, C1, C2, C3 are the geographic parameters.
2.8.5
Deterministic sediment yield models
The deterministic sediment yield models are developed based on
fundamental erosion processes.
Simons, et al., developed a model based on the conception that a basin may be divided into two portions: surface runoff
and river channel systems. Four main processes of runoff are:
interruption, infiltration, flowing routing and sediment routing
(Walling, 1981).
22
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
The flowing routing was based on the continuity equation of flow and the approximate momentum equation of kinetic
wave, as well as a set of resistance functions under different
hydraulic conditions. Sediment routing requires calculating the
detachment amount of soil accounted for by rainfall splash and
land runoff, the amount of detachment and transport of wash load
by runoff, and the transported bed load. It is assumed that the
detachment soil amount due to rain splash is a singular exponent
function of rain intensity. The sediment transportation function is
the combination of the Meyer-Peter and Muller bed-load equation
and the Einstein suspended-load equation (Walling, 1981).
The Upland Soil Erosion Model for Unstable Land Flow
can simulate the processes of sediment transportation and relief
evolution. The relief trends to the concave shape which is widely
observed in the natural environment. The model is applied when
the land flow erosion is mainly caused by sheet erosion; the
kinetic-wave approximation is effective for flow solution, and the
slope gradient is less than 25 per cent; or the detachment and
transportation by splash is negligible. The simulation may be realized by the continuity equation and momentum equation of soil
and water.
In the sediment transport equation, sediment consists of
bed load and suspended load. The bed load may be calculated by:
qb = β (τ – τc)β2
(2.58)
where qb is the bed load discharge per unit width, τ is the boundary shearing stress, τ0 is the critical shearing stress, and β1 and β2
are the constants.
The suspended sediment discharge may be calculated
by:
Sq =
qb
G
w −1
11.6 (1 − G )
+ 2.5
∫
w
[(
V
U*
I
G
+ 2.5)
ln r (
1− r
∫
I
G
(
1− r
r
w
) dr
(2.59)
w
) dr ]
r
where Sq is the suspended sediment discharge per unit width,
G = d50/Y, W = Vs /(0.4 U*), Vs is the settling velocity of sediment,
U* is the shearing velocity defined as √τ / ρ, ρ is the water density,
V is the mean velocity, r = ξ /Y, and ξ is the measurement distance
from the river bed.
The total sediment discharge is the sum of bed load and
suspended discharges.
qs = qb + Sq
(2.60)
The Water Erosion Prediction Project (WEPP) was
carried out by USDA. There are three versions of the WEPP: the
hillslope, watershed and grid versions.
The hillslope version can directly replace USLE and
RUSLE. Only the function of slope sediment silting is added in
the WEPP.
The watershed version includes the hillslope version
calculating erosion on slopes. It can be used to predict the
sediment yield in watersheds and to calculate sediment
transportation, siltation and scour in river channels, sheet erosion
on terraces, and shallow gully erosion and sedimentation in
reservoirs.
The grid version can be used for any geographical
regions which do not correspond to the boundary. These regions
may be divided into a number of units. The hillslope version can
be used to calculate the erosion for each unit area.
The WEPP is a model consisting of various procedure
modules related to soil erosion.
(1) Module on erosion process. The soil erosion process
in the WEPP is divided into three stages: erosion, transport and
deposition. There are two types of erosion: rill erosion and interrill erosion. The inter-rill erosion caused by splash and thin-layer
flow is the function of the gradient and square of the rain intensity.
The rill erosion caused by runoff is the linear function of the
shearing force of flow.
(2) Module on hydrologic process. This has several
sub-modules such as meteorology, infiltration and freezingthawing. The meteorology sub-module consists of volume and
duration of rainstorm, ratio of peak rain intensity to mean rain
intensity, duration of peak rain, daily maximum and minimum
temperature, wind speed and solar radiation. These meteorological factors include the duration of runoff, peak runoff coefficient,
total runoff including snow-melt, growth amount of vegetation,
and resolving ratio in residues and water content in different soil
layers. The sub-module of infiltration uses the Green-Ampt
Equation to describe the rule of infiltration. The freezing-thawing
sub-module has been used for frost, snow-melt and snow accumulation in soil.
(3) Module on plant growth and residue process: This
module is used to estimate the effects of plant and soil residues on
soil erosion.
(4) Module on water use process. Based on the
sub-modules of meteorology, plant growth and infiltration, it
simulates the dynamic variation of water content in soil, taking one
day as a time step. It can also estimate the potential or real
evapotranspiration.
(5) Module on hydraulic process. Using the data of
runoff, hydraulic roughness, duration of runoff and peak runoff
coefficients, this module can estimate the process of runoff by
dynamic-wave equation.
(6) Module on soil process. This module presents the
dynamic variation of soil and its ground characteristics by daily
tracing. The concerned variables are natural surface roughness,
artificial roughness (height of ridge culture), bulk density of soil
and ability of saturation conduction water, soil erodibility and
shearing force of critical flow. This module also considers the
effects of cultivation, weathering and aggregate rainfall on the
ground characteristics (Liu, 1997).
Models for Erosion and Sediment Yield in Small
Watersheds in the Loess Hilly-gullied Regions of Shaanxi
Province, China.
The new ERODE model, combining erosion models and
GIS, is useful in planning soil conservation measures and watershed management. This model is designed to estimate annual
runoff and soil loss. In the Loess Plateau, most rainfall takes the
form of storms. The runoff and sediment yield are calculated for
each rainstorm. The model consists of three sub-models:
slopeland, gully and slope, and gully.
(1) Slopeland sub-model: This considers only the splash
erosion when there is no rill. It considers both rill erosion by flow
in the dominant act and the inter-rill erosion by splash erosion. It
includes two main processes. One is the effect of crust on top soil;
CHAPTER 2 — SOIL EROSION
23
Amount of sediment:
the other simulates the processes of soil erosion and transportation
on rills.
for red clay area:
SG = 106.57QG1.138
(2.72)
for loess area:
SE = 225.2QE1.196
(2.73)
Erosion force of rainfall:
Ek = (P – Z) (28.83 + 13.51gI)
(2.61)
Amount of sediment caused by cave erosion:
where P is the total rainfall in mm, Z is the volume intercepted by
vegetation, and I is the mean rain intensity in mm min–1.
0.373J1.02
Ss = 91.84R1.04
s L
St = 169.02RT1.04L0.13
Erosion force of runoff:
Ps = 0.001ρgRA sin θ
(2.62)
where A is the area in m2, g is the gravity acceleration, R is the runoff
in mm, ρ is the water density, and θ is the gradient in degrees.
Splash erosion:
Di = 0.015 J
Ek
τ
e
( 2.68 sin θ − 0.48 Cv )
(2.63)
ST = St – Ss
4.8
Dr = 7.9 Ps
(
τ
τb
)
−0.5
⋅ 10
−8
(2.64)
(2.75)
(2.76)
where Ss is the amount of soil erosion on the topsoil of the cave, St
is the gross sediment yield from the cave in kg, L is the length of
the cave in m, J is the height of the cave in m, ST is the amount of
net erosion of the cave in kg, RT is the runoff depth of the cave in
mm, and Rs is the runoff depth on the land surface slope in mm.
(3) Sub-model on gully erosion. Sediment delivery ratio
(Rsd):
where Di is the splash erosion in kg m–2, J is the soil crust factor (it
takes 0.7 with crust and 1.0 without crust), τ is the shearing force
on top soil, and Cv is the degree of vegetation cover in per cent.
Rill Erosion (Dr):
(2.74)
0.59
Rsd = 0.0277P −0.29C 0.19 Sm
(
E a 0.44
)
E
(2.77)
where P is the amount of rainstorm in mm, C is the runoff coefficient, Sm is the sediment concentration during the flood peak in
kg m–3 and Ea/E is the proportion of kinetic energy by rainfall of
intensity exceeding 0.15 mm min–1.
Sediment transport force by flow (Tc):
2.9
Tc = 0.0081P1.55
s
(kg m–2)
(2.65)
Soil erosion (SL):
if Di < Tc, then SL = Di
(2.66)
if Di > Tc, then SL = Tc
(2.67)
if Si + Dr < Tc, then SL = Si + Dr
(2.68)
if Si + Dr > Tc, then SL = Tc
(2.69)
where SL is the amount of soil erosion, and Si is the amount of
inter-rill erosion, 2.5 kg m–2.
(2) Gully-slope sub-model: Because of the steep slope
of€gullies, gravitational erosion occurs often. But the sediment
yield caused by landslides, even in large scale, may not be so
enormous. In fact, the collapse of shallow layers often has more
effect on sediment yield. Another erosion process is cave erosion.
Amount of runoff:
1.04 E1.14 (mm) (2.70)
for red clay area: QG = 1.086 • 10–4 PA0.164 I30
for loess area:
QE = 1.29 • 10–4 PA0.225 E1.509
(mm) (2.71)
where PA is the antecedent affecting rainfall, and E is the kinetic
energy of rainfall in joule m–2.mm.
SOIL EROSION CONTROL AND WATERSHED
MANAGEMENT
Soil erosion control is a complex engineering system to promote
the sustainable development of agricultural production and social
economies. It concerns a number of aspects such as environment,
scientific techniques, economies, societies, policies and regulations. Soil erosion control includes planning and management of
soil and water conservation measures closely related to watershed
management.
2.9.1
Soil and water conservation planning
Soil and water conservation planning is to control soil erosion and
regulate river channels in a certain area. It is based on the situation
of soil and water loss, conditions of natural resources and social
economy, and the strategic goals of national economic development following the principle of soil and water conservation and
ecology.
(1)
Categories and tasks of planning
The categories of soil and water conservation may be divided
into the following: (1) National scale — taking a large river
basin or large natural region as a unit; (2) Large river scale —
taking large rivers with an area ranging from several dozen to
hundreds of thousands of km2 as a unit; (3) Large tributary scale
— taking an area of several thousand to several tens of thousands of km2 as a unit; (4) Small watershed scale — several to
several hundred km2.
The first two categories require strategic middle- and
long-term planning. The basic tasks are systematically to carry out
24
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
environmental recognition and resource evaluation based on a
comprehensive investigation of nature and society. They should
predict the economic, ecological, environmental and social benefits of soil and water conservation, and coordinate agriculture,
forestry and husbandry in planning. The third and fourth categories are the practical planning for middle and short terms
covering medium and small scopes. The basic tasks are to produce
local plans on land use, tillage, vegetation and engineering
projects.
(2)
Planning approaches
The land use plan is the core plan. Land evaluation is based on
land specifications. The definition of land evaluation in The Sketch
on Land Evaluation edited by the Food and Agricultural
Organization of the United Nations (FAO) is to compare and illustrate the basic conditions of soil, vegetation, meteorology and
other aspects of the land, and to carry out appraisals and comparisons for the prospective land.
At present, the Land Potential Gradation developed by
the USDA Soil Conservation Service has been widely adopted
under the recommendation of FAO. The land is divided into eight
degrees based on the limited intensity of crop or grass which has
been acted upon by soil. Optimum land use planning usually
adopts the method of linear planning. Linear algebra is used for
resource distribution. That is, with the existing natural resources,
resources of social economy and technology, a scientific decision
should be made to achieve the best social, economic and ecological benefits.
2.9.2
Measures for soil and water conservation
Cultivation measures for soil and
water conservation
Increasing the roughness of land surface, changing the microrelief of land slope and improving vegetation cover can foster
soil and water conservation and improve soil texture (Hudson,
1981).
Tied ridging: Closely spaced ridges are arranged on the
ground surface to form a series of rectangular depressions. When
the soil becomes saturated and the depressions are full, overflow
occurs and the ridges break. On slope ground, once a ridge is
broken, a small flood is released and bursts the next ridge, storing
more water, and so on down to the end of the slope. This measure
has been used successfully on deep permeable soils of East Africa
and in western Gansu Province, China.
Contour cultivation and grass strips: On gentle slopes
or where erosion risk does not warrant major earth-moving
works, it may be sufficient to slow surface runoff by carrying out
all tillage operations on the contours. Another protection measure
involves using grass strips when the soil erosion is not severe.
Surface runoff moving down the slope is intercepted by the grass
strips, the velocity is slowed, and silt is deposited in the grass
strips.
Ridge and furrow: The ground is tilled into wide parallel
ridges approximately 10 m wide, with intervening furrows about
0.5 m deep. Surface runoff moves across the ridges to the furrows,
then down the furrow, which is on a gradient of about 1:400. This
method is particularly suitable to large areas of gentle sloping
land, but for channel terraces it requires some controlled surface
drainage.
(1)
(2)
Engineering measures for soil and water conservation
(a)
Agricultural arable land. Channel terrace: If surface
runoff flows down the slope of arable land without any impediment, it not only carries away the soil dislodged by splash erosion,
but also scours the soil down. To avoid this, terraces are used to
intercept the surface runoff. In some African countries, a broadbased contour ridge has a wide (15 m) and low bank and a shallow
channel with gentle sloping sides; a narrow-based contour ridge
has a steep-sided bank with a width of only 3 to 4 m (Hudson,
1981).
Bench terrace: Bench terraces entail converting a steep
slope into a series of steps with horizontal or near horizontal
ledges. To hold up the vertical face, some structures are necessary.
Usually these are stone structures, but bricks or timber are also
used. There are different types of bench terraces, such as outwardsloping, inward-sloping and reverse-sloping ones. It is desirable
for each bench to be as wide as possible for cultivated crops.
Small terraces for fruit-trees, coffee plants and vines are equally
effective and require less earthmoving.
Irrigation terrace: A flat bench terrace has a raised lip at
the outer edge to retain irrigation water. It is extensively used for
the production of rice, and also for tea, fruit trees, and other high
value crops. For paddy fields, the terraces are level so that each
terrace becomes a shallow pond.
Orchard terrace: If the soil is too shallow or the slope is
too steep, bench terracing may not be practical. In this case, the
land may be developed for tree crops by using intermittent
terraces, otherwise known as orchard terraces. These are small,
level or reverse-slope terraces, each having one line of trees. The
important feature of any of these development techniques for steep
erosion-prone slopes is that the land between the terraces must be
planted with a vigorous cover crop, such as a creeping legume. In
Kenya, one type of intermittent terrace is used. The excavated soil
is used to build a bank above the ditch with the purpose of catching silt to form a more level terrace.
Terrace systems: Terraces, as mechanical erosion-control
measures in slope cropland, are used to alter flow length, provide
temporary runoff storage, and reduce slope gradient. Terrace
systems can meet water management and erosion-control needs
for intensive slope cropland.
(b)
Non-arable land. Mechanical protection of forest soils:
Mechanical protection is not usually required for natural forests,
but commercial planting may well need some protection during
establishment and after harvesting. Two forms are most common:
contour trenches and contour furrows. Both are similar to the
structures used for arable land. Contour trenches are commonly
used in America on steep land from 30 to 75 per cent. The
trenches are usually built without any gradient in the channel,
since the objective is to hold runoff until it infiltrates the soil.
Cross-ties are added every 10 to 15 m to further restrict water
movement. Contour furrows are similar in form, but smaller, and
are used on gentle slopes up to about 35 per cent. They have a
smaller water-holding capacity (Hudson, 1981).
Mechanical controls on grazing land: Poor grazing land
has such low production levels that only very simple and inexpensive measures are economically justified. Such measures are not
designed to control soil movement directly, but to improve the
vegetation by reducing runoff and increasing infiltration. Two
types of structures are used. Pasture furrows are small and have
level open drains that follow the real contours and are fairly close
CHAPTER 2 — SOIL EROSION
together, like the large channel terraces used on arable land. The
other approach is to form many small surface depressions which
hold and store runoff.
Erosion control on roads: Siting and alignment: The
siting of a new road can be established efficiently using aerial
photography. The first rule of road siting is to place roads on
crests wherever possible. When it is impossible, the next alignment is on a gentle gradient fairly close to the real contours.
Gradients of the order of 1/100 to 1/500 are desirable for the openchannel drains required along roads. A gradient of 1/100 to 1/20
may cause some problems for controlling soil erosion on side
drains. For a gradient steeper than 1/20 it is usual to adopt a
zigzag layout or the combination of one reach on gentle slope and
some reaches straight down the slope.
Road drainage: In siting roads, swamp and permanently
wet areas should always be avoided. Roads straight up and down
the steepest of slopes need side drains only to deal with the runoff
from the road surface, and this water can be easily dealt with by
mitre drains. A wide shallow cross-section with a gentle side slope
will provide the best hydraulic design, and regular mowing of the
cover grass has been shown to be the most effective and cheapest
maintenance.
(c)
Structures for gully erosion control. Temporary structures: If the objective is to slow down the water and so cause
deposition of silt, there is no need for the structures to be watertight. These are called porous checks.
Wire bolsters: If there is plenty of loose rock available
nearby it can be used to build a loose rock-fill dam anchored in
place by wire netting. Galvanized wire netting of a fairly stout
gauge and two metres or more in width is laid out flat across the
gully bed. Loose rock is packed on one half of the width of the
netting, and the other half is wrapped over the stones and laced to
the other edge.
Netting dam: Another use of wire netting is to form
small check dams, usually near the top end of gullies. Wooden
posts are driven into the bed of the gully and used to support a
strip of wire netting which forms a low wall across the gully. Light
brush or straw is piled loosely against the upstream side of the
netting wall.
Brushwood dam: In wooded areas, two types of silt
retaining dam are adopted. The brushwood dam uses small
branches, up to two or three cm in diameter, packed across the
direction of flow. They can be anchored by packing them between
rows of vertical stakes.
Log dam: When heavier timber is available it can be
used for log-piling dams. Two rows of vertical posts are driven
into the bed of the gully, extending up the side to above flood
level, and then logs are packed in between them. In a wide,
shallow river it is best to drive in all the vertical posts to the same
height above the ground, so that the top of the dam follows the
section of the river bed.
Brick weirs: The shape that gives the best strength-toweight ratio is an arch weir. A single thickness of brickwork can
be built to a height of 1 to 1.5 m over a circular span of about 2 m.
A straight wall of similar size would need three or four times more
brickwork to be of comparable strength.
Permanent structures:
Silt-trap dam: A quick positive reduction in sediment
movement can be achieved by building permanent silt trapping
dams.
25
Regulation dam: This is a useful application of permanent dams to regulate flash floods, using the leaky bath-tub
principle. A permanent dam is built with sufficient storage for the
runoff from a single storm. The outlet consists of a pipe which
allows the flood water to drain away in one or two days, leaving
the storage reservoir empty for the next storm.
Gully-head dam: This is used when an active gully is
developing steadily in an upstream direction and must be stopped
before it threatens roads, bridges or similar structures. An effective
way of controlling the erosive force of runoff over the gully head
is to submerge the gully head in the pond of a permanent
impounding dam. The energy of the inrushing water is then dissipated as it flows into the pond.
Drop structure: This is built using masonry, bricks, or
concrete to allow the flood runoff to pass over harmlessly. The
capacity of drop structures is controlled by the size of the inlet. It
acts as a rectangular weir with the flow proportional to the length
of the weir.
Cabion: The main problem with rigid structures is that
they cannot adapt to the conditions of surrounding soil. One
construction that can overcome this difficulty is a more sophisticated version of the wire netting bolsters. This method was
developed in Italy and uses pre-fabricated rectangular baskets
called cabions. Its main advantage is that there is sufficient flexibility for the structure to adjust to settling resulting from scouring
the foundation (Hudson, 1981).
Sediment controlling reservoir: This reservoir can trap
sediment (rock, silt and floating material) scoured down by a
flood. There are four types of sediment controlling dams: (a)
sluicing gate dams; (b) open mouth dams; (c) grid dams; and (d)
net dams.
(3)
Vegetation measures for soil and
water conservation
Soil and water conservation forests: Any artificial or natural
forests having the function of improving the ecological environment, conserving water resources, preventing soil erosion or
regulating the hydrological status of rivers, lakes, and reservoirs
are called soil and water conservation forests.
One hectare of forest can store about 300 m3 of rainfall.
The forestry canopy intercepts the rainfall, and the layer of withered branches and falling leaves absorbs the surface runoff.
According to measurements, the canopy can intercept 15 to 40 per
cent of rainfall, and 1 kg of withered fallen material can absorb 2
to 3 kg of water. Also, the permeability of forest soil is 3 to 10
times higher than grassland or arable land.
Forests can lower the mean annual temperature, reduce
the temperature difference and increase humidity. Each hectare of
forests can absorb 192 kg of CO2 in a day. The dust content in
forests is 20 to 40 per cent lower than in open country.
Mixed forests of male and female trees, bush and tree or
complex mixed forests should be arranged according to local
conditions.
Grass for soil and water conservation: Planting grass can
conserve soil and water and improve the physical and chemical
quality of soil. The functions of grass are to: (a) store water,
preserve soil moisture and prevent soil erosion; (b) improve soil
and increase its fertility (one hectare of alfalfa can fix 225 kg of
nitrogen in three years, which is about 750 kg of ammonium
nitrate); and (c) provide forage, fertilizer and fuel.
26
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
(4)
Wind erosion control on cultivated land in arid areas
Stubble mulching and less ploughing: Stubble mulching is one of
the most effective methods to prevent wind erosion and conserve
soil moisture. It is mainly used for crops of wheat, sorghum, and
other millets. The crop is planted directly in the field where the land
is covered with stubble, thus using less ploughing for intertill crops.
Contour strip cropping: The strip is perpendicular to the
common wind direction. It also needs appropriate crop stubble to
resist wind erosion.
Wind breaks and forest belts to resist wind: The efficiency of this method depends on wind speed and direction, and
on the shape, width and spacing of the wind break. If the wind
direction is perpendicular to forest belts, the wind speed will
decrease by 70 to 80 per cent. It is usual to take an interval of
1 300 ft for forest belts. In areas with high and medium intensive
wind erosion, the intervals between belts may be 350 to 450 ft and
500 to 650 ft, respectively (Woodruff, et al., 1981).
(5)
Erosion and sediment control for surface mining
Improper surface mining or waste piles from deep mining will
cause serious soil erosion. The key is to have proper planning
before mining starts. Provisions must be made for the placement of
overburden, controlling head cutting and sheet erosion, sediment
retention, and land stabilization. Suitable soil should be placed on
the surface to facilitate the growth of new vegetation. Spoil piles
should be kept away from the system. Land stabilization during
mining and after reclamation must be an integral part of the planning. Settlement basins may be used to trap sediment. In mining
processes, backfill and reclamation should be carried out simultaneously. Measures for soil and water conservation should be part of
these processes. The entire area should be protected by vegetation
or using other methods. Drainage systems and settling ponds are
adopted to eliminate the impacts of surface mining on water quality
of the adjacent areas. A monitoring system should be established to
monitor the dispensing polymer electrolytes and flocculation of
suspended sediment (ASAE, 1977).
(6)
Erosion and sediment control in urban areas
Extension of urban area requires the construction of roads and
buildings on a large scale that disturbs the original environment
and removes a lot of soil, thus causing soil erosion. Erosion and
sediment control in urban areas in Malaysia reduced by 68 to 80
per cent the sediment yield from construction sites between 1966
and 1974 according to a report by United States Geological
Survey (USGS). Much sediment yield comes from urban
construction. The report by Wolman and Schick shows that the
sediment yield from urbanized or developing areas ranged from
several hundred to 55 000 t km–2 per year. Yorke and Herb indicated that the annual average sediment yield from cultivated land
was 620 t km–2, while that from construction sites ranged from
1 610 to 22 600 t km–2, and the average was 7 330 t km–2.
Desirable practices for erosion and sediment control
include the following:
Temporary structural practices: They can be divided into
two groups: water control and sediment control. Water control
includes small diversion terraces (dikes), small waterways
(swales), and grade stabilization structures. The sediment control
practices consist of sediment traps for drainage areas smaller than
2 ha and sediment basins for drainage areas of 2 to 40 ha.
Permanent structural practices: Diversion, grassed
waterways, level spreaders and subsurface drains are used for
agriculture, and storm drain outlet protection, land grading and
riprap are used only for urban areas.
Vegetation practices: These include both temporary and
permanent practices for establishing ground surface cover to
control soil erosion. They include seeding, sodding, mulching, as
well as criteria on ground cover, vines, shrubs and trees.
Special practices: These include vegetation tidal bank
stabilization using original topsoil, protection of tress in urban
areas, seeding strip-mine areas, dune stabilization, dust control
and protective material for channel and steep slopes (ASAE,
1977).
2.10
SUMMARY ON GLOBAL SOIL EROSION
The problem of soil erosion has been given more and more attention in the world. The total erosive area in the world is 25 million
km2, accounting for 16.8 per cent of the total continent area. One
third to one fourth of the topsoil of arable land suffers from
serious soil erosion. About 60 billion tons of fertilized top soil is
eroded and about 17 billion tons of sediment flow into the ocean
or sea annually. Fournier pointed out that the maximum sediment
yield occurs in semi-arid regions. Based on an estimation by
Table 2.6
Present status of soil erosion and global trends
Water erosion
Regions
Africa
Wind erosion
Area
(106 hm2)
Annual denudation
(mm)
Annual losses
(106t)
Trend
Area
(106 hm2)
Trend
227
0.023
201
+
186
+
Asia
441
0.153
1592
+
222
+
South America
123
0.067
603
+
42
–
Central America
46
0.055
758
+
35
–
North America
60
0.055
758
+/–
35
–
Europe
114
0.032
425
+/–
42
+/–
Oceania
83
0.390
293
+
16
+
1094
0.079
3872
+
548
+
World
NOTES: 1. Area after Oldman, 1991–1992; 2. Denudation after Lal, 1994; 3. Soil losses after Walling, 1987; 4. “+” increasing, “–” decreasing.
CHAPTER 2 — SOIL EROSION
UNDP, 5 to 7 million km2 of arable land is lost annually due to
soil erosion, and the yearly economic losses reach $10 billion
(Zhao, et al., 1997).
The mean annual precipitation and rain intensity are the
most important factors affecting water erosion. In tropical regions,
intensive downpours can cause much more damage than in
temperate climates. In general, water erosion in the regions
between latitudes 40° North and 40° South is the most serious in
the world. It includes North America and part of South America,
most of Africa, except the dry and desert areas and the equatorial
forest, Asia up to 40° N, as well as the dry central areas of
Australia.
As for wind erosion, the main regions are North America
(the Great Plains Dust Bowl), the Sahara and Kalahari deserts in
Africa, north-western China, Central Asia (particularly the steppes
of Russia), and central Australia (Hudson, 1981).
Global sediment maps: Different methods have been
used to survey water erosion on a global scale. The map by
Fournier is a map of suspended sediment yield in a basin area
larger than 2 000 km2. Another world map of erosion rates was
contributed by Strakhov on the basis of the suspended load in 60
rivers. Table 2.6 shows the present status of soil erosion and global
trends (Jansson, 1982).
A total of about 50 million km2 land is in arid, semiarid, and dry sub-humid regions. In these regions, about 3.1 billion
ha and 3.1 billion ha of grassland are undergoing medium and
serious desertification, respectively; 335 million ha and 170
million ha of rainfed cropland are suffering medium and serious
desertification, respectively; 40 million ha and 13 million ha of
irrigated cropland are being subjected to medium and serious
desertification, respectively (Wang, 1997).
REFERENCES
21st Century United Nations Conference on Environment and
Development, 3–14 June 1992 (Chinese version translated
by China Environment Bureau).
ASAE, 1977: Soil erosion and sedimentation. Proceedings of the
National Symposium on Soil Erosion and Sedimentation by
Water, ASAE Publication 4–77, Palmer House, Chicago.
Bennet, H.M., 1939: Soil Conservation. New York-London.
Chen Qibo and Fei Xiliang, 1996: The new progress in prediction
of soil erosion. Journal of Chinese Soil and Water
Conservation, Number 2.
Chen Yongzong, Jing Ke and Cai Guoqiang, 1988: Modern Soil
Erosion and Management of the Loess Plateau. Science
Press (in Chinese).
China 21 Century on Population, Environment and Development
White Book (Chinese version), Beijing, 1994.
Division of Sediment, Chinese Hydraulic Engineering Society
(CHES), 1992: Manual of Sediment. China Environmental
Science Press (in Chinese).
Encyclopedia of Chinese Agriculture (Volume on Water
Engineering), 1987 (in Chinese).
Flaxman, E.M., 1963: Channel stability in undisturbed cohesive
soils. Journal of the Hydraulics Division, American Society
of Civil Engineers (ASCE), 89 (Hy2), Proc. 3462, March,
pp. 87–96.
Foster, G.R. and W.H. Wischmeier, 1973: Evaluating Irregular
Slopes for Soil Loss Pediction. Paper 73–227, ASAE,
Lexington.
27
Frazier, B.E., D.K. McCool and C.F. Engle, 1983: Soil erosion in
the Palouse: an aerial perspective. Journal of Soil and Water
Conservation, Volume 38, Number 2.
FAO, 1965: Soil Erosion by Water, Some Measures for its Control
on Cultivated Lands.
Gong Shiyang, 1998: Soil Erosion on the Loess Plateau of the
Yellow River Basin (in Chinese).
Gottschal, L.C., 1975: Nature of Sedimentation Problems.
Guo, Tingfu, et al., (ed.) 1998: Standards for Classification and
Gradation of Soil Erosion. Trade Standard, Ministry of
Water Resources of China (in Chinese).
Holy, M., 1982: Erosion Environment. Technical University of
Prague, Czechoslovakia (translated by Janaa Ondrachova).
Hua Shaozu, 1990: Planning on Soil and Water Conservation.
Regional training course on soil erosion and its control.
Hudson, N., 1981: Soil Erosion. Batsford Academic and
Educational, London.
Jansson, M.B., 1982: Land Erosion by Water in Different
Climates. UNGI Report Number 57.
Liu Tungsheng, 1985: Loess in China. Springer Series in Physical
Environment.
Liu Zengwen, 1997: Introduction to the WEPP model on prediction of water erosion. Journal of Chinese Soil and Water
Conservation, Number 12.
Margan, R.P.C., 1980: Soil Conservation Problems and Prospects.
Mayer, L.D., 1984: Evaluation of the Universal Soil Loss
Equation. Journal of Soil and Water Conservation,
Volume 39, Number 2.
Meng Qingmei, Hua Shaozu, et al., 1996: Soil and Water
Conservation in the Loess Plateau. Yellow River Hydraulic
Press (in Chinese).
Meyer, L.D., 1984: Evolution of the Universal Soil Loss Equation.
Journal of Soil and Water Conservation, Volume 39, Number 2.
Ministry of Water Resources of China, 1998: Standards for
Classification and Gradation of Soil Erosion-Trade Standard.
Mutchler, C.K., 1963: Runoff plot design and installation for soil
erosion studies. Agricultural Research Service Report
Number 41–79, USDA, August 1963.
Norman, H., 1981: Soil Conservation. Batsford Academic and
Educational Ltd, London.
Onstad, C.A. and G.R. Foster, 1974: Erosion and deposition
modeling on watershed. Scientific Journal Series 8537,
Minnesota Agricultural Experiment Station.
Piest, R.F. and C.R. Miller, 1975: Sediment sources and sediment yield.
Chapter IV of Sedimentation Engineering, (ed.) V.A. Vanoni.
Schertz, D.L., 1983: The basis for soil loss tolerances. Journal of
Soil and Water Conservation, Volume 38, Number 1.
Sharma, 1998: CTA of UN participatory watershed management
programme in Asia. Proceedings of the International
Symposium on Comprehensive Watershed Management
(ISWM-’98), 7–10 September 1998, Beijing.
Spomer, R.G. and R.L. Mahurin, 1984: Time-lapse remote sensing
for rapid measurement of changing landforms. Journal of
Soil and Water Conservation, Volume 39, Number 6.
Vanoni, V.A. (ed.), 1975: Sediment Engineering. ASCE, New York.
Walling, D.E., 1998: Opportunities for using environmental
radionuclides in the study of watershed sediment budgets.
Proceedings of the International Symposium on
Comprehensive Watershed Management (ISWM-’98),
7–10 September 1998, Beijing.
28
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
Walsh, S.J., 1985: Geographic information systems for natural
resource management. Journal of Soil and Water
Conservation, Volume 40, No 2.
Wang, Lixian, 1995: Soil and Water Conservation. Beijing Forest
University, June.
Wang Lixian, 1997: Land Degradation in Globe and its
Preventing and Harnessing Measures, Soil and Water
Conservation of China (in Chinese).
Williams, J.R. and H.D. Berndt, 1972: Sediment yield computed
with universal equation. Journal of the Hydraulics Division,
ASCE, 98 (Hy12), Proc. 9426, December, pp. 2087–2098.
Wu Changwen, Liu Weichang, and Zhan Dingsheng, 1997:
Principles and methods of planning of soil and water conservation in urban areas. Journal of Chinese Soil and Water
Conservation, Number 1.
Xin Shuseng and Jiang Deqi, 1982: An Introduction to Soil and
Water Conservation in China. Agricultural Press (in
Chinese).
YRCC, 1991: Remote Sensing in the Yellow River Basin-Soil
Erosion (in Chinese).
Zachar, D. 1982: Soil Erosion. Forest Research Institute, Zvolen.
Zhang Shengli, et al., 1998: The Reasons and Trends of Changes
in Runoff and Sediment Yield of the Middle Yellow River
Basin. Yellow River Hydraulic Press (in Chinese).
Zhao Yi and Liang Weilu, 1997: Soil erosion and its control in the
USA. Journal of Chinese Soil and Water Conservation,
Number 5 (in Chinese).
Zhu Pengcheng (ed.), 1992: Sediment Manual. Chinese
Environment Science Press (in Chinese).
CHAPTER 3
SEDIMENT TRANSPORT IN RIVERS
PATTERNS OF SEDIMENT TRANSPORT IN
RIVERS
3.1.1
Bed material load and wash load
Sediment is classified as either bed load or suspended load according to the patterns and laws of movement. It can also be classified
as bed material load and wash load according to the particle size,
its origin and effect in fluvial processes.
The ratios of fine to coarse sediment in river bed material sediment are quite different. Sediment in river beds is often
(but not always) composed of much coarser and much less fine
sediment than moving sediment. There is always an exchange
between coarse sediment and bed material during transport.
Incoming coarse sediment may originate directly from the river
bed of an upstream reach. It is directly supplied from the bed
and therefore is called bed material load. In contrast, fine sediment, eroded and washed from upland watersheds, is
transported through a channel over a long distance and is
scarcely ever deposited in the channel; therefore, it is called
“wash load”. Thus, the amount of coarse sediment carried by
flow depends on sediment transport capacity and exhibits a
well-defined relationship with the flow discharge. In contrast,
the concentration of fine sediment depends only on the supply
of sediment from the upstream reach, and no obvious relationship with flow discharge is found.
Sediment can be classified as bed material load and
wash load, or bed load and suspended load. It should be emphasized that the two sets of classification of sediment are distinct and
should not be intermingled. Bed material load may move as both
bed load and suspended load, and the same is true for wash load.
Of course, wash load is fine and mainly moves as suspended load.
It is not correct to identify the bed material load with bed load and
wash load with suspended load.
3.1.2
Bed load, saltation and suspended load
It should be pointed out that only bed material load, not wash
load, is discussed here.
At low flow, although some sediment moves in suspension, most sediment particles move in the form of sliding, rolling
and saltation in a zone close to the bed surface with a thickness of
1 to 3 times the particle diameter. Such sediment is called bed
load. This zone is called the bed surface layer.
With increasing flow velocity, some particles are caught
by turbulent eddies. Entering the main flow region, these particles
are transported downstream by flow. Sediment supported by turbulent eddies and moving downstream in suspension is called
suspended load.
With a high level of shear stress, however, not only can
the particles enter into motion on the bed surface, but also those
in the subsurface layer of the bed can do so as well. This motion
penetrates further into the bed in response to further increases in
shear stress. The velocity of the moving sediment is significantly
smaller in a deeper bed. The sediment that moves in such a way
is called the laminated load.
3.1.3
Continuity of sediment movement
Sediment motion can be viewed as a continuum even though the
sediment is classified in categories such as bed load and
suspended load according to its mode of movement. There are
continuous exchanges between these loads as well as between the
material in the bed and that being transported. That is, there is an
exchange between suspended load and bed load, and between bed
load and bed material. When a large eddy sweeps over the river
bed, a direct exchange between suspended load and bed material
can occur.
3.1.4
Relative importance of bed load and suspended load
The relative importance of bed load and suspended load depends
on sediment size and flow velocity. For the same composition of
bed sediment, sediment slides, rolls or moves in saltation if flow
velocity is low. As velocity increases, part of the sediment is
carried into the main flow zone and becomes suspended load. The
rest remains in the bed surface layer and moves as bed load, but
the thickness of the bed surface layer is augmented. Following still
further increases in flow velocity, the suspended load is greater,
and it exceeds the bed load. In general, for ordinary river flows,
sediment coarser than a certain diameter moves mainly as bed
load, and sediment finer than that diameter moves mostly in
suspension.
If the critical conditions for sediment incipient motion,
the fall in velocity of the sediment and the nature of the turbulence
of flow are known, the patterns of sediment motion in flow can be
roughly predicted, as shown in Figure 3.1 (Chien and Wan, 1983).
The condition for sediment initiation using the shear velocity as
Fall velocity ω or threshold shear velocity U* (cm s-1)
3.1
Gilbert
US waterways
Experiment Station White
Diameter D (mm)
Figure 3.1 — Zoning of sediment movements.
1 – Grain Reynolds number U*D/v = 3.5; 2 – Form resistance dominates;
3 – Skin friction dominates; 4 – Fall velocity ω; 5 – Sliding, rolling and saltating;
6 – Threshold shear velocity U*
30
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
the main parameter, which will be discussed later, is shown as
COD in the figure. A conclusion from the data by Nikijin is that
the shear velocity in most zones of flow equals roughly the rootmean-square of the vertical component of the fluctuating velocity,
except in the zone close to the boundary. Curve EOF in Figure 3.1
represents the fall velocities of sediment of various sizes. Curves
COD and EOF divide Figure 3.1 into several zones, and each of
them is characterized by a different kind of sediment movement:
1. In the zone below curve DOE, the fall velocity of the sediment coming from upstream is larger than the vertical
component of the fluctuating velocity and, therefore, sediment will settle. Because the shear velocity of flow is lower
than the critical value for sediment initiation, i.e. U*< U*c,
the settled sediment will accumulate on the bed.
2. In the zone between CO and OE, the sediment in flow can
remain in suspension because the fall velocity of the sediment is less than the upward component of fluctuating
velocity. However, the sediment on the bed of the same size
cannot be picked up by the flow because of the influence of
the laminar sublayer and cohesive forces. One can say for
simplicity that sediment coming from upstream is transported through the river channel without any exchange with
the bed sediment.
3. In the zone between DO and OF, the shear stress of flow is
over the threshold value for initiation but the turbulence is
not strong enough for sediment suspension. Sediment moves
in this zone as bed load.
4. In the zone above CO and OF, sediment cannot resist movement by flow and is likely to be suspended once it begins to
move. Bed load and suspended load coexist in this zone.
The higher the shear velocity, the more suspended load there
will be.
3.2
3.2.1
3.2.1.1
BED LOAD
Incipient motion of sediment
STOCHASTIC PROPERTY OF INCIPIENT MOTION OF
SEDIMENT
Incipient motion is an important critical condition that determines
which sediment starts to move under the action of flow. If flow
intensity exceeds a certain value, sediment particles begin to
move. The flow condition that corresponds to this critical limit is
called incipience.
Although the flow condition for which the sediment
grains on the bed start to move is a well-defined physical concept,
many difficulties are encountered in determining the actual threshold condition for specific cases. A typical bed surface is composed
of innumerable sediment grains of various combinations of sizes,
shapes, specific gravities, orientations, packing and locations.
Besides, water flow also has fluctuation characteristics. Therefore,
the forces exerted on sediment grains vary with both time and
space. Thus, even for uniform sediment, the grains do not all start
to move or come to rest together. For non-uniform sediment, the
conditions are much more complicated. Even for given flow
conditions, one cannot define a specific grain size such that larger
particles remain at rest and smaller particles are all in motion.
Also, the spatial distribution of sediment movement at a certain
instant is such that grains move at some places and remain at rest
at others. And at certain locations of the bed, sediment moves
during one time interval, and fails to move during another. The
incipient motion of sediment is clearly a stochastic phenomenon.
Obviously, if the criterion for incipient motion is determined according to any one of two, three or four different
conditions, the results will be quite different.
Dou (1962) used the velocity near the bed as the
hydraulic parameter to determine the incipient motion of sediment. According to his analysis, in which the fluctuation of flow is
considered, three probabilities for incipient motion that correspond to Kramer’s criteria (1935) for bed load movement are as
follows:
1. Occasional individual motion,
pc1 = p [u0 > uc = u–c + 3σu0 = 2.11u–c] = 0.00135
(3.1)
2.
3.
Sparse motion,
pc2 = p [u0 > uc = u–c + 2σu0 = 1.74u–c] = 0.0227
(3.2)
Strong motion,
pc3 = p [u0 > uc = u–c + σu0 = 1.37u–c] = 0.159
(3.3)
where u–c is the time average of the critical velocity near the bed,
and σu0 is the standard deviation of the velocity fluctuation near
the bed.
3.2.1.2
CONDITION OF INCIPIENT MOTION FOR NON-COHESIVE
UNIFORM SEDIMENT
Here, the simplest case of non-cohesive uniform sediment is
discussed. The hydraulic parameters for this condition can be
expressed by shear stress (drag force) or average velocity.
(A) Shear stress approach. Early in 1936, starting with
the balance of forces acting on a particle on the bed, Shields
(1936) deduced the following function for the incipient motion of
non-cohesive uniform sediment:
τ0
U D
= f( ∗ )
(γ s − γ )
υ
(3.4)
This is the formula Shields used for threshold drag force, where τc
is threshold drag force, γs is unit weight of sediment, and γ and υ
are unit weight and kinetic viscosity of water, respectively. The
form of the function f in Equation 3.4 must be determined by
experiment. Based on data from experiments by Shields and other
investigators, an average curve shown in Figure 3.2 was obtained.
Actually, there were no data for grain Reynolds numbers
smaller than 2 when Shields drew the curve. Compared with the
relationship between the drag factor and the grain Reynolds
number for a settling particle, Shields deduced that in that range
τc/ (γs – γ) D was proportional to the reciprocal of grain Reynolds
Curve for
laminar flow
Figure 3.2 — Condition for incipient motion for non-cohesive
sediment (Shields curve and its modification).
CHAPTER 3 — SEDIMENT TRANSPORT IN RIVERS
number. After Shields’ work, a number of other researchers,
including Tixon, Li, White and Mantz studied the incipient motion
of sediment. Their results are included in Figure 3.2. A belt for the
incipient drag force can be drawn to represent the data. This curve
has the following characteristics:
(a) It has a saddle shape. A minimum value of τc/ (γs – γ) D
occurs for Re* of about 10.
(b) For Re* smaller than 2, τc/ (γs – γ) D is proportional to Re*
with an exponent of –0.3.
(c) If U*D/ ν >10, the incipient drag force increases with the
increase of grain weight. If Re* is larger than 1 000,
τc/ (γs – γ) D has a constant value of about 0.045.
(B) Incipient velocity approach. A relationship between
the velocity field and shear stress field exists. Therefore, if the
drag force for incipient motion is known, the velocity for incipient
motion can be deduced. For instance, if the logarithmic velocity
formula:
U = 5.75U∗ log 12.27
χR
= 5.75
τ0
Ks
ρ
log 12.27
χR
Ks
(3.5)
is adopted and substituted into Equation 3.4, the latter can be
transformed into the following:
U D
χR
= 5.75 f ( ∗ ) log 12.27
Ks
υ
Uc
γs −γ
gD
γ
(3.6)
where R is hydraulic radius Ks is roughness, and χ is the coefficient. For the belt zone in Figure 3.2 with Re larger than 60, the
f(U*D/ν) has a value in the range of 0.03 to 0.06. Hence,
Uc
γs −γ
γ
= (1 ~ 1.4 ) log 12.27
gD
(3.7)
For natural sediment, (γs – γ)/γ may be taken as 1.65, and the
formula is then:
Uc
χR
= (1.28 ~ 1.79 ) log 12.27
gD
Ks
(3.8)
Many formulae for the critical velocity take this form. They are
slightly different because the structure and coefficients of the
velocity formulae they used are somewhat different.
For instance, the Goncharov (1962) formula is:
Uc
γs −γ
γ
= 1.06 log
= 1.4 log
gD
12 R
(3.10)
D90
for R/D90 = 10~40,
Uc
gD
= 1.04 + 0.87 log
gD
10 R
D90
 h
= 1.47
 D
1/ 6
(3.12)
where h is the water depth.
(C) Comparison of the two approaches. Although the
drag force and velocity for incipient motion provide two different
expressions for the same phenomenon and can be mutually transformed from one to another, they represent two study approaches
based on two different concepts. Each has advantages and disadvantages. The following discussion is primarily a comparison of
them.
The incipient motion of sediment is a dynamic process.
The force causing sediment motion, in the final analysis, is the
drag force exerted by flow on the particles. In practical applications, an important advantage of the formula for the critical drag
force is that it can be taken as a constant for a particular flow
condition and for a specific grain size, even though it is a function
of the grain Reynolds number. In contrast, the corresponding critical velocity varies with the grain Reynolds number, and it depends
greatly on the water depth. A serious disadvantage of the drag
force concept is that the slope is included in the formula. Because
the measurement of slope in rivers requires high precision, the
results obtained are less reliable than those based on the average
velocity; the latter is measured regularly at hydrological stations.
Furthermore, the concept of velocity and water depth is easier for
people to visualize. Thus, the concept of critical velocity also has
convenient features.
3.2.1.3
CONDITION FOR INCIPIENT MOTION OF COHESIVE
Sediment finer than a critical size becomes harder and harder to
move because of the cohesion between the finer grains.
In the study of critical conditions for the motion of cohesive sediment, two different cases arise. One is that of
unconsolidated sediment newly deposited during the natural
process of siltation. Another is the cohesive sediment formed
during a long-term process of deposition that has undergone physical and chemical action.
(A) Incipient conditions for newly deposited cohesive
sediment. The forces acting on a particle include weight, drag,
uplift and cohesive force. Several such semi-empirical equations
are as follows:
Tang’s equation for incipient motion of cohesive
sediment:
τc =
(3.9)
and the Levy (1956) formulae are:
for R/D90 > 90,
Uc
Uc
8.8h
D95
gD
The Shamov(1952) formula:
SEDIMENT
χR
Ks
31
(3.11)

3.2(γ s
77.5 

1
 γb 
− γ )D + 

 γ b0 
10


D

k
(3.13)
where γb is the unit weight of sediment on the bed, γb0 is the unit
weight of consolidated sediment (= 1.6 g cm –3 ), and k is a
constant equal to 2.9 × 10–4 g cm–1 The distinguishing feature of
the formula is that the relative consolidation of the sediment on
the bed is included in the cohesive term.
The Wuhan Institute of Hydraulic and Electric
Engineering (1961) equation for incipient motion of cohesive
sediment:
Uc =
 h
 D
0.14
γs −γ

 17.6

γ
D + 6.05 × 10
−7
10 + h 
D
0.72


1/ 2
(3.14)
32
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
The Dou (1960) equation for incipient motion of cohesive sediment:
2
Uc
gd
=
γs −γ
γ
( 6.25 + 41.6
h
h
) + (111 + 740
ha
ha
)
haδ
D
2
(3.15)
where ha is atmospheric pressure in the water column, and d is the
thickness of the water molecule, 3 × 10–8 cm.
Detailed studies on the incipient motion of clayey mud
were conducted by Migniot (1968, 1977). Clayey mud belongs to
the category of Bingham fluid. Migniot found that the incipient
friction velocity is closely related to the Bingham shear stress, as
shown in Figure 3.3. If τB is less than 15 dyne cm–2, the clayey
mud is in a plastic state:
U*c = 0.95τB1/4
(3.16)
For τ B larger than 15 dyne cm –2 , the clayey mud
becomes consolidated,
U*c = 0.50τB1/2
(3.17)
3.2.2.1 DEVELOPMENT OF BED FORMS
With increasing flow velocity, bed forms will experience several
different stages, as shown in Figure 3.4.
Soon after some particles are in motion, a few particles
may gather on the bed and form a small ridge; this ridge gradually
moves downstream and tends to increase in length. Finally, the
ridges connect with each other and ripples with a regular shape
form, as shown in Figure 3.4(b). The longitudinal cross-sections of
ripples are usually not symmetrical. The upstream face is long and
has a gentle slope, and the downstream face is short and steep. The
former is generally between 2 and 4 times as long as the latter.
Ripple height is usually between 0.5 and 2 cm; the highest ripple
is not more than 5 cm. The wave length normally does not exceed
30 cm, and they are usually within the range of 1 to 15 cm.
With increasing flow velocity, ripples develop further and
eventually become dunes (Figure 3.4(c)). Dune size is closely related
to water depth. Figure 3.5 shows that the heights and lengths vary
significantly in different rivers (Chien and Wan, 1983).
Incipient friction velocity (cm s–1)
Elevation (m)
(B) Incipient motion of consolidated cohesive sediment. The cohesion among clayey grains is quite complicated.
Knowledge in this respect is still limited. Up to now, no property of consolidated cohesive sediment has been found to
estabish a good relationship with incipient shear stress or incipient velocity.
3.2.2
Bed form and resistance in fluvial streams
The bed of an alluvial stream changes with flow conditions. In the
Fox example in sand bedded rivers, when sediment particles are
set in motion, ripples form on the bed. With the change of flow
conditions, different bed forms appear, such as dunes, flat bed and
sand waves, etc. Different bed forms have different roughnesses of
bed surface, consequently, this changes the resistance to flow and
affects the flow and sediment transport accordingly. Variations of
bed form and resistance are the main characteristics of fluvial
streams, and they should be studied in depth.
Elevation (m)
Distance (m)
(a) Nanjing Reach, Yangtze River
Distance (m)
(b) Volga River
Bingham yield stress τB (dyne cm–2)
Figure 3.3 — Relationship between incipient friction velocity of clayey
mud and Bingham yield stress (after Migniot).
Distance (m)
(c) Mississippi River
Distance (m)
(d) Klaralven River
Distance (m)
(e) Huayuankou Reach, Yellow River
Figure 3.4 — Various phases of bed form development.
Figure 3.5 — Longitudinal profiles of dunes in various rivers.
CHAPTER 3 — SEDIMENT TRANSPORT IN RIVERS
If the dune reaches a certain height and the flow velocity
is then increased further, the dune decays; its wave length
increases and its height gradually decreases to the form shown in
Figure 3.4(d). With still further increases in velocity, the bed
becomes flat again (Figure 3.4(e)).
The sediment transport rate is quite high in the second
flat bed phase. If the velocity continues to increase, the flow
approaches or becomes supercritical (Froude number of about
unity or even larger), and the bed forms a sand wave
(Figure 3.4(f)). A sand wave is a type of bed configuration that is
in phase with the wave on the water surface, and these two waves
interact strongly. The differences between a sand wave and a dune
are as follows. The shape of dune is non-symmetrical, and the
streamlines of the flow separate at the dune peak; in contrast, a
sand wave is symmetrical, more like a surface wave; the streamlines are almost parallel to the river bed and no separation occurs.
Sand waves can move either in the same direction as the
flow, as do ripples and dunes, or in the opposite direction. The
former is called a “downstreamward sand wave” and the latter is
called a “upstreamward sand wave” or “antidune”. Antidunes
often form in shallow flows that are moving at high velocities.
Even though the sand wave as a whole profile moves upstream,
the movement and transport of every particle is in the direction of
the flow.
In the development of antidunes, the amplitude of the
surface wave may exceed that of the sand wave by a factor of 1.5
to 2. The trough of surface waves can even be below the crest of
the sand waves (Figure 3.6) (Simons and Richardson, 1960). In
this instance, the waves on the water surface are unstable and
break (Figure 3.4).
33
3.2.2.2 FLOW RESISTANCE IN ALLUVIAL STREAMS
As discussed above, with the change of flow conditions, various
bed configurations form on the bed surface of alluvial streams.
During the ripple and dune phases, the flow separates at the crest
of the bed form so that the pressure on the downstream and
upstream sides differ. The net force thus produced is the form
resistance. During the sand wave phase, the undulation of the sand
bed is much more pronounced than it is in the ripple and dune
phases, but the sand waves have a symmetrical shape with no flow
separation at their crests. Therefore, their form resistance is
smaller and the energy loss is less than that for ripples and dunes.
The corresponding energy loss is only a little more than that for a
plane bed, because the breaking of the sand waves generates a
strong local turbulence that dissipates parts of the flow energy. As
a major component of resistance, form resistance changes as the
flow conditions change. Hence, the friction factor in an alluvial
river is not just a constant, but varies with flow conditions.
Einstein (1950) suggested that the resistance of an alluvial stream consists of bed resistance and bank resistance.
Furthermore, the bed resistance consists of grain friction and bed
form resistance. Although grain friction and bed form resistance
both act on the bed surface, the ways in which they affect the
movement of bed material are different. The formation of bed
form resistance is the result of the separation of flow at the peaks
of sand waves and the unsymmetrical distribution of pressure on
the stoss and lee faces. The turbulence created by bed form resistance occurs mainly in the separation on the lee face, and it occurs
at some distance from the bed grains. The role of the eddy created
by bed-form resistance on bed load movement is thus not as direct
as that from the grain friction. The eddy created by the corresponding flow potential energy from grains on the channel bed
plays a large role in the transportation of bed material for grain
friction only.
A. Einstein’s approach. The bed roughness of an alluvial
channel consists of two parts, namely, grain roughness or skin
roughness due to the sediment particle size, and form roughness
due to the existence of bed forms. According to Einstein, the shear
stress or drag force acting along an alluvial bed can be divided
into two parts, i.e.,
τ = τ' + τ''
(3.18)
= γJ (R'b + Rb'')
Figure 3.6 — Antidune on the verge of breaking (after Simons and
Richardson).
If the velocity is higher than that for which sand waves
form, the undulating bed resembles that of a mountain stream,
with chutes and pools. The flow is supercritical at the chutes and
subcritical in the pools. The transition from supercritical flow to
subcritical flow is achieved through a hydraulic jump
(Figure 3.4(h)), and the entire bed form migrates slowly upstream.
Severe erosion occurs at the chutes, and the sediment particles
eroded from these regions are deposited in the pools. In natural
rivers on plains, the velocity is seldom high enough for this
phenomenon to occur.
In ordinary rivers, the most common bed features are
ripples and dunes. Sand waves, chutes and pools occur much less
often. In natural rivers, the process described above may not occur
in a normal progression; various types of bed forms can exist at
the same time, and the process of development may differ from
one instance to another.
where τ is the total drag force acting along an alluvial bed, τ' and
τ'' are the drag force due to grain roughness and form roughness,
respectively, γ is the specific weight of water, J is the energy or
channel slope, and R'b and Rb'' are the hydraulic radii due to grain
roughness and form roughness, respectively.
The grain friction denotes the resistance to a two-dimensional flow, which is not affected by side banks, with a plane bed.
The grain friction can be described by the following equation:
R 'χ
U
= 5.75log (12.27 b )
U*
Ks
(3.19)
where R'b is the hydraulic radius due to grain friction, Ks is a
representative roughness, which is taken as D65, the particle size
of bed material of which 65 per cent by weight is finer, by
Einstein, χ is a function of Ks/δ, and δ = 11.6υ/U* the thickness of
laminar sublayer U* = gRb' J. The relationship between χ and Ks /δ is
shown in Figure 3.7.
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
Per cent
34
stationary sand
wave and flat bed
Figure 3.9 — Relationship between grain friction and total bed
resistance (after Engelund and Hansen).
Figure 3.7 — χ versus Ks/δ.
Based on data from 10 rivers in the United States,
Einstein and Barbarossa (1952) established a relationship for bed
form resistance U/ U*'' = F(Ψ’), as shown in Figure 3.8. Where:
Ψ ′=
γ s − γ D35
γ Rb′ J
(3.20)
where D35 is the particle size of sediment of which 35 per cent by
weight is finer, R ' b the hydraulic radius due to grain friction,
U*′′ = gRb′′ J , and R '' b is the hydraulic radius due to bed form
'/U''*
resistance. With an increase of flow intensity, i.e. a decrease of ψ ',
dunes tend to diminish, and the dune resistance decreases
correspondingly.
Among the 10 rivers analysed by Einstein and
Barbarossa, eight had values of D35 smaller than 0.5 mm, and the
other two had D35 values of 0.7 and 1.0 mm. In later experiments
with coarser bed materials, the result departed from the mean
curve of the 10 rivers of the United States, as shown in Figure 3.8.
The bed form resistance for coarse sand was shown to be smaller
than that for medium and fine sand.
The following procedures are for the computation of
total hydraulic radius due to grain and form roughness when the
water discharge and bed material are given, or vice versa.
1. Assume a value of Rb'.
2. Apply Equation 3.19 to determine U by R'b and D65 (= Ks).
3. Compute Ψ ' using Equation 3.20 and the corresponding
value of U/U*'' from Figure 3.8.
Missouri River near Fort Randell, S.D.
Missouri River near Pierre, S.D.
Missouri River near Omaha, Nebr.
Elkhorn River near Waterloo, Nebr.
Big Sioux River near Akron, Iowa
Platte River near Ashland, Nebr.
Niobrara River near Butte, Nebr.
Salinas River at San Lucas, California
Nacimiento River near Junction, Calif.
Salinas River at Paso Robles
(easily vegetated)
Figure 3.8 — Relationship between bed form resistance and flow parameter compared with experiment (after Einstein and Barbarossa).
Compute U*'' and the corresponding value of Rb''.
Compute R b = R b' + R b'' and the corresponding channel
cross-sectional area A.
6. Verify using the continuity equation Q = UA. If the
computed Q agrees with the given Q, the problem is solved.
Otherwise, assume another value of Rb' and repeat the procedure until agreement is reached between the computed and
the given Q.
B. Engelund and Hansen’s approach (1972). The relationship between grain friction and total resistance shown in
Figure 3.9 is considered to be reliable and is popularly used.
According to Engelund and Hansen, the shear stress of
drag force acting along an alluvial bed can be divided into two
parts, i.e.,
τ = τ' + τ''
(3.21)
4.
5.
or
γhJ = γh (J' + J')
(3.22)
where J' and J'' are energy loss or friction slope due to grain friction and that due to bed form, respectively. Engelund thinks the
two expressions of grain friction are equivalent and interchangeable.
τ' = γhJ' = γh'J
(3.23)
Divided by (γs – γ)D, Equation 3.22 turns into:
Θ = Θ' + Θ''
where
Θ=
hJ
γs − γ
D
γ
(3.24)
(3.25)
This parameter is simply the inverse of the Einstein flow
parameter, Ψ.
The abscissa is the parameter for the flow intensity due
to grain friction:
h' J
Θ' =
γs − γ
(3.26)
D
γ
And the parameter for the flow intensity related to bed form Θ'' is:
h" J
Θ "=
= Θ –Θ '
(3.27)
γs − γ
D
γ
CHAPTER 3 — SEDIMENT TRANSPORT IN RIVERS
35
The lower branch in Figure 3.9 corresponds to the dune phase.
Θ ' = 0.06 + 0.4Θ2
(3.28)
As Θ decreases, Θ ' gradually approaches the constant value of
0.06, which corresponds to the condition of incipient motion. If
Θ >0.4,
Θ ' = 0.4Θ2
(3.29)
In contrast, for high transport rates of sediment and with sand
waves forming on the bed, the data fall near the other curve. For
flat bed or for stationary sandwaves without local enlargement
loss,
Θ' = Θ
(3.30)
But in the sand wave phase, as a result of the additional energy
loss caused by the breakage of the water surface, Θ ' is smaller
than Θ. Engelund was able to express the resistance losses for all
phases of bed configuration, except for the ripple phase, in a
single figure.
The following procedure is to determine a stagedischarge relationship by using Figure 3.9.
Step 1: Determine J and h from a field survey of slope and
channel cross-section.
Step 2: Compute Θ from Equation 3.25 for the given sediment
size D.
Step 3: Determine Θ' from Figure 3.9 with Θ from Step 2.
Step 4: Compute h' from Equation 3.26.
Step 5: Compute U from Equation 3.19. In the case of two
dimensional flow, Rb'= h'. Correspondingly, Rb'' = h'',
Rb = h.
Step 6: Determine the channel cross-sectional area A corresponding to the h value selected in Step 1.
Step 7: Compute Q = UA. The stage-discharge relationship can
be determined by selecting different h values and repeating the processes.
Figure 3.10 — Comparison of the Meyer-Peter formula with measured
data.
where Kb is a coefficient for bed resistance, Kb' is the roughness
coefficient due to grain resistance, gb is the rate of bed load transport per unit width by dry weight, and a and b are constants. The
formula was calibrated against measured data, as shown in Figure
3.10, in order to determine the two constants.
The Meyer-Peter formula is based on a large quantity of
experimental data. The main variables in the experiments varied
within the following ranges:
Width of flume:
0.15–2 m
Flow depth:
0.01–1.2 m
Energy slope:
0.04–2 per cent
Density of sediment:
1.25–4 g cm–3
Diameter of sediment:
0.40–30 mm
The Meyer-Peter formula is more reliable than some
others for rivers carrying coarse sand and gravel. It has been widely
used and the results obtained from it are generally satisfactory.
(b) Einstein bed load theory. Einstein noticed the stochastic
nature of bed load motion and combined statistics with
modern fluid mechanics. Applying probability theory and
making some hypotheses, Einstein (1950) derived a mathematical expression for the relationship between the bed load
transport intensity Φ and the flow parameter Ψ:
1−
1
π
∫
B*ψ −1 / η 0
e
−t
2
− B*ψ −1 / η 0
FORMULAE WITH SHEAR STRESS AS THE MAIN PARAMETER
(a) The Meyer–Peter formula. Based on the data of a great
number of experiments, Meyer-Peter (1934, 1948) developed
the following bed load formula by isolating involved parameters one by one.
γg Qb
Q
K
 b
 Kb′



3/2
1/3 
γ 
hJ = a 4 (γ s − γ ) D + b4  
g
γ −γ 
 s

 γs 
2 /3
gb2 / 3 (3.31)
where Q (= BhU) is the total discharge through the cross-section,
and Qb is the part of the discharge pertaining to the bed:
Qb = BRbU
(3.32)
A*Φ
1 + A*Φ
(3.33)
where
1
3.2.3
Bed load transport
3.2.3.1 TRANSPORT OF UNIFORM BED LOAD
A number of formulae for bed load transport have been proposed
by scientists. These formulae are based on different modes of
motion and employ different parameters, including shear stress
and flow velocity. Several representative ones are briefly introduced, as follows.
dt =
Φ=
 γ   1 

  3
γ s  γ s − γ   gD 
gb
ψ =
2
γ −γ
D
γ
Rb J
s
/
1
2
(3.34)
(3.35)
where the constants, determined through experiments, are as
follows:
(3.36)
1/η0 = 2.0
A* =
1
0.023
B*=1/7
(3.37)
(3.38)
Figure 3.11 is a comparison of the function with
measured data, and it shows that the function represents the data
quite well.
(c) The Engelund formula. Engelund and Fredsφe (1976) treated
sediment particles as spheres of diameter D, so that there are
approximately 1/D2 spherical particles in a unit area of the
bed surface. For a certain flow intensity, the proportion of
36
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
Y =
γ SWb h
(3.45)
γ sD
where SWb represents the average concentration of sediment load
in weight per unit volume.
BED LOAD FORMULAE WITH VELOCITY AS THE MAIN PARAMETER
In the former USSR, scientists employed the average velocity
instead. These formulae can be rewritten in a more general form as
follows:
gb = gshbSvbu–b
(3.46)
the particles on the bed surface that are moving is p. The
mean velocity of the bed load particles is ub. Hence, the rate
of bed load transport gb is given by:
g
b
=
π
6
p
3
D γs
D
2
ub
(3.39)
Based on the balance of forces acting on particles moving as bed
load, the following equation can be derived:
us
U*

Θc

Θ
= α 1 − 0.7



(3.40)
where Θc is the Θ value at the incipient motion of particles, α is a
constant for a sandy river bed, and α = 9.3. They deduced the
conclusion that the proportion of the particles on the bed surface
that are moving, p, is:
6
p=
(Θ − Θ c )
(3.41)
πβ
where β is a kinetic frictional coefficient. Combining
Equations 3.39 and 3.40 with Equation 3.41, Engelund and
Fredsφe obtained their bed load formula:
gb =
9.3 Dγ s
β
Θ
U* ( Θ − Θ c )( Θ − 0.7 Θ c )
(3.42)
Based on data from flume experiments, Θc = 0.046, and β = 0.8.
(d) The Ackers-White formula. Ackers and White (1973)
collected 1 000 sets of experimental data from previous
researchers. Following Bagnold’s approach, they derived a
functional relationship between dimensionless parameters.
Then they conducted a regression analysis with the data to
determine a functional relationship. Their formula includes
both bed load and suspended load. Nevertheless, the formula
was simplified into a bed load formula for natural sand
coarser than 2.5 mm, in the following form:
Y = 0.025
where
 M − 1
 0.17 
1.5
U
M=
g
γs −γ
γ
(3.43)
where hb stands for the thickness of bed load layer and Svb for the
volume concentration of bed load in the layer.
Different researchers made various assumptions about
hb, Svb and ub, and thus they obtained different formulae. Table 3.1
presents three representative ones. The three share much in their
approach, although they differ in details.
COMPARISON OF BED LOAD FORMULAE
Having thoroughly analysed the bed load formulae presented in
the preceding section, Chien (1980) pointed out that these formulae have common properties and give similar results under certain
conditions, even though they have different forms.
The comparison is based on the following conditions:
First, the channel bed is flat and the characteristic roughness is the
sediment diameter. Second, except for the Ackers-White formula,
the threshold condition of the initiation of bed load motion is
taken as Θc = 0.047.
Figure 3.12 shows a comparison of the Meyer-Peter,
Bagnold, Einstein and Yalin formulae. The figure shows that
for Ψ > 2, the Bagnold, Einstein and Meyer-Peter formulae are
close together, but the Yalin formula yields smaller values for the
bed load transport rate. Figure 3.12 also shows that for low intensity of bed load transport, the Φ – Ψ curves slope gently. That is,
if Θ << 1, a slight variation in Θ responds to great change in Φ.
This trend is more apparent if Θ is close to Θc. In other words, bed
load transport is quite sensitive to the flow if the transport intensity is low. Figure 3.13 also shows that the bed load formulae
diverge for Ψ < 2. The Φ – 1/Ψ curves in this range approach
straight lines on a log-log plot. The Meyer-Peter, Bagnold, Yalin
Meyer-Peter
Bagnold
Yalin
Einstein
Ψ (Θ )
Figure 3.11 — Comparison of Einstein bed load function with
measured data (uniform sediment) (after Einstein).
Symbol
Material
Diameter
(mm)
Gravel
28.65
Sand
5.20
Brown coal
5.20
Bary grains 5.20
Sand
0.785
Plastic 4.75 × 3.18 × 2.38
Sand
D42 – 1.26
Plastic
3.88
Specific
gravity
2.68
2.68
1.25
4.22
2.68
1.052
2.68
1.13
Author
Meyer-Peter
Gilbert
Chien
Wilson
1
D
32
10 h
D
(3.44)
Φ
Figure 3.12 — A comparison of Meyer-Peter, Einstein, Bagnold and
Yalin formulae.
CHAPTER 3 — SEDIMENT TRANSPORT IN RIVERS
and Engelund formulae approach lines indicating an exponent of
1.5 on the (1/Ψ) term. In contrast, the Einstein bed load function
approaches the line:
37
Slightly non-uniform natural sand
Extremely non-uniform natural sand
Slightly non-uniform plastic material
(a) Dm as the
representative dameter
Non-uniform natural sand
Φ=
7.9
(3.47)
Ψ
(a) Dm as the
representative dameter
and the exponent is therefore 1. The exponent for the AckersWhite formula is 1.35–1.45, a value that falls between the other
two values. A serious difficulty arises from the suspension of the
material with high intensity bed load transport; in this case, one
cannot readily separate suspended load from bed load. So far, the
data for high intensities of bed load motion are insufficient, therefore one cannot conclude which formula is the best one to use.
3.2.3.2 TRANSPORT OF NON-UNIFORM BED LOAD
The foregoing is only for uniform sediment. However, sediment in
natural rivers is always non-uniform. Two techniques are used to
deal with non-uniform bed load motion. If only the total bed load
transport rate is required, the bed load formulae introduced in
section 3.2.3.1 can be used directly, but a representative diameter
must be determined. If instead the transport rates of various diameters are required, the mutual effects of the various particle sizes
must be studied.
DETERMINATION OF REPRESENTATIVE DIAMETER FOR
CALCULATING TRANSPORT RATE OF NON-UNIFORM BED LOAD
Einstein found from data measured in both small streams and
flume experiments that D35 can be used as the diameter in the bed
load formulae. D35 stands for the diameter for which 35 per cent
of the bed material is finer. Meyer-Peter (1948) suggested another
form of representative diameter:
Einstein bed load function
Meyer-Peter formula
Figure 3.13 — Comparison of measured bed load transport rates for
non-uniform sediment with results calculated using different
representative diameters.
Dm =
∑ D ∆p
i
i
(3.48)
100
where ∆pi stands for the percentage of particles of diameter Di in
the bed material. Chien examined the two representative diameters, with the results presented in Figure 3.13.
The results show that Dm is preferable to D35 for low
intensities of bed load motion, but no difference was found
between the two for high intensities.
BED LOAD TRANSPORT RATES OF VARIOUS GRAIN SIZES
Many engineering situations require not only calculating the total
bed load transport rate but also the transport rates of the various
Table 3.1
Bed load formulas with velocity as the main parameter
Author
u–b
hb
Svb
gb (kg/m/s)
Valid range
 
U 

0.95 D 
 Uc 
 
1.2
3
Sharmov
(1959)
Levy (1957)
 U − Uc   D 
 1.2   h 
 
U 
K ′′
 Uc 
 
 1.2 
1/ 4
K'D
α' (U – Uc)
3
 U 3

 gD 


α''D
α ′′′
D 
⋅  
h
1/4
Uc
Goncharov
(1962)
  Uc 
  1.4 
α k U 1 −
3
U


3





(α1 + αζ )D
U−
⋅
Uc
1.4
Uc
1.4
1 + α 6 1.4
U
α4
Uc
α3
1.4
1+
ζU
⋅
U
2
 U c 2
 
1.4 

⋅ U −

U c  D 
1/4
 
1.2  h 
 U 3

 gD 


2D
Note
0.2 < D < 0.73 mm,
and 13 < D < 65 mm
1.02 < h < 3.94 m
0.18 < h < 2.16 m
0.4 < U < 1.02 m s-1
0.8 < U < 2.95 m s-1
K' and K'' are
function of D.
Therefore the
formula includes
the square root of
D instead of D.
0.25 < D < 23 mm
5 < h/D < 500
1 < U/Uc < 3.5
______
D 
⋅ (U − U c ) 
h
1/4
(3.0 – 5.3)(1 + ζ )D




 U3

⋅ 
− 1
3
 U 

 c 

 1.4 

 Uc 
⋅ U −

 1.4 
0.08 < D < 10 mm
10 < h/D < 1 550
0.72 < U/Uc < 13.1
The coefficient 3
is suitable for
river flow and 5
for flumes; ς is a
coefficient related
to turbulence.
38
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
grain sizes of non-uniform sediment. For example, both the fining
process in the upstream reaches of a dam and an armouring layer
in the downstream reaches require such a calculation.
Few researchers have studied the movement of various
sizes of non-uniform sediment because the mutual interactions
between the various sizes are complex. Some results of Einstein
(1950) and Chien are presented here.
The following formula was obtained by Einstein:
1−
1
∫
B*ψ* −1 / η 0
− B*ψ* −1 / η 0
e
−t 2
dt =
A*Φ*
1 + A*Φ*
π
and it is suitable for various groups of grain sizes:


i0


 β2  
Ψ* = ξY 
 β 2 Ψ
 X 
Φ* =
ib
(3.49)
Φ

X

∆
β x = log10.6

(3.50)
(3.51)
where i0 and ib are the percentage of sediment with a size D in bed
load and that in bed material, respectively. Y, a function of
∆ = Ks/χ as shown in Figure 3.14, is a correction factor for lift
force, and X is the maximum grain size subject to the hiding effect
in a sediment mixture.
Under the conditions of a rough bed, i.e., ∆/δ >1.8:
X = 0.77∆
(3.52)
Under the conditions of a smooth bed, i.e., ∆/δ <1.8:
X = 1.39δ
(3.53)
β = log 10.6 is a constant. ξ , a function of D/X as shown in
Figure 3.15, is a factor concerning the hiding effect for particles
finer than X.
The procedure for the computation of bed load of different grain sizes from Einstein’s bed load transport function is as
follows:
Step 1: From the given bed material and flow condition, compute
Ψ* from Equation 3.50. The values of ξ and Y can be
determined from Figures 3.14 and 3.15. The value of βx
can be determined from Equations 3.51 and 3.52.
Step 2: From Figure 3.11, determine Φ*.
Step 3: Bed load by weight per unit width of a given size ibgb can
be computed from Equations 3.48 and 3.50.
Step 4: Repeat the preceding steps for each size fraction and get
ibgb for each size fraction.
Step 5: Sum up the results over the size range for bed material
and get a total bed load.
TRANSPORTATION OF EXTREMELY NON-UNIFORM SEDIMENT
Einstein and Chien (1953) carried out experiments with sediments
with a wide size range. Their experiments revealed that the bed
material was sorted by the flow, with large and small particles
gathering at different places. If sorting occurs, coarse particles are
covered by a layer of fine sediment, so that the coarse sand shielding zones are much fewer and the sheltering effect on the
movement of fine particles is also less. Such an effect is considered by introducing a factor θ, which is a function of the grain
Reynolds number, into the lift force. And after some modifications, the flow parameter Ψ* is redefined as:
Ψ∗ =
(
ξY β / β X
θ
)
2
Ψ
(3.54)
Details can be found in the references (Einstein and
Chien, 1953).
Ks/δ
D/X
Figure 3.14 — Y – Ks/δ.
ξ
Figure 3.15 — Hiding effect on movement of fine particles in nonuniform sediment.
3.2.3.3
CHARACTERISTICS OF TRANSPORT OF GRAVEL BED LOAD
FLUCTUATION AND BURSTING
Sediment transport is a stochastic phenomenon, and the stochastic
characteristics of gravel bed load are even more obvious. Under
almost unchanged flow conditions, the transport rate of gravel bed
load might vary within a wide range.
A parameter ∆ξ is used to denote the variation of gravel
bed load:
∆ξ = (gbmax – gbmin)/gb
(3.55)
where gbmax is the measured temporal value corresponding to a
95 per cent frequency of varying gravel bed load, gbmin is the
measured temporal value corresponding to a 5 per cent frequency
of varying gravel bed load, and gb is the average value of varying
gravel bed.
According to field data obtained from gauging stations
along the Upper Yangtze River, ∆ξ varies within the range of 5 to
8. In Tan’s (1983) paper, field data obtained from Inner River,
Dujiangyan, were cited. Under conditions with almost constant
CHAPTER 3 — SEDIMENT TRANSPORT IN RIVERS
39
U (m s-1)]
4b [kg (s m-1)]
coarsening process will stop once a layer of coarse material
completely covers the streambed and protects the finer materials
beneath it from being transported. After this process is completed,
the streambed is armoured and the coarser layer is called the
armour layer. A definition sketch of armouring is shown in
Figure 3.17. From this,
Ya = Y – Yd
where Ya is the thickness of the armour layer, Y is the depth from
original streambed to the bottom of the armouring layer, and Yd is
the depth from the original streambed to the top of the armouring
layer or the depth of degradation.
Based on the definition of armouring layer thickness,
Dmin (mm)
Elevation (m)
Ya = (∆p)Y
Yd = Ya (1/∆p – 1)
SELECTED EROSION
In his study of incipient motion, Gessler (1971, 1972, 1976)
considered the effect of flow fluctuations. In turbulent flow,
and also
τ '0 τ c
<
−1
τ0 τ0
sediment cannot be entrained.
τc/τ0
Figure 3.17 — Definition sketch of streambed armouring.
€
(3.59)
where τ0, τ–0 and τ'0 are the instant, temporal average and fluctuating value of the drag force exerted by the flow on the bed,
respectively. If:
τ0 < τc
(3.60)
Probability (q)
ARMOURING PROCESS (YANG, 1997)
When the sediment transport capacity of a channel exceeds the
rate of sediment supply from upstream, the channel may be
degraded. Because of the non-uniformity of the bed material size,
finer materials will be transported at a faster rate than the coarser
materials, and the remaining bed material becomes coarser. This
(3.58)
The required armour layer thickness varies with the size
of the armouring material. Usually, two to three armouring particle
diameters or 0.5 ft, whichever is smaller, should be sufficient.
τ0 = τ–0 + τ'0
LATERAL DISTRIBUTION OF GRAVEL BED LOAD DISCHARGE
Owing to the uneven distribution of flow, bed material and incoming sediment, the transport of gravel bed load in the lateral
direction is uneven too. In many cases, the transport of gravel bed
load is limited to a certain zone within the full width, as shown in
Figure 3.16. The region where the transport of gravel occurs is
sometimes called the belt of gravel transport. The belt of gravel
transport changes with variations in flow conditions, as shown in
Figure 3.16. The highly three-dimensional characteristics of gravel
transport should be noticed while dealing with field data.
(3.57)
where ∆p is the decimal percentage of material larger than the
armouring size.
From Equations 3.56 and 3.57,
Figure 3.16 — Lateral distribution of gravel transport.
discharge, the rates of gravel bed transport varied greatly. The
ratio of rates of gravel bed load of two neighbouring measurements was taken as an index: more than half of the indices were
larger than 5. The maximum ones reached 500 to 700.
Consequently, field data of gravel bed load should be treated carefully. In order to avoid inaccurate results, a long-term series of
frequent measurement may be required.
(3.56)
Figure 3.18 — q versus τc/τ0.
(3.61)
€
40
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
From numerous flume experiments, Gessler concluded
that the fluctuation of drag force follows a normal error distribution. Thus, the probability of sediment staying on the bed is:
q=
∫
1
σ 2π
τc
τ0
−∞
−1
 x2
exp 
2
 −2σ

dx


(3.62)
where σ is the standard deviation of τ ' 0 / τ– 0 ; the relationship
between q and τc/τ0 is shown in Figure 3.18.
Because the probability of incipient motion of the sediment with grain diameter D is (1-q), the size distribution curves of
the bed materials washed away and remaining on the bed can be
derived. If the maximum and minimum grain diameters of the
original bed material are known and the weight percentage of
grain with diameter of D is p0 (D), the accumulated percentage of
the sediment with a diameter less than D is:
∫D
Dmin p0 (D)dD
(3.63)
For the armouring layer of the bed after scouring, the frequency of
grains with diameter D is:
pa (D) = C1qp0dD
(3.64)
where the coefficient C 1 can be determined by means of the
following equation:
∫DDmax
p (D)dD = 1
min a
(3.65)
Thus, the sediment size distribution of the armouring layer is:
∫
∫
D
D min
D max
D min
qp0 ( D ) dD
(3.66)
qp0 ( D ) dD
∫
∫
D min
D max
D min
3.3
Diffusion equation and vertical distribution of
suspended sediment
In turbulent flow, the movement of water elements, and how they
change positions between water layers, also causes sediment
exchanges between the layers. At the same time, sediment particles, because of their greater specific weight, tend to settle and
move toward the bed. As a result, the sediment concentration is
greater near the bed than it is at a point some distance above the
bed. Because of this variation in concentration, water elements
moving upward carry a greater amount of sediment than the water
bodies moving downward. Thus, the exchange between the
upward and downward water elements of the same volume results
in a net transport of sediment in the upward direction. The amount
of the upward sediment flux per unit horizontal area is proportional to the concentration gradient dS v /dy and is written as
–εydSv/dy. The amount of the downward sediment flux per unit
horizontal area due to settling is written as ωSv. Here, only the
vertical concentration profile of suspended load carried by a twodimensional flow in a state of equilibrium is studied. Under such
conditions, the upward sediment flux equals the downward sediment flux:
εy
dSv
dy
+ Svω = 0
(3.68)
where ω is the fall velocity of sediment particles, Sv is the sediment concentration in volume, and εy is the sediment exchange
coefficient. In order to solve the differential Equation 3.68, one
must determine the vertical distribution of εy. The simplest procedure is to assume that it is a constant. The solution is then:
=e
−ω ( y − a ) / ε y
(3.69)
Sva
(1 − q ) p0 ( D ) dD
(1 − q ) p0 ( D ) dD
3.3.2
Sv
and the size distribution of the sediment washed out is:
D
portion of the particles that are lifted up. In this way, some sediment
is kept in suspension. However, in the process a continuous
exchange occurs between the suspended sediment and the sediment
in the near-bed region.
(3.67)
Table 3.2
Observed phenomena and related mechanism of particle
suspension for a smooth bed
Phenomenon
Mechanism
Sediment particles are
lifted up from the bed
Sediment in the near-bed region is
picked up and lifted by the upward
moving low-speed band of flow
The highest position is
reached by particles
being lifted
Sediment reaches its highest
position as the burst breaks up
Particles start to fall
Particles are entrained by water
bodies with large momentum,
and swept away
SUSPENDED SEDIMENT TRANSPORT
3.3.1
Mechanism of sediment moving in suspension
Suspended sediment transportation is closely related to the turbulent
bursting phenomenon. The following frames show the observed
phenomena and the related mechanism of particle suspension for a
smooth bed (Table 3.2). If the low-speed streak of flow near the bed
is lifted due to a burst of turbulence, the sediment there is carried
upward. If the fall velocity of a particle is large, the particle will
quickly fall back to the bed. Such particles are part of the saltation
load. If, in contrast, the fall velocity is small, the sediment can be
carried upward along with the low-speed water element until the
latter breaks up; at that moment the sediment has reached its highest
position and begins to settle back down. As the particles fall, some
of them, caught in the downward moving part of the high-speed
streak of flow, will return to the near-bed region, while others,
caught in an upward-moving eddy, are lifted again. The higher the
turbulence intensity and the smaller the particle size, the greater the
Some particles fall into
the near bed region
Other particles are lifted
up again before entering
the near bed region
As a high-speed region of the flow
reaches the bed, it spreads toward
both sides (in the z direction), and
carries sediment into the
neighbouring low-speed region
Sediment falls into another eddy
that is moving upward
CHAPTER 3 — SEDIMENT TRANSPORT IN RIVERS
where S va is a reference concentration of the suspension at
distance a above the bed. The experimental results of Hurst agree
well with the formula in Equation 3.69. Later on, Rouse obtained
the results shown in Figure 3.19 using a series of grids moving in
simple harmonic motion in a cylinder. The figure shows that the
experimental results essentially follow the theoretical curve for all
except the coarsest particles. Lane and Kalinske made analyses of
data from a natural river and found that Equation 3.69 also gave
satisfactory results for this practical case. They suggested that εy
could be expressed as follows:
εy =
κU* h
(3.70)
where τ0 stands for the shear stress at the bed. For a logarithm
velocity profile:
u
1
κ



 y0 
ln
(3.71)
The sediment exchange coefficient is not nearly a
constant, but is a function of position in space. From the theory of
turbulent flow, the diffusion coefficient is equivalent to the
momentum exchange coefficient εm, and it is related to the velocity gradient in the following way:
τ
εm =
ρ
du
du
dy
=
U∗ 1
κU* y

y

h

(3.74)
Ln
Sv
Sva
=
ω (y − a)
εy
(3.77)
h
h − y dSv
h
+ Svω = 0
dy
(3.78)
after integration, this gives the vertical concentration profile of
suspended load:
Sv
For simplicity, one assumes that:
τ = τ 0 1 –
h−y
By integrating this expression and taking the average, one obtains
Equation 3.70.
The substitution of Equation 3.77 into Equation 3.68
yields:
Sva
For two-dimensional flow, the shear stress is linearly distributed
along the depth, so that:
(3.76)
κ y
ε y = ε m = κU* y
dy
(3.73)
(3.75)
Then, substituting Equations 3.74 and 3.76 into Equation 3.72
yields:
(3.72)
εy = εm
y
Differentiating, one obtains:
6
εy = 0.067U*h
ω (y − a)
εy
=
U*
where κ is the Karman constant in the logarithmic formula for the
velocity distribution. If the usual value of κ = 0.4 is taken, then:
ω=
41
where
h − y
=
 y
z=
z

h − a
a
(3.79)
ω
κU*
(3.80)
For a dune-covered bed, and in the absence of more
experimental data for U * , Einstein suggested that U * can be
replaced by the shear velocity relevant to grain friction
U'*= (Rb'gJ)0.5.
The exponent z in the expression for suspended load
affects the distribution of the sediment concentration. Figure 3.20
compares the relative vertical distributions of suspended load
concentration obtained from Equation 3.79. The figure shows that
a smaller value of z results in a more uniform distribution. Thus,
the height of the suspension is also a function of z. In the case of
z = 5, the amount of sediment carried in suspension is very small;
1/4 mm
1/8 mm
1/16 mm
1/32 mm
Sv/Sva
Figure 3.19 — Vertical distribution of sediment concentration for
various particle sizes in a sediment mixture (Rouse experiments with
uniform stirring) (after Rouse).
Figure 3.20 — Relative distribution of suspended load obtained from
the diffusion theory (after Rouse).
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
the discharge ratio of suspended load to bed load is then 1:4,
according to an estimation based on the Einstein method. From
the practical point of view,
κU*
=5
3.3.3
Transport rate of suspended load
If vertical profiles of both the concentration Svy and the velocity uy
are known, the discharge of suspended load passing through a
cross-section of unit area at y per unit time is uySvy; integration of
uySvy over the depth yields the discharge of suspended sediment
per unit width. In practical applications, one difficulty remains
because the diffusion theory gives only a relative quantity of sediment concentration. From Equation 3.79, the concentration at any
position remains unknown unless Sa, the concentration at the
reference position at distance a above the bed, is known.
Another difficulty is that the upper and lower limits for
the integration need to be determined. The simplest way is to integrate from the bed to the free surface to get the total sediment
discharge. But both velocity and sediment concentration approach
infinity at y = 0 according to the logarithm velocity distribution
formula and Equation 3.79.
Here the Einstein (1950) method of dealing with these
two difficulties is introduced. According to Einstein’s concept, the
region near the bed is called the bed layer. In the bed layer, sediment
particles move as bed load by sliding, rolling or saltating. The law
of bed load motion is completely different from that of suspended
load. Since the bed load motion is dominant in the bed layer, i.e.,
the layer below the suspension region and above the bed, the
extension of the concentration distribution for the suspended load
to the near-bed region is not theoretically feasible.
If a in Equation 3.79 denotes the thickness of the bed
layer, then the suspended sediment discharge per unit width can be
expressed as:
gs = γs ∫ahSvyuydy
A=
z −1
(1 − A )
(3.81)
can be taken as the threshold value for sediment suspension.
However, various researchers have used other threshold values.
Bagnold (1966) used the value 3 and Engelund (1965) the value 2;
these values yield ratios of suspended load to bed load of 2:1 and
0.9:1, respectively.
Since Equation 3.79 was derived analytically in the
1930s, a number of studies have been conducted to test the diffusion theory against field observations and laboratory data. The
verification has two aspects: whether the formula structure is
correct, and whether the analytical expression for the exponent z is
valid. The conclusion is as follows. The formula structure is essentially correct; but there is a certain deviation between the
measured exponent z1 and the analytical expression z.
If
I1 = 0.216
A
I 2 = 0.216
A
 1 − y z
∫ A  y  dy
1
z
z −1
(1 − A )
 1 − y z
∫ A  y  ln ydy
z
Figure 3.21 — Relationship of I1 and A for suspended sediment
discharge with z as a parameter (after Einstein).
(3.82)
(3.83)
is used, then after substituting the logarithm velocity distribution
formula and Equation 3.79 into Equation 3.82 and simplifying,
one obtains:


 30.2 h 
 ⋅ I1 + I 2 

 ∆ 
gs = 11.6 γ sU* Sva a  2.303 log

A
(3.84)
(3.86)
Clearly, I1 and I2 are functions of A and z, and their values can be
obtained by numerical integration with the results shown in
Figures 3.21 and 3.22.
Einstein’s equations can be applied to compute the
suspended load discharge for given flow and sediment with the
following procedure.
a
h
(3.85)
1
I1
ω
where
I2
42
Figure 3.22 — Relationship of I2 and A for suspended sediment
discharge with z as a parameter (after Einstein).
CHAPTER 3 — SEDIMENT TRANSPORT IN RIVERS
Step 1: Compute a = 2D, U* = (ghJ)1/2
Step 2: Compute ∆ = Ks/χ, where Ks= D and χ can be obtained
from Figure 3.7
Step 3: Compute a = 2D, A = a/h
Step 4: Compute z = ω/(κU*)
Step 5: Get I1 from Figure 3.22 and I2 from Figure 3.23
Step 6: Compute Sva = ibgb/(11.6 × 2DU*) (details of determining Sva are discussed in section 3.4)
Step 7: Compute gs from Equation 3.84
43
Hou, et al. studied conditions when the velocity profile
at inflow was uniform, and they defined the boundary conditions
as follows:
1.
Free surface condition. At the free surface, y = h, the
upward transport by turbulent diffusion is the same as that due to
sediment settling, so that no sediment crosses the free surface.
εy
∂Sv
∂y
+ ωSv = 0
(3.92)
2.
Channel bed condition. The sediment concentration at
the bed approaches the saturation value Sv0 within a relative short
distance. Thus, at y = 0,
Sv = Sv0
3.
S/Sv0
(a)
(3.93)
At the entrance to the section, x = 0,
Sv = Sv0 f (y)
S/Sv0
(b)
Figure 3.23 — Variation of sediment concentration in a channel with a
movable bed starting with clear water at the point of inflow.
3.3.4
Non-equilibrium transport of suspended sediment
The vertical concentration distribution of suspended load for
steady uniform flow is treated in the preceding section. This
section treats the special case of non-equilibrium sediment transport in which the distribution of concentration varies in the
streamwise direction even though the flow of water is steady and
uniform. Typical examples of such a transport are the degradation
process induced by clear water erosion downstream of a newly
built dam and the aggradation process in a settling basin.
For simplicity, the following approximations are introduced.
1.
Sediment motion is steady:
(3.94)
If the inflow water is clear, then f (y) = 0.
The boundary conditions and the process of recovery of
sediment concentration in the direction of flow are shown in
Figure 3.23. The objective is to determine the sediment concentration distribution Sv (x,y) throughout the flow field.
For these conditions, the solution to the differential
equation has the form:
 ω
Sv ( x, y) = Sv0 exp −
 2ε y

y 



ω2x  ∞
∑ A exp
exp −
 4ε yU  n =1 n



 xp




− ωy  −
 2ε 
y

 ε β2 x 
− y n  sin β y
n

U 







(3.95)
where
∂Sv / ∂t = 0
(3.87)
An =
2.
The streamwise variation of the sediment exchange coefficient is negligible:
∂εx / ∂x = 0
(3.88)
For these conditions, the diffusion equation of sediment
transport becomes:
u
∂Sv
2
= εy
∂ Sv
2
+
∂ε y ∂Sv
+ω
∂Sv
(3.90)
∂y ∂y
∂y
∂y
For uniform sediment, the equation of non-equilibrium
sediment transport is the solution to this differential equation with
suitable boundary conditions.
The recovery of sediment concentration along the flow
direction by scouring is discussed first. If the variation of sediment
exchange coefficient with elevation can be neglected and its depthaveraged value is used, then Equation 3.90 can be further
simplified,
∂x
u
∂Sv
∂x
2
= εy
∂ Sv
∂y
2
+ω
∂Sv
∂y
(3.91)
(3.96)
n
y
−
(3.89)
 2 
 ω  + 4+ ω
β h
 ε β   ε β
2




ωy 
1
sin β ydy
f ( y) ⋅ exp 
∫
 2ε y 
ε ω


1
3.
The second derivative of sediment concentration with
respect to x is negligible compared to that in the y direction.
∂2Sv / ∂x2 << ∂2Sv / ∂y2
4
n
y
n
h
h
y
−
2 ω 2 + 4ε y2 β n2
n
0
and the coefficient βn can be calculated from:
tan β n h = −
2ε y β n
(3.97)
ω
The depth-averaged concentration can be obtained from
the integral of Equation 3.95 with respect to y over the depth h.
An example of the recovery of sediment concentration
resulting from clear water erosion as calculated from Equation
3.95 is shown for flow with a slope of J = 0.0001, depth h = 2.4 m,
mean velocity U = 1.9 m s–1, and particle sizes of 0.04 mm and
0.1 mm. The computed results, shown in Figure 3.24, indicate that
the distance required for the recovery of concentration from clear
water to the saturation state is generally not long if the sediment is
uniform and the streamwise variation of sediment size gradation
caused by clear water erosion is negligible. In the example shown
44
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
in Figure 3.24, the concentration recovers 89 per cent of the saturated value within a distance of 800 m.
In the foregoing discussion, the sediment was supposed
to be uniform. Hence, the bed material does not change during
degradation, i.e. the sediment-carrying capacity of the flow does
not vary along the river course unless the cross-section of the flow
changes. For this condition, the recovery distance is the distance
over which the streambed is scoured. The studies conducted by
various authors confirm that this distance is usually not long. In
nature, however, the bed material is composed of sediment with
mixed particle sizes. Because the flow can carry fine particles
more readily than coarse ones, most of the fine sediment is carried
away while the coarse sediment stays in place. The result is called
the armouring of the bed, and it causes a decrease in the sedimentcarrying capacity. This phenomenon starts upstream and
progresses downstream. For this reason, the distance for the sediment concentration to recover differs from that for erosion.
Although the former is rather short, the latter distance is quite long
(Chien, et al., 1986).
The analysis of deposition is quite similar to that of the
recovery of sediment concentration; only the boundary and initial
conditions are different. As an example, Zhang’s (1980) paper can
be referred to; details will not be discussed here.
3.4
TOTAL SEDIMENT LOAD
The total sediment load should include both bed load and
suspended load. In the previous paragraphs, relationships and
characteristics of bed load and suspended load are discussed. The
sum of the amount of bed load and suspended load is the total bed
material load that can be transported for a given flow and in given
boundary conditions. The characteristics of bed material load are
different from those of wash load. Consequently, formulae and
methods for calculating the bed material load and the wash load
are also different. Only sediment discharge in the form of bed
material load can be calculated on the basis of mechanics. This
will be discussed first.
3.4.1
Einstein’s bed load function
Einstein’s (1950) bed load function provides a method for computing the bed material load, and considers bed material, bed load
and suspended load in combination. For the sake of convenience,
one can assume that the transition from bed load to suspended
€
load occurs entirely at one elevation, i.e. below a given elevation
bed load movement prevails, and above this, suspension prevails.
The results of flume experiments reveal that unless the movement
of sediment is quite intense, this critical elevation is about two
grain diameters above the river bed. Einstein’s formulae for the
sediment carried as bed load and that carried as suspended load
are given as follows:
For bed load:
1−
where
1
π
B*ψ* −1 / η 0
∫
e
−t 2
dt =
− B*ψ* −1 / η 0
φ* =
ib
γ
(
i0 γ s − γ
)
2
1
(
3
)
1/ 2
γs − γ D
θ
(3.98)
1 + A*φ*
gD
( β / β)
ψ * = ξY
1/ 2
A*φ*
(3.99)
(3.100)
'
γ
Rb J
For suspended load:
isgs = 11.6U*Sva (PI1 + I2) γs
where
P=
1
h
log( 30.2
)
0.434
Ks / X
(3.101)
(3.102)
1
I1 = 0.216
A
z −1
(1 − A) z
∫ ( 1 −y y ) dy
z
(3.103)
ln ydy
(3.104)
A
1
I 2 = 0.216
A
z −1
(1 − A)
∫ ( 1 −y y )
z
z
A
A=
a
h
z=
(3.105)
ω
kU*
(3.106)
The quantities i0, ib and is are the portions of sediment
with median diameter D in the bed material, bed load, and
suspended load respectively; gb refers to the sediment discharge
of bed load and g s to the suspended load by weight per unit
width.
The next question is how to determine sediment concentration at the interface between the two (at the elevation a = 2D); it
is to be used as the specific reference concentration S va in
Equation 3.101. The mean sediment concentration (expressed in
percent by volume) in the bed surface layer is:
ib g b
(3.107)
2 Du b γ s
where u–b denotes the mean velocity of bed load movement.
If the sediment concentration at the top of the bed
surface layer is proportional to the mean value of sediment
concentration in the layer, and it is proportional to the friction
velocity, then:
Sua = ξ
ib g b
2 DU*γ s
(3.108)
The coefficient ζ has been shown in artificial flume
experiments to be the reciprocal of the well known constant 11.6.
Thus, the above expression can be rewritten as:
Figure 3.24 — Recovery of sediment concentration by clear water
erosion along a channel.
ibgb = 11.6SuaU*αγs
(3.109)
CHAPTER 3 — SEDIMENT TRANSPORT IN RIVERS
Clay concentration (ppm)
Substituting it into Equation 3.101, one obtains:
isgb = ibgb (PI1 + I2)
45
(3.110)
If gT denotes the total discharge of bed material expressed by
weight per unit width, including both bed load and suspended
load, and iT denotes the portion of sediment with diameter D in
bed material load, then:
(3.111)
This is Einstein’s formula for the sediment transport capacity as
bed material load, in which the term ibgb can be deduced from the
Einstein bed load function. If sand waves exist on the bed surface,
the term U* should be replaced by U*' = (Rb'gJ)0.5. Details of the
computation method and procedure can be found in the original
works of Einstein.
3.4.2
Colby’s method (1964)
Einstein’s procedure is complicated and laborious for practical
use. Guided by Einstein’s theory, a few methods for calculating
sediment transport were established using field data observed at
hydrometric stations. Among these formulae, the Colby method,
the modified Einstein procedure (Colby and Hembree, 1955) and
the Toffaletti formula (1969) have been widely used in western
countries.
The Colby method is suitable for rivers with beds of
medium to fine sand. The sediment transport capacity of a river
depends mainly on three factors: velocity, flow depth and sediment diameter (or fall velocity). Instead of using regression
analysis or an empirical curve fitting to express the effects of these
factors on sediment transport capacity, Colby developed a set of
graphs shown in Figure 3.25. Altogether 24 curves are included,
and they correspond to values of h varying by factors of 1 000 and
to various values of median diameter. The curves in Figure 3.25
are for a temperature of 60°F, D50 = 0.2 to 0.3 mm, and for flows
with a negligible amount of fine silt and clay. If the conditions are
not such, then the sediment transport found on the chart should be
multiplied by a correction factor:
'
eb
tan α
gs = 0.01τ 0U
He verified Equation 3.115 using various flume data for
which D was within the range of 0.11 to 5 mm, with satisfactory
results.
3.4.4
The Engelund-Hansen formula (1972)
The Engelund-Hansen formula is broadly recognized as one of the
most reliable formulae. They applied Bagnold’s stream power
concept and the similarity principle to obtain a sediment transport
formula:
fΦT = 0.4Θ5/2
(3.116)
f = 8ghJ / U2
where
(3.117)
Φ = gT [γs (γs – γ) gD3]–1/2
(3.118)
Θ = τ / (γ s – γ ) D
(3.119)
h = 0.03 m
h = 0.3 m
h=3m
h = 30 m
0.1 mm
0.2 mm
0.3 mm
mm
0.2 mm
3.4.3
Bagnold’s work (1966)
Bagnold’s formulae for sediment transport capacity for both bed
load and suspended load, which are in submerged weight, are as
follows:
'
(3.115)
(3.112)
where k1, k2 and k3 are correction coefficients for temperature,
content of fine silt and clay and median diameter of bed material,
respectively, as shown in Figure 3.26. Colby’s method is based on
measured data and consequently it cannot be used for a designed
purpose.
g b = τ 0U
eb
U
+ 0.001 )
tan α
ω
g 'T = τ 0U (
m
m m
m
1 + (k1k2 – 1) 0.01k3
Figure 3.25 —Work chart for the relationship for sediment transport
capacity (after Colby).
mm
iTgT = ibgb (1 + PI1 + I2)
mm
(3.113)
U
(3.114)
ω
where eb is the efficiency of bed load movement. Then, the transport rate of total bed material load by submerged weight is:
Mean velocity (m s–1)
Figure 3.26 — Correction factors (after Colby).
46
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
where g is the gravitational acceleration; h is the water depth, U*
is the average flow velocity; g is the total sediment discharge by
weight per unit width, γs and γ are the specific weights of sediment and water, respectively, D is the median particle diameter,
and τ is the shear stress along the bed.
Strictly speaking, Equation 3.116 should be applied to
flows with dune beds in accordance with the similarity principle.
Data from flume experiments show that the Engelund-Hansen
formula fits well not only for the dune-covered bed configuration,
but also for one with antidunes. If the mean velocity of sediment
movement is taken to be proportional not to the friction velocity
U*, but to the friction velocity U**′ relevant to grain resistance on
the bed surface, then the final expression for the Engelund-Hansen
sediment transport capacity formula takes the form:
fΦ T = 0.3Θ
2
2
Θ + 0.15
fφT ~ φ2
If Θ is large,
fφT ~ φ3
1
γ =
n
U*
gD
γs − γ
γ
[
SwT h U* n
( )
γs
U
D
γ
(3.123)
where SwT is the sediment concentration in percentage by weight
for a water column above a unit element of the bed surface. From
a large amount of flume data, they found this parameter of sediment transport to be a function of M and X. By analysing 1 000
sets of flume data, they obtained the final expression as:
3
γ =c(
(3.121)
D
If X > 60, the sediment is coarse, and the value corresponds to D > 2.5 mm for natural sediment; if X < 1, the sediment
is fine, and it corresponds to D < 0.04 mm for natural sediment; if
X is in between, 1 ≤ X ≤ 60, the sediment is in the transitional
region between the two for natural sediment.
They postulated that only part of the shear stress on the
channel beds is effective in causing the movement of coarse sediment, while in the case of fine sediment, suspended load
movement predominates, and the total shear stress is effective in
causing sediment movement. They suggested the mobility number
for sediment as follows:
M =
Figure 3.27 — Comparison of Engelund-Hansen formula against
flume data (after Engelund and Hansen).
intensity of sediment transport is related to the power provided by
flow. They assumed the efficiency of sediment transport to be
proportional to M. By combining the efficiency in M, they attained
the following parameter for sediment transport:
3.4.5
The Ackers-White formula (1973)
Based on Bagnold’s river power concept, Ackers and White
applied dimensional analysis to express the transport rate of sediment in terms of some dimensionless parameter. They used a
dimensionless parameter X to divide all sediment into three
groups: coarse, fine and medium,
 γs − γ 
g

γ 

X=
 v2 




Dune
(3.120)
and it is shown by the dotted line in Figure 3.27.
If Θ is small,
Ripple
U
32 log(
1− n
)
n
(3.124)
− 1)
A
The condition of incipient motion of sediment is M = A,
where for coarse sediment:
n=0
A = 0.17
c = 0.025
m = 1.5
For sediment in the transition region:
n = 1 – 0.56 log X
A=
0.23
+ 0.14
(3.125)
(3.126)
X
log c = 2.86 log X – (log X)2 – 353
]
10 h
M
(3.127)
(3.122)
D
where n = 0
for coarse sediment
n=1
for fine sediment
n = f(X)
for sediment in the transition region.
The parameter of sediment mobility M is no different
from the parameter of flow intensity Θ , which is referred to
frequently in the preceding sections and chapters.
In establishing the formula for sediment transport capacity, Ackers and White also adopted the Bagnold concept that the
m=
9.66
+ 1.34
(3.128)
X
If fine sediment exhibits any effect of cohesion among
the particles, the above-mentioned formulae are not applicable.
Figure 3.28 is a graphical representation of Equation 3.124.
3.4.6
Yang’s approach (1996)
Yang defined the unit power as the velocity-slope product. His
approach considers that the rate of work carried out by a unit of
CHAPTER 3 — SEDIMENT TRANSPORT IN RIVERS
47
suspended sediment transport capacity is approximately equal to
the total transport capacity. Among such formulae, the formula
most widely used in China is that developed at WUHEE:
Svm = k (
U
3
ghω
)
m
(3.134)
A similar formula is the Velikanov formula:
Svm = k (
U
3
(3.135)
)
ghω
The principle parameter in these formulae is the product
of U2/gh and U/ω. In Figure 3.29, a comparison of Equation 3.134
with field and laboratory data displays some scatter.
Figure 3.28 — Ackers-White formula for sediment transport capacity
(after Ackers and White).
water in transporting sediment must be directly related to the rate
of work available to a unit weight of water. Thus, the total sediment
concentration or total bed material load St must be directly related to
unit river power. Using dimensional analysis and considering that a
critical unit of river power UcrJ is required at incipient motion, he
found the best form of expressing the total bed material load:
UJ
ω
−
Ucr J
ω
(3.129)
)
where I1 and I2 are dimensionless parameters reflecting the flow
and sediment characteristics U*, v, ω and D. Running a multiple
regression analysis for 463 sets of laboratory data, he obtained the
final expression, as follows:
U
ωD
− 0.457 log *
ν
ω
U
UJ − Ucr J (3.130)
ωD
+(1.799 − 0.409 log
− 0.314 log * ) log (
)
ν
ω
ω
log St = 5.435 − 0.286 log
where St is the total sediment concentration in ppm by weight.
The critical dimensionless unit of river power Ucr J/ω is
the product of the dimensionless critical velocity Ucr /ω and the
energy slope J, where:
Ucr
=
ω
2.5
+ 0.066
U* D
log(
) − 0.06
ν
U*
ω
= 2.05
U D
for 1.2 ≤ * < 70
ν
(3.131)
for 70 ≤
U* D
ωD
+(1.780 − 0.360 log
ν
ωD
ν
− 0.297 log
U*
ω
– 0.480 log
U*
ω
(3.133)
) log (
3.4.8.1
ANNUAL SEDIMENT LOAD EVALUATED BY THE
RELATIONSHIP BETWEEN FLOW DISCHARGE AND
SEDIMENT TRANSPORT RATE
Regular measurements of flow discharge and sediment sampling
are the routine work of hydrometric stations. In most cases, data
are obtained for sediment transport rates related to various flow
rates. By means of the relationship between flow discharge and
sediment transport rate and the frequency curve for flow, one can
evaluate the total sediment load at a given hydrometric station.
However, this approach is affected by three procedural difficulties.
(3.132)
ν
As the rate of sediment transport increases, the need to
include incipient motion criteria in a sediment transport equation
decreases. For sediment concentrations higher than about 100 ppm
by weight, Yang introduced the following unit river power equation:
log St = 5.165 − 0.153 log
Estimation of total sediment load including
wash load
The sediment transport capacity formulae presented in the preceding sections that were established on the basis of mechanics
should be used to compute only the sediment discharge in the
form of bed channel-derived load. For the wash load, the relationship between sediment transport rate and flow rate is based on
factors related to the common background of the watershed. Such
a relationship can be established only from data observed in the
field, including: (i) data of sediment load measurement at hydrometric stations; (ii) information of sediment yield in drainage
basins; and (iii) measurements of sediment deposits in reservoirs.
The following section contains a discussion of the nature of such
data and methods for processing the three kinds of data.
UJ
ω
Sediment concentration, Sm (kg m–3)
log St = I1 + I 2 log (
3.4.8
Yangtze River
Yellow River
People’s Canal
Qingtong irrigation area
Sanmenxia Reservoir
)
Formula of the Wuhan University of Hydraulic and
Electric Engineering (WUHEE)
For rivers flowing over alluvial plains, suspended load predominates, and the bed load is generally negligible. In such cases, the
Guanting Reservoir
Flume data by WUHEE
3.4.7
U3/(ghω)
Figure 3.29 — Comparison of Equation 3.134 with observed data.
48
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
(1) The observed field data usually do not include the measurement of bed load, and the portion of suspended load near the bed
surface is not easy to measure, so the measured data do not fully
reflect the total sediment load carried by the flow. (2) In some
rivers, the measured data points are widely scattered; thus it is
difficult to establish a relationship of flow discharge versus sediment transport rate by conventional methods of curve fitting. (3)
Fewer data for sediment transport rate are available than for flow
discharge, and data series may be too short to be representative of
average conditions. Efforts made to find rational solutions to these
difficulties are discussed as follows.
A.
Evaluation of total sediment load based on measurement of
suspended load
Because of its size, a suspended load sampler is designed to
exclude the main zone of bed load transport close to the bed. Also,
the bed load is well outside the scope of suspended load sampling.
For the wash load, which is mainly composed of fine sediment
that is uniformly vertically distributed, the mean sediment concentration obtained by conventional sampling methods should
represent satisfactorily the true mean value. However, for the bed
material load, especially particles coarser than fine sand, a considerable part is concentrated near the streambed, and it is not
included in the results of suspended sediment sampling. How to
estimate the unmeasured sediment load is a major concern.
Samples are taken at points and by depth-integration devices, and
the necessary corrections for these two methods are different.
Chien and Wan (1956) proposed a correction method for point
samples, and Chien (1953) proposed a correction method for
depth-integrating samples. Details are not discussed here, but one
can refer to Chien and Wan (1983).
Method of establishing a relationship for discharge-sediment
transport rate from scattered data points
If the wash load in the drainage area is large, and both regional
factors (such as vegetation cover, topography and soils, etc.) and
the rainfall distribution are strongly non-uniform, the data points
for the measured sediment transport rate plotted against measured
discharge usually display a wide band of scatter. If the relationship
is established by following the trend of the data, considerable error
will result in the computation of the annual sediment load using
that relationship and the corresponding frequency curve for river
discharge.
The wide scatter of the data points can result from two
circumstances. First, owing to the large spatial differences, runoff
formed in different areas may lead to quite different sediment
concentrations, sometimes high, and other times low. Second, the
scatter may be due to temporal differences in runoff. For example,
in early spring there is a high volume of runoff because of melting
ice and snow; in summer and autumn, heavy rainstorms cause
floods. The sediment concentrations for these two cases differ
greatly. In some drainage areas, both conditions occur and the
situation is then even more complicated. In addition, heavily sediment-laden rivers, because of the self-regulation of the channel,
are characterized by the ‘more sediment may be released if more
sediment is supplied’ phenomenon. Such a situation can enhance
the extreme scatter of data points in a plot of sediment transport
rate against water discharge.
In analyses of hydrological data, one can sometimes
determine the concrete causes of the scatter of data points. One
can then calculate a set of relationships for discharge vs. sediment
transport rate and the corresponding discharge frequency curve for
the specific events of runoff originating from different source
areas or occurring in different seasons. The total sediment transport rates for given time periods are then evaluated separately. An
example is illustrated in Chien and Wan (1983).
3.4.8.2
ESTIMATION OF SEDIMENT LOAD BASED ON FACTORS IN
RIVER BASINS
If soil erosion is the source of sediment, the amount of sediment
conveyed in the river system is naturally related to the various
factors that affect soil erosion in the watershed. If such relationships can be shown graphically or expressed by mathematical
equations, the amount of sediment originating from the watershed
and conveyed into the river can be deduced from the characteristics of river basin factors; such a process can be useful if there is a
lack of field data.
In practical applications, two approaches are possible.
The first is to establish a direct relationship for the sediment load
conveyed into the river expressed in terms of the characteristics of
the given watershed, and based on measured data from the hydrometric network. The second is to estimate the amount of soil
eroded from the ground surface, and then to estimate how much of
that material can be carried into the river (see Chapter 1).
Anderson (1951) analysed measured data for 29 watersheds in Oregon, United States (watershed areas ranging from 145
to 18 850 km2), to establish a relationship between suspended load
and various regional characteristics of watersheds that had hydrometric stations (he assumed that the bed load was negligible). The
included factors comprise a set that appears to be quite complete.
The standards relating to the measurement, units and physical
interpretation of these factors are given in Table 3.3.
B.
3.4.8.3
ESTIMATION OF SEDIMENT YIELD OF A WATERSHED FROM
RESERVOIR DEPOSITION
If a large reservoir is constructed in a river, all of the sediment
load from the upstream areas will be intercepted by the reservoir.
Thus, measuring the amount of deposition in the reservoir is a reliable way to assess the sediment yield of the drainage area.
If, on the contrary, the storage capacity of the reservoir
is not large relative to the volume of runoff, then part of the sediment load may be carried downstream. Thus, the sediment yield
based on deposition in the reservoir must include the efficiency
of the reservoir in trapping sediment. Figure 3.30 shows the relationship between the sediment outflow to inflow ratio during
flood events and the characteristics of the reservoir, with sediment size and concentration as additional parameters (Xia, Han
and Jiao, 1980), where V is the storage volume of the reservoir,
Q i is the inflow and Q 0 is the outflow. The abscissa of the
diagram, VQi /Q02, has the dimension of time; it reflects the time
of flood detention in the reservoir. In addition, the efficiency of
sediment release is also related to sediment size and sediment
concentration. Fine sediment can be released much more easily
than coarse sediment. If the fine sediment concentration exceeds
50 kg m–3, the fall velocity of the sediment is less, and more
sediment is released.
In addition to the above approaches, physical and
mathematical models have been used recently to study the
formulation and confluence of runoff, including the concept of
sediment yield.
CHAPTER 3 — SEDIMENT TRANSPORT IN RIVERS
49
distinct features: at a given gauging station, the higher the
concentration, the coarser the suspended sediment (Chien and
Wan, 1986).
When the total concentration exceeds a certain value,
clay content no longer increases with concentration, but rather
maintains a certain value. The persistence of fine-material
Release efficiency
3.5
HYPERCONCENTRATED FLOW
In an ordinary sediment-laden flow, sediment is carried by the
flow and it has little effect on flow behaviour. Therefore such an
effect can be ignored. In hyperconcentrated flow, however, the
existence of large amounts of solid particles remarkably influences or changes fluid properties and flow behaviour. In such
cases, the above-mentioned influence or change must be considered. In many cases of hyperconcentrated flow, sediment together
with water, forming a pseudo-one-phase fluid, moves as its own
entity, and the sediment can no longer be considered as material
carried by water.
The existence of hyperconcentrated flow cannot be
judged simply by concentration alone. Grain size composition
and mineral content of sediment play a very important role. As
regards the Yellow River where the incoming sediment has
similar mineral content and grain size composition, flow with a
concentration higher than 200 kg m –3 can be considered as
hyperconcentrated.
In a natural environment, debris flows, turbidity flows
along the sea bottom and hyperconcentrated density flows can be
considered as hyperconcentrated flow. Hyperconcentrated
hydrotransport is a kind of hyperconcentrated flow in industry.
In hyperconcentrated flow in the Yellow River basin,
the size composition of the suspended sediment exhibits some
Si > 50 kg m–3
Si > 50 kg m–3
(s)
Figure 3.30 — Relationship between efficiency of sediment release and
characteristics of reservoir and sediment load.
Table 3.3
Principal factors affecting sediment yield in western Oregon, United States (after Anderson)
Factors in watershed
Flow
Soil
Geography
Unit
Average value
Range
MAq
FOp
m3 km–2
–
0.325
3.56
0.0114–0.0817
1.98–.30
Magnitude of runoff
Intensity of runoff
SC
per cent
23.0
19.1–22.0
Representing the source of
suspended load easily suspended
and carried away by the flow
Aggregate ratio*
B
per cent/
(cm2 g–1)
1.37
0.56-3.84
It reflects the permeability and
ability of the soil to withstand
erosion
Area of watershed
River gradient
A
J
km2
m km–1
2.00
172
145–18 850
40–286
—
Average gradient of surface soil in
watershed
Road
R
per cent
0.3
0.05–0.6
Road construction includes water
soil erosion
Forest cut within
last ten years
RC
per cent in
ten years
6.0
0–30.4
Cutting down forests destroys
protection provided by forest
BC
per cent
4.0
0–22
Some erosion will be greatly
reduced if land surface is covered
by plants
OC
per cent
12
0–48
As above
Cultivated land =
(BC + OC) × A
C
km2
20.7
0–173.5
—
Eroded bank
EB
m
5 180
62 500
Soil resulting from bank erosion
directly enters the river
Average annual
runoff steepness
of flood discharge
The content of
silt and clay in
the surface soil
layer (15 cm)
Cultivated land
Land utilization with thin cover
Cultivated land
different from BC
Symbol
* Definition of soil aggregate ratio B and the technique for measuring the term B are given in reference (Anderson, 1951).
Physical meaning
50
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
concentration is a general characteristic not only of the hyperconcentrated flow in a river system, but also of the hyperconcentrated
lahar-runoff flow.
The features noted above are important in the transport
of hyperconcentrated flow. A certain amount of fine particles form
an intricate network of the floc structure which effectively reduces
the fall velocity of coarse particles, thereby ensuring a high sediment transport capacity. When the concentration rises beyond a
certain limit, further increases in concentration will only make the
sediment composition coarser. The clay content does not increase
with the high concentration, thereby ensuring the flow will not
transform into a laminar one which requires a much larger slope to
be kept in motion.
It is well known that clear water is a Newtonian fluid
with a viscosity of m. Water with a low concentration of sediment
remains Newtonian fluid, but the viscosity increases with increasing concentration. As sediment concentration exceeds a certain
value, particularly for sediment containing clay particles, the
water-sediment mixture no longer behaves as a Newtonian fluid.
The critical concentration Sv0 varies according to the size composition and mineral composition of sediment as well as the water
quality.
Data from rheological measurements indicate that most
hyperconcentrated flows can be described as Bingham fluid. The
mixture of water and finer granular particles carried by debris
flows can also be described as Bingham fluid. The rheological
equation of Bingham fluid is:
τ = τB + η
du
(3.137)
In many cases m = 3 is adopted, but in some cases m is
smaller in a low concentration region but larger in a high concentration region.
Fall velocity of sediment particles is an important parameter in sediment transport. The fall velocity of sediment particles
in hyperconcentrated flow may be reduced many times due to the
increase in viscosity, the backflow caused by other settling particles and a reduction in the effective weight, etc.
The most widely adopted formula of the gross fall velocity for uniform discrete sediment particles is suggested by
Richardson and Zaki (1954):
ω0
ν
y
2η
( 2 γ m HJ − γ m yJ − 2 τ B ), 0 ≤ y ≤ H −
m
(3.138)
Chien (1980) suggested that the exponent m is a function
of the grain Reynolds number.
τB
γmJ
(3.140)
Equation 3.140 can be rewritten as:
up − u
= (1 −
up
γ m yJ
γ m HJ − τ B
)
2
0≤y≤H−
,
τB
γmJ
(3.141)
where γm is the specific weight of the mixture, H is the depth, and
y is the distance from the bed; up is the maximum velocity in the
profile and equal to the velocity of the plug zone. In the plug zone,
y > H – τB/γmJ, where the shear stress is smaller than the Bingham
yield stress, there is no relative motion between layers, and the
whole fluid moves as an entity with velocity up, as shown in
Figure 3.31.
u = up =
γmJ
2η
(H −
τB
γmJ
2
H−
) ,
τB
γmJ
<y≤H
(3.142)
A hyperconcentrated flow transforms into turbulent flow
if the Reynolds number is large. The flow begins to develop into
turbulence at Rem = 2 000 and develops fully into turbulence if
Rem >10 000. The flow is in a transitional region if Rem = 2 000 to
10 000.
Re =
4 ρ m HU
m
= (1 − Sv )
(3.139)
)
There are several patterns of hyperconcentrated flow.
Neutrally buoyant load motion. If a flow carries enough clay
material, the mixture may exhibit strong yield strength, and
most sediment in the flow will belong to neutrally buoyant load.
Mud flow in the Loess Plateau is an example of such flow.
B. Neutrally buoyant load motion + suspended load + bed load.
A part of fine sediment moves as neutrally buoyant load,
while coarse sediment is transported as suspended load and
bed load.
C. Suspended load + bed load motion. If a hyperconcentrated
flow carries very little clay material, sediment moves mainly
as suspended load along with a small part of bed load.
D. Laminated load motion and neutrally buoyant load + laminated load. If the energy slope of a flow is sufficiently high
and there is only cohesive material available, laminated load
motion may develop. Water debris flow is essentially a laminated load motion. On the other hand, in viscous debris flow
gravel, cobbles and big stones may move as laminated load
and sand and silt may be neutrally buoyant load.
For hyperconcentrated flow, laminar flow or turbulent
flow may occur, depending on relevant conditions.
If the concentration is high enough, laminar flow might
appear in small rivers or canals. Considering Bingham fluid
flowing in an open channel with slope J, a theoretical velocity
profile is obtained as follows:
u=
τB = KSvm
ω0 D
A.
(3.136)
dy
where τB and η are called the Bingham yield stress and the coefficient of rigidity, respectively.
Bingham yield stress and rigidity vary with the size
composition and mineral composition of sediment and the sediment concentration. The higher the content of fine particles, the
larger the Bingham yield stress and the rigidity. They increase
rapidly when sediment concentration increases. Researchers have
drawn up various empirical formulae for describing the relationship between rheological parameters and sediment concentrations.
Exponential formulae are widely used, such as:
ω
m = f(
η(1 +
τ BH
2 ηU
(3.143)
)
In a fully developed turbulent flow, the velocity distribution still follows the logarithmic formula, but the velocity gradient
CHAPTER 3 — SEDIMENT TRANSPORT IN RIVERS
Figure 3.31 — Velocity distribution of a laminar flow.
or the Karman constant κ is different from that of clear water a
flow. The κ constant varies with concentration, as shown in
Figure 3.32.
In the case of a pipe flow, the Reynolds number should
be modified as:
clay slurry than in clear water. Wan and Sheng (1978) found that
in a region of high concentration, the relationship S – U3/gHω0,
which is used to describe sediment carrying capacity has a reverse
tendency and has a hook-like outline, as shown in Figure 3.33.
Here, S is the average concentration of a flow under equilibrium
conditions, and ω0 is the fall velocity of a single particle in still,
clear water.
In the region of high concentration (about S >
200 kg m–3) more sediment can be carried by flow, with even
weaker intensity. In other words, high concentration does not
require high flow intensity to be carried. This is a very useful
concept.
The reason for the reverse tendency of the S – U3/gHω0
relationship is the obvious reduction in the fall velocity at high
concentrations, particularly at high concentrations of fine particles. If the reduction in fall velocity due to concentration has been
taken into consideration, the hyperconcentrated flow follows the
same law as that followed by an ordinary sediment-laden flow.
4 ρ m UR
2
η(1 +
2 τ BR
3 ηU
(3.144)
)
and the f vs. Re1 relationship shifts a little from the f vs. Rem curve
for open channel flow.
As mentioned above, in hyperconcentrated flow, the fall
velocity of sediment particles is reduced quite substantially.
Consequently, in such a flow, sediment is easier to transport. If all
the sediment particles belong to a neutrally buoyant load, the flow
can be maintained, provided the potential energy of the flow is
sufficient for overcoming the resistance. If not all the sediment
particles belong to a neutrally buoyant load, due to the reduction
of their fall velocities, coarse particles are easier to transport in
S (kg m–3)
Re =
51
(
U 3 0.92
)
Hw 0
S (kg m–3)
Figure 3.33 — Sediment carrying capacity S – U3/gHw0.
€
x
Figure 3.32 — Variation of constant κ with sediment concentration.
REFERENCES
Ackers, P. and W.R. White, 1973: Sediment transport: new
approach and analysis. Journal of the Hydraulics Division,
Proceedings. ASCE, 99 (Hy11), pp. 2014-2060.
Agricultural Research Service, USDA 1975: Present and prospective technology for predicting sediment yields and sources.
Proceedings of the Sediment Yield Workshop, USDA
Sedimentation Lab., p. 285.
Anderson, H.W., 1951: Physical characteristics of soils related to
erosion. Journal of Soil and Water Conservation,
pp. 129-133.
Bagnold, R.A., 1966: An approach to the sediment problem from
general physics. USGS, Professional Paper 422-I, p. 37.
Boyce, R.C., 1975: Sediment routing with sediment-delivery
ratios. Present and Prospective Technology for Predicting
Sediment Yields and Sources, Agricultural Research Service,
USDA, pp. 61–73.
Cao, Ruxuan, 1975: Discussion on sediment-carrying capacity of
hyperconcentrated flow. Selected Papers of the Symposium
on Sediment Problems on the Yellow River, Volume 2,
pp. 249-258 (in Chinese).
52
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
Chien, Ning, 1980: Comparison of bed load formulae. Journal of
Hydraulic Engineering, Number 4, pp. 1-11 (in Chinese).
Chien, Ning, 1980: Preliminary study on the mechanism of
hyperconcentrated flow in the north-west region of
China. Selected Papers of the Symposium on Sediment
Problems on the Yellow River, Volume 4, pp. 244-267 (in
Chinese).
Chien, Ning and Zhaohui Wan, 1983: Mechanics of Sediment
Motion. Science Press.
Chien, Ning, Ren Zhang and Zhide Zhou, 1986: Fluvial
Processes. Science Press (in Chinese).
Chien, Ning and Zhaohui Wan, 1986: A Critical Review of the
Research on Hyperconcentrated Flow in China. International
Research and Training Center on Erosion and Sedimentation,
p. 42.
Colby, B.R., 1964: Discharge of sands and mean-velocity relationships in sand bed streams. Professional Paper 462-A, USGS,
p. 47.
Colby, B.R. and C.H. Hembree, 1955: Computation of total sediment discharge, Niobrara River near Cody, Nebraska. Water
Supply Paper 1357, USGS.
Dou, Guoren, 1960: On threshold velocity of sediment particles.
Journal of Hydraulic Engineering, April, pp. 22-31 (in
Chinese).
Dou, Guoren, 1962: Theory on threshold motion of sediment
particles. Scientia Sinica, Volume 11, Number 7, pp. 9991032 (in Russian).
Einstein, H.A., 1950: The bed load function for sediment transport
in open channel flows, Technical Bulletin 1026, p. 71,
USDA.
Einstein, H.A. and N.L. Barbarossa, 1952: River channel roughness. Transactions, ASCE, Volume 117, pp. 1121-1146.
Einstein, H.A. and Ning Chien, 1953: Transport of sediment
mixtures with large ranges of grain sizes. Sediment Series
Number 2, Missouri River Division, U.S. Army Corps of
Engineers, p. 49.
Engelund, F., 1965: A criterion for the occurrence of suspended
load. La Houille Blanche, Number 6, p. 607.
Engelund, F. and E. Hansen, 1972: A Monograph on Sediment
Transport in Alluvial Streams. Teknisk Forlag, Copenhagen,
p. 62.
Englund, F. and J. Fredsφe, 1976: A sediment transport model for
straight alluvial channels. Nordic Hydrology, Volume 7,
pp. 293-306.
Fei, X., 1982: Viscosity of the fluid with hyperconcentration-coefficient of rigidity. Journal of Hydraulic Engineering,
Number 3, p. 57-63 (in Chinese).
Fei, X., 1983: Grain composition and flow properties of heavilyconcentrated suspension. Proceedings of the International
Symposium on River Sedimentation, pp. 307-308 (in
Chinese).
Foster, G.R. (ed.): Soil Erosion: Prediction and Control. Special
Publication Number 21, Soil Conservation Society of
America, 1977, p. 393.
Gessler, J., 1970: Self-stabilizing tendencies of sediment mixtures
with large range of grain sizes. Journal of the Waterways and
Harbor Division, ASCE, 96 (WW2).
Gessler, J., 1971: Critical shear stress for sediment mixtures.
Proceedings of the Fourteenth Congress. International
Association for Hydraulic Research, Volume 3, pp. 1-8.
Gessler, J., 1976: Stochastic aspects of incipient motion on
riverbeds. Stochastic Approach to Water Resources,
Volume 2, (ed.) H.W. Shen, p. 26.
Goncharov, B.N., 1962: River Dynamics. Hydrology and
Metrology Press, Leningrad, pp. 226-236 and p. 252 (in
Russian).
Goncharov, V.N., 1962: Basic River Dynamics. HydroMeteorological Press, Leningrad (in Russian).
Gong, Shiyang and Guishu Xiong, 1980: The origin and transport
of sediment of the Yellow River. Proceedings of the
International Symposium on River Sedimentation, Volume 1,
pp. 43-52 (in Chinese).
Jiang, Zhongshan and Wenjin Song: Sediment yield in small
watersheds in the Gullied-Hilly Loess areas along the middle
reaches of the Yellow River. Proceedings of the International
Symposium on River Sedimentation, 1980, Volume 1, pp. 6372 (in Chinese).
Kramer, H., 1935: Sand mixtures and sand movement in fluvial
models. Transactions, ASCE, Volume 100, pp.798-838.
Levy, E.E., 1956: River Mechanics. National Energy Press,
Moscow (in Russian).
Levy, I.I., 1957: River Dynamics. National Energy Resources
Press, Moscow, pp.127-140.
Li, Changhua and Meixiu Sun, 1964: Criteria for threshold shear
stress and ripple formation. Proceedings of the Nanjing
Hydraulic Research Institute (River and Habour Division)
(in Chinese).
Meyer-Peter, E., H. Favre and H.A. Einstein, 1934: Neuere
Versuchsresultate uber den Geschiebetrieb. Schweiz
Bauzeitung, Volume 103, Number 12, pp. 147-150.
Meyer-Peter, E. and R. Muller, 1948: Formula for bed load transport. Proceedings of the Second Meeting. International
Association for Hydraulic Research, Volume 6.
Migniot, C., 1968: Study on physical properties of fine sediment
(silt) and their properties under flow dynamics. La Houille
Blanche, Volume 7 (in French).
Migniot, C., 1977: Effects of flow, wave and wind on sediment. La
Houille Blanche, Number 1 (in French).
Renfro, G. W., 1975: Use of erosion equations and sedimentdelivery ratios for predicting sediment yield. Present
and Prospective Technology for Predicting Sediment
Yields and Sources, Agricultural Research Service, USDA,
pp. 33-45.
Richardson, J.F. and W.N. Zaki, 1954: Sedimentation and fluidization, Pt. 1. Transactions, Institute of Chemical Engineers,
Volume 32, pp. 35-53.
Rouse, H., 1938: Experiments on the mechanics of sediment
suspension. Proceedings of the Fifth International Congress
of Applied Mechanics, pp. 550-554.
Scott, K.M. and R.L. Dinehart, 1985: Sediment transport and
deposit characteristics of hyperconcentrated streamflow
evolved from lahar at Mount St. Helens. Proceedings of the
International Workshop on Flow at Hyperconcentrations of
Sediment, International Research and Training Center on
Erosion and Sedimentation, pp. 3-2-33.
Shamov, G.E., 1952: Formulae for determining near-bed velocity
and bed load discharge. Proceedings of the Soviet National
Hydrology Institute, Volume 36 (in Russian).
Sharmov, G.I., 1959: River Sedimentation. Hydrology and
Metrology Press, Leningrad, pp. 84-93.
CHAPTER 3 — SEDIMENT TRANSPORT IN RIVERS
Shields, A., 1936: Anwendung der Aechlichkeitsmechanik und der
Turbulenzforschung auf die Geschiebewegung, Preussische
Versuchsanstalt fur Wasserbau und Schiffbau, Berlin.
Simons, D.B. and E.V. Richardson, 1960: Resistance of flow in
alluvial channels. Journal of the Hydraulics Division, ASCE,
86 (Hy5), pp. 73-99.
Tang, Cunben, 1964: Law of sediment threshold. Journal of
Hydraulic Engineering (in Chinese).
Tan, Ying, 1983: Inttermittent surges of gravel transport in rivers.
Proceedings of the Second International Symposium on
River Sedimentation, Water Resources and Electric Power
Press.
Tixon, L.J., 1948: Etude des conditions dans lesquelles les particules solides sont transportées dans les courants à lit mobiles.
Proceedings, Association internationale des sciences
hydrologiques, Volume 1, Oslo, pp. 293-310.
Toffaletti, F.B., 1969: Definitive computations of sand discharge in
rivers. Journal of the Hydraulics Division. ASCE, 95 (Hy1),
pp. 225-246.
Yang, Chih-Ted, 1973: Incipient motion and sediment transport.
Journal of the Hydraulics Division, ASCE, 99 (Hy10),
pp. 1679-1704.
Yang, C.T., 1996: Sediment Transport — Theory and Practice.
McGraw-Hill.
Yang, Wenhai and Wenlin Zhao, 1983: An experimental study of
the resistance to flow with hyperconcentration in rough
53
flumes. Proceedings of the Second International Symposium
on River Sedimentation, pp. 45-55.
Wan, Zhaohui and Shoubai Sheng, 1978: Phenomena of hyperconcentrated flow on the stem and tributaries of the Yellow
River. Selected papers from the Symposium on Sediment
Problems on the Yellow River, Volume 1, pp.141-158 (in
Chinese).
Wan, Zhaohui, Qingsong Li, Tison Yang and Jungfeng Tian, 1998:
Incipient motion of fine particles under large water depth.
Proceedings, InterCoh 98.
White, C.M., 1940: Equilibrium of grains on bed of stream.
Proceedings, Royal Society of London, Series A,
Volume 174, pp. 322-334.
White, S.J., 1970: Plain bed thresholds for fine grained sediments.
Nature, Volume 228, Number 5267, pp. 152-153.
Wuhan University of Hydraulic and Electric Engineering
(WUHEE), 1961: River Dynamics. China Industrial Press (in
Chinese).
Xia, Zhenhuan, Qiwei Han and Enze Jiao, 1980: The long-term
capacity of a reservoir. Proceedings of the International
Symposium on River Sedimentation. Volume 2, pp. 753-762
(in Chinese).
Zhang, Qishun, 1981: Study of sediment-diffusion mechanism for
open-channel flow and its application. Chinese Journal of
Sediment Research, Volume 1, pp. 37-52 (in Chinese).
CHAPTER 4
FLUVIAL PROCESSES
4.1
INTRODUCTION
Fluvial processes, broadly speaking, involve the study of entire
historical processes of formation and evolution of various parts of
a river valley from its origin to estuary and belong to the
geomorphologic category. But in a narrower sense, fluvial
processes relate to river changes that occur owing to natural
conditions or human activities and belong to the category of river
dynamics. The latter is more spectacular from an engineering point
of view.
The fluvial processes of alluvial rivers are the result of
the interaction of flow, sediment and channel bed. The channel
bed influences the current structure and sediment movement, and
the flow and sediment transport enhance changes in the channel
bed. They are interdependent and condition each other. Because
the flow and sediment transport are ever changing, the fluvial
processes are quite complicated, which can benefit humans or lead
to disasters, so rivers should be monitored. River regulation and
training works must take into account the characteristics of fluvial
processes of rivers so that the river training works can help rivers
to do what they would do naturally rather than force them into an
unnatural situation, which would ultimately lead to failure. Fluvial
processes are quite different from one river to another and river
training measures are also multifarious. The purpose of this
chapter is to introduce the main aspects of fluvial processes, river
training and river sediment management, including categories of
rivers, classification of river patterns, river morphology, fluvial
processes for rivers with different patterns, and the operational
measures of channel stabilization and rectification, so as to meet
planning and design requirements for river regulation and river
training works.
4.2
CATEGORIES OF RIVERS
According to their geometrical position, rivers can be divided into
two major types: mountainous rivers and plain rivers. The upper
reaches of large rivers are always the mountainous or upland
rivers, while the lower reaches are plain rivers.
(a)
4.2.1
Mountainous and upland rivers
Mountainous and upland rivers have the following features:
(1) The flood peak rises rapidly and falls sharply, and the
maximum discharge might be hundreds or thousands of
times higher than the minimum. In South China, most mountainous rivers have a sediment concentration of less than
1 kg m–3 in flood seasons. However, for the rivers in the
Loess Plateau in North China, the maximum sediment
concentration might be over 1 000 kg m–3, and debris floods
often occur in the mountainous rivers in southwest China.
(2) Under the effects of the geological structure and flow
actions, well-developed terraces exist along both sides of
such rivers, but there is no wide flood plain. Diluvial fans
and mouth bars often occur at the outfalls of their tributaries.
(3) The longitudinal profile is steep, the torrents wind through
shallow shoals, and the channel bed manifests itself rising
and falling along the river.
(4) The valley cross-section is V- or U-shaped (Figure 4.1).
(5) The river bed is composed of base rock and gravel. When
earthquakes occur, landslides, mountain slides and rapid bed
deformation take place, and the channel may often be
blocked. Dammed and falling water is formed upstream and
downstream of the block.
4.2.2
Plain and piedmont rivers
The features of plain rivers can be described as follows.
(1) These rivers have large catchment areas and smooth flood
hydrographs. The ratio of the maximum to minimum
discharge at the Yichang Station on the Middle Yangtze
River in China is only 26, and 10 at Bahadurabad Station on
the Brahmaputra River in Bangladesh. However, for rivers
with less runoff and concentrated rainstorms, such as the
Lower Yellow River, the flood peak still rises and falls
rapidly. The average ratio of maximum to minimum
discharge at Huayuankou Station on the Lower Yellow River
is as high as 446.
(b)
Figure 4.1 — Morphology of mountainous river valleys on Maohu Reach of the Beipanjiang River (China) (a) V-shaped valley,
(b) U-shaped valley, ∇1 high water level.
CHAPTER 4 — FLUVIAL PROCESSES
55
1, 2, 3 — Flood, middle and low flow; 4 — Valley slopes; 5 — Flood plains; 6 —Lips of flood plains; 7 — Side bar;
8 — Levees; 9 — Sediment deposit; 10 — Original rock bed.
Figure 4.2 — Morphology of plain river valleys.
(2)
(3)
(4)
(5)
The incoming sediment load is determined by the
characteristics of the river basin. For example, the longterm annual incoming sediment load at Yichang Station
amounts to 0.521 × 10 9 t, with an average sediment
concentration of 1.18 kg m –3, while for the Yellow River
flowing through the Loess Plateau, a seriously eroded
region, the long-term annual incoming sediment load is
1.62 × 109t, with an average sediment concentration of 37.6
kg m–3 at Shanxian Station.
The river valleys have deep alluvial layers with thicknesses
of tens or hundreds of metres. The channel beds are
composed of loose sediment deposits which can be easily
eroded.
The fully developed valleys have a main channel and wide
flood plains (Figure 4.2).
The longitudinal profile is even and smooth. The channel
slope of the Middle and Lower Yangtze River is
0.1–0.027‰, the Lower Yellow River is 0.1–0.2‰, and
channel slopes of the Upper, Middle, and Lower
Brahmaputra River are 0.086–0.071‰, 0.072–0.047‰, and
0.038–0.034‰, respectively. The annual runoff and sediment
load for the major rivers in the world are listed in Table 4.1
(Chien and Dai, 1980; Sedimentation Committee, 1992;
China–Bangladesh Joint Expert Team, 1991).
4.3
CLASSIFICATION OF RIVER PATTERNS
4.3.1
River patterns
In China, rivers are often categorized in four basic patterns
according to their static and dynamic characteristics.
(1) Straight: Straight rivers are usually relatively short reaches
having negligible sinuosity at the bankfull stage. At low
stages, there are sand bars on both sides of the stream, and
the thalweg meanders in a sinuous path along the bars
(Figure 4.3 (a)). The alternate sand bars move downstream
and the thalweg also shifts simultaneously. Long, straight
rivers rarely occur naturally, and are often engineered.
Table 4.1
Annual runoff and sediment load of some rivers in the world
State
River
Station
Drainage
area (km2)
Annual runoff
(109 m3)
Annual
sediment load
(109t)
Average sediment
concentration
(kg/m3)
Bangladesh
Brahmaputra River
Bahadurabad
535 000
618
0.499
0.81
Bangladesh
Ganges
Harding Bridge
963 000
344
0.196
0.57
Pakistan
India River
Kodli
969 000
175
0.435
2.49
Burma
Irrawaddy River
Polom
430 000
427
0.299
0.70
Viet Nam
Red River
Hanoi
119 000
123
0.130
1.06
United States
Mississippi Rver
Estuary
322 000
561
0.312
0.56
United States
Missouri River
Herman
1 370 000
61.6
0.218
3.54
United States
Colorado River
Grand Canyon
356 000
5.6
0.182
11.67
Brazil
Amazon River
Estuary
5 770 000
5 710
0.363
0.06
Egypt
Nile River
Gfla
2 978 000
89.2
0.111
1.25
China
Yellow River
Shanxian
688 384
43.2
1.62
37.6
China
Yangtze River
Datong
1 700 000
921.1
0.478
0.52
China
Pearl River
Wuzhou
329 725
227
0.0718
0.32
China
Yongding River
Guanting
50 800
1.40
0.081
57.8
56
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
(a) Straight (Guankou Reach, Xishui River, China)
(b) Meandering (Chencun Reach, Weihe River, China)
(c) Braided (Wandering) (Huayuankou Reach, Yellow River)
(d) Branched (Maanshan Reach, Yangtze River) 1, 2, 3, islands
Figure 4.3 — River patterns.
(2)
(3)
Meandering: Meandering rivers consist of a series of bends
of alternate curvature connected by straight crossings
(Figure 4.3 (b)), and the slopes are usually relatively flat.
The natural meandering channels are unstable, with bank
caving at the downstream part of concave bands. There are
deep pools in the bends and high velocities along the outer
concave banks. The depth at crossings is relatively shallow
compared to the depth at bends.
Wandering: The river channels of wandering rivers are wide
and shallow and divided by numerous unstable mid-bars.
(4)
They have a braided appearance at low flow, but all the bars
are inundated or destroyed at the flood stage. The banks are
poorly defined and unstable, and the main stream frequently
and rapidly shifts from one side to the other. The subsidiary
channels are also unstable and often change in flood seasons
(Figure 4.3 (c)).
Anabranched or branched: The appearance of anabranched
rivers is similar to that of branched rivers, but the mid-sand
bars are higher and more stable, and some of them become
the islands lived on and cultivated by local people, and can
Table 4.2
Classification of river patterns by different authors
River pattern
Author
Meandering
Non-meandering
Leopold (USA)
Meandering
Straight
Braided
Rosinski (Russian
Federation)
Meandering
Periodic widening
Wandering
Contragies (Russian
Federation)
Free
meandering
Non-free
meandering
Single channel
Branched
Xie (China)
Meandering
Straight
Fang (China)
Meandering
Mid-island
Chien (China)
Meandering
Straight
Anabranched
Wandering
Lane, Chang (USA)
Meandering
Straight
Steep slope braided
Mild-slope braided
Simons (USA)
Meandering
Straight
Ling (China)
Stable
meandering
Unstable
meandering
Straighteningmeandering
Branched
Wandering
Shifting
Braided
Stable
branched
Shifting
branched
CHAPTER 4 — FLUVIAL PROCESSES
be inundated only by extraordinary floods. The channels of
anabranched rivers are divided by stable and high islands
into more than two branches. One is the main channel and
the others are subsidiary channels. The main channel and the
subsidiary channels are also relatively stable, but can be
changed under some flow and sediment transport conditions
(Figure 4.3 (d)).
4.3.2
Methods for classification of river patterns
A prerequisite for the systematic study of fluvial processes is to
classify the river patterns according to the plan morphology (static
condition) and the features of evolution (dynamic condition) of the
river. However, until now, there has been no unified method used
for such classification. For example, Leopold and Wolman (1957)
classified rivers into the categories of meandering, straight and
braided according to the plane morphology of the rivers. Fang
(1964) classified rivers as mid-island, meandering and shifting
based on the coefficient variation of peak flood discharge (Cv);
the ratio of incoming sediment concentration to sediment carrying
capacity, and the ratio of the maximum width of water surface
during floods to the width of the channel. Chien, et al. (1987) and
Xie, et al. (1987) stressed the static and dynamic features of rivers
and classified rivers as straight, anabranched or branched, meandering and wandering. The static features of rivers denote the
planform, configuration, mega-bedform, and topography of river
channels. Dynamic features include scope and intensity of main
current shifting, migration of the main channel, strength of deposition and erosion in the main channel and of the banks, etc. Table
4.2 shows the classification of river patterns suggested by different
authors (Chang, 1988; Ling, 1963; Xie, 1980; Simons, 1979;
Rosinski, 1950).
Leopold’s classification is much more simple and
generalized. However, according to the experience of Chinese
river scientists, if the Middle and Lower Yangzte River and the
Lower Yellow River were classified into the same river pattern, the
braided pattern, this would be quite inappropriate and the
classification of river patterns would lose its significance. The
Middle and Lower Yangzte River, having 41 branched reaches
with a total length of 817 km, has high lands, high mid-bars, and
its channels are relatively stable, while the Lower Yellow River
has its wandering main current with large shifting scopes in a
transversal direction, and a changeable and unpredictable
configuration over a length of 275 km, with a wide channel bed
and dense and scattered mid-bars. These two rivers, in fact, reflect
two different river patterns with different fluvial processes.
Therefore, Chinese scientists prefer to classify the two rivers into
the anabranched and wandering categories, respectively, rather
than the braided category. As for piedmont rivers with a large
slope, coarse bed sediment, low mid-bars and stable filaments in
low water periods, but with apparent deformation of channel in
high water periods, they are still classified in the wandering
category.
4.3.3
Characteristics of rivers with different patterns
Sometimes a river has the features of two river patterns. For
example, the Brahmaputra River at the India-Bangladesh border
has the appearance of a braided pattern. It is a wandering river, and
while the dynamic features of the river in Bangladesh are those of
a wandering river, the stable islands occupied by local people are
3–4 m above the low water level, and some of them have existed
57
for more than 100 years. It is thus a wandering-anabranched river
(Zhou and Chen, 1998). The upper part of the Lower Yellow River
is a typical wandering river, and its lower part is a typical
meandering river.
4.3.4
Causes for formation of river patterns
The pattern of a river is determined by the characteristics of its
watershed, i.e. (i) incoming runoff and its hydrograph; (ii)
incoming sediment load and its hydrograph, and size distribution
of sediment; and (iii) boundary conditions such as the topography of the valley, geological structure, sediment particles, and
soil composition of the channel and banks. For most alluvial
rivers, boundary conditions play a significant role in the formation of river patterns. If the boundary, including channel bed and
banks, is composed of sand or silt, a wandering river such as the
Lower Yellow occurs. When the channel bed is composed of
sand and silt and the banks have some clay or sandy clay, a
meandering river such as the Jingjiang Reach of the Middle
Yangtze occurs. This conclusion was also proved by the experiments conducted by Ying (1965) and Schumm (1972). Their
experiments were carried out on channel beds with uniform
slopes, and the channel beds and banks were composed of sand.
When bed materials were added, the wandering river ultimately
occurred. Once the bed materials and clay or white bole were
added, the clay or white bole settled on the banks, and then the
meandering river occurred.
In addition to the boundary conditions, supplementary
factors, such as the sedimentation of river bed, scope of discharge
variation, floods features and geographic conditions, etc., also led
to subsidiary effects on the formation of river patterns. Chien
(1987) gave the summarization as shown in Table 4.3.
4.3.5
Transformation of river patterns
The pattern of a river defined by the definite conditions of its river
basin can be transformed when remarkable changes occur in the
natural conditions in the river basin, or after large-scale human
activities. For example, the Murrumbidgee River in Australia was
a typical wandering river in ancient times, when it had a dry
climate, less runoff, worse vegetation, more sediment load and
less clay and silt in its bed material. However, the meandering
river was later transformed, because the climate became wet, rainfall increased, vegetation grew and the incoming sediment load
decreased (Schumm, 1968). The Missouri River in the United
States was a meandering river in the 19th century. The vegetation
and forests on the banks and floodplains were destroyed by floods,
especially the 1881 flood, and the river gradually widened and
straightened (Schumm, 1971). If a reservoir is put into operation,
the river pattern in the reach downstream of the reservoir is
changed. For example, after the Sanmenxia Reservoir on the
Yellow River was impounded in the early 1960s, the number of
regulated low floods increased, and the channel bed downstream
of the reservoir was eroded by the clear water released from the
reservoir. The wandering reach of the Lower Yellow River thus
tended to be transformed into a single, meandering channel.
4.3.6
Critical relationships between different river patterns
There are some critical conditions in distinguishing between
different river patterns. If a factor of a river pattern such as the
longitudinal slope is close to a critical value, a small change of the
factor may result in a great change in the river pattern.
58
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
Table 4.3
Conditions for different river patterns
Condition
Composition of
band material
Boundary
condition
Node point
Meandering river
Straight river
Loose particles
with low-erosionresistance
Materials lying
between wandering
and meandering rivers
on both side banks
Two-layer structures
having erosion
resistance on both
side banks
Banks composed of
more clay or having
more vegetation
–
Node point control
lies at entrance or
exit of branch,
and transversal
free shifting
is restricted
–
Side banks controlled
by node points with short
intervals or wide distribution of exposed bedrock
on both side banks caused
by geological tectonic
movement
–
–
–
Relatively large
amount of
incoming bed
material load
Low incoming bed
material load with a
certain amount of
wash load
Low incoming bed
material load with a
certain amount of
wash load
Longitudinal erosion
and deposition
are basically in
equilibrium
Longitudinal erosion
and deposition
are basically in
equilibrium
Channel deposition
in moderate and low
water encourages
the development of
a wandering river
–
Weakened erosion
in flood season
and weakened
deposition in
non-flood season
–
Range of
discharge
variation
Large range
of discharge
variation
Small ranges of
discharge variation
and coefficient
variation of
flood discharge
Small ranges of
discharge variation
and coefficient
variation of
flood discharge
–
Rising and
falling of flood
Sharp rising and
falling of flood
Slow rising and
falling of flood
Slow rising and
falling of flood
Incoming
sediment load
from watershed
Equilibrium of
longitudinal
erosion and
deposition
Yearly erosion
and deposition
Incoming
runoff
condition
Branched river
Standing of downstream water level
in flood season
benefits maintenance
of meandering river
Water level
withstanding
Incoming
sediment
condition
Wandering river
Slope of valley
Geographical site
Accumulation
in past.
Channel
aggradation
is beneficial to
formation of
wandering
Steep slope
Smooth slope
On alluvial fan
out of gorge or
upper part of
alluvial plain
On middle and
lower part of
alluvial plain
Smooth slope
On middle and
lower part of
alluvial plain
withstood by main
river or lake in
flood season
–
–
–
Slope of straight rivers
on estuarine is
small, but that having
exposed bedrock or dense
vegetation on both sides
of banks could be formed
and developed under
various slopes
Straight river with exposed
bedrock or dense vegetation on both sides of banks
could be formed and
developed under different
geographic sites
CHAPTER 4 — FLUVIAL PROCESSES
4.3.6.1
RELATIONSHIPS BETWEEN LONGITUDINAL SLOPE AND
RIVER PATTERNS
The geological site of a river plays a great role in the formation of
its river pattern. A river flowing out of a gorge, with a steep slope,
would easily develop into a braided-wandering river, while a river
on a plain, with a smooth slope, would be a sinuous (meandering)
river. An empirical relation was established by Chien and Zhou
(1965).
S = 0.01Qn–0.44
(4.1)
where Qn is the bankfull discharge in m3 s–1, and S is the longitudinal slope in 1/10 000. The rivers in the region above the S-Q line
belong to the wandering pattern, while those in the region below
the line belong to the meandering pattern. For rivers of the same
size, the rivers develop from a meandering to a wandering pattern
as the slope increases.
∆Q
Θ=(
0.5 TQn
)(
ds
59
)
0.6
D35
(
Qmax − Qmin
Qmax + Qmin
)
0.6
b 0.45 bmax 0.3
( )
(
) (4.6)
d
b
where ∆Q is the rising range of flood discharge in m3 s–1, Qn is
the bankfull discharge in m3 s–1, T the duration of floods in days,
d is the depth under bankfull discharge in m, S is the slope, D35 is
the grain size of bed material for 35 per cent finer in mm, Qmax
and Qmin are the maximum and minimum daily discharges in the
flood season in m3 s–1, b is the channel width under bankfull
discharge in m, and bmax is the surface width under the historical
highest water level including width of flood plains in m. The
ranges of available data are: Q, 242–92 600 m3 s–1; Qn, 35–58 500
m3 s–1; d, 0.32–17.0 m; b, 85–3 010 m; D50, 0.06–0.32 mm; and
C, 1.7–1 010 kg m–3. Θ > 5 is a wandering (braided) river, Θ < 2
is a non-wandering river, and Θ = 2–5 is a transitional river.
4.3.6.5
RELATIONSHIPS BETWEEN LONGITUDINAL SLOPE, BED
SEDIMENT AND DISCHARGE
4.3.6.2
RELATIONSHIPS BETWEEN LONGITUDINAL SLOPE AND
MEAN DISCHARGE
Lane (1957) established two relationships between longitudinal
slope and mean discharge:
S=
0.0041Qm–0.25
S=
0.0007Qm–0.25
Based on the theory of flow power, Chang (1988) established four
regions by three critical straight lines for the reactions among
longitudinal slope, bed sediment and discharge.
Critical straight line 1:
Sc/d1/2 = 0.00238Q–0.51
(4.7)
Critical straight line 2:
S/d1/2 = 0.05Q–0.55
(4.8)
Critical straight line 3:
S/d1/2 = 0.047Q–0.51
(4.9)
(4.2)
(4.3)
m3 s–1,
and S is the
where Qm is the average annual discharge in
longitudinal slope. The rivers in the region above Equation 4.2
belong to the wandering pattern, while those below Equation 4.3
are meandering. The rivers in the region between Equations 4.2
and 4.3 are in transition from meandering to wandering.
4.3.6.3
RELATIONSHIPS BETWEEN LONGITUDINAL SLOPE AND
MAXIMUM DISCHARGE
Romashen (1977) analysed the data of valley slopes and average
maximum discharge from 250 reaches of the rivers in the former
USSR and divided the rivers into branched, un-shaped meandering, and meandering patterns. He concluded that the critical
condition between branched and un-shaped meandering rivers is:
Qmax S = 1.4
(4.4)
and the critical condition between un-shaped meandering and
meandering rivers is:
Qmax S = 0.35
(4.5)
where Qmax is the maximum discharge in m3 s–1, and S is the
longitudinal slope. According to Romashen’s analysis, branched
rivers could also be divided into the channel bed branched pattern
(corresponding to wandering) and the floodplain branched pattern
(corresponding to branched). A channel bed branched river has a
steep valley slope and low discharge, while the flood plain
branched river has a mild valley slope and large discharge.
4.3.6.4 WANDERING INDEX
Chien, et al. (1965) analysed the data from 21 stations on the
Yangtze River, the Yellow River and other plain rivers in China,
and found the following wandering index:
where Sc is the critical slope corresponding to bed load, d is the
medium size of bed material in mm, Q is the bankfull discharge in
cfs, and S is the longitudinal slope. If the unit of discharge is in
m3 s–1, the coefficient in Equation 4.7 is 0.000386; in Equation
4.8 it is 0.00704; and in Equation 4.9 it is 0.00763.
The rivers in the region between critical straight lines
1 and 2 are meandering or straight; the rivers in the region
between critical straight lines 2 and 3 are straight or braided, and
the rivers in the region above critical straight line 3 are mild slope
braided and steep slope braided, which are separated by a hypothetical straight line.
4.3.7
Indexes of river stability
The characteristics of alluvial processes are determined by the
added conditions of the river basin. The index of river stability
is a mark to express the local, temporal and relative variation of
the river channel when the incoming runoff and sediment load
from the watershed change over time. The stability of a river
and its equilibrium are two different concepts. The latter
denotes that, as a whole, no erosion or deposition occurs in the
river when the incoming sediment load carried by the flow from
upstream reach is equal to the sediment carrying capacity of the
flow. Obviously, a stable river is not often a river in equilibrium. Similarly, an equilibrium river is not necessarily a stable
river (Chien, 1958).
The index of river stability can be divided into indices of
longitudinal (river channel) stability and transversal (river bank)
stability.
4.3.7.1 LONGITUDINAL STABILITY OF RIVER CHANNELS
Longitudinal stability denotes the variability of channel bed due to
aggradation and degradation of the bed along the river. It depends
60
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
on the erodibility of bed sediment and flow intensity, and can be
expressed by the Rohkin and Chien numbers.
(1)
Rohkin Number. The longitudinal stability of a river
channel depends on the ratio of the tractive force acting on a sediment particle by flow to the resistance force against the motion of
the particle, and can be expressed by the Rohkin Number, i.e.:
Φ = d/s
(4.10)
where d is the grain size of bed material (d35 or d50 in mm), and S
is the channel slope in 1/1 000. The larger the parameter, the more
stable the channel.
(2)
The Chien Number (1958, 1987). The stability of an
alluvial river is determined by the incoming runoff and sediment
load from its river basin. For a quasi-equilibrium river, the sediment carrying capacity is equal to the incoming bed material load
and depends on the incoming flow and boundary conditions. The
relative stability can thus be expressed by the hydraulic parameter
of sediment carrying capacity, i.e.,
K = D/dS
(2)
The Xie Number (1987). The transversal stability of a
natural channel is related to the channel banks and can be
expressed as follows:
C = b/B
(4.13)
where b is the channel width for low water in m, and B is the
channel width under the dominant discharge in m (see section
4.3.1). The larger the parameter C, the narrower the main channel,
thus, the more stable the banks. The parameters of Ψ and C for the
Yangtze River and the Yellow River are listed in Table 4.5.
4.4
MORPHOLOGY OF RIVERS
Under the effects of flow action over a long period, an alluvial
river may be in a quasi-equilibrium state through the self-adjusting
action of the channel. Some functional relationships exist between
the river morphology, including cross-sectional geometry and
longitudinal profile, and river basin factors. These relationships
are called hydraulic geometry equations, i.e.,
(4.11)
b = F1 (Q, Go, Do); d = F2 (Q, Go, Do); s = F3 (Q, Go, Do) (4.14)
where D is the grain size of bed material (d35 or d50 in mm), d is
the depth under dominant discharge in m, and S is the longitudinal
slope in 1/1 000. The larger the parameter K, the more stable the
channel. The parameters of Φ and K of the Yangtze and Yellow
Rivers are listed in Table 4.4.
Obviously, the physical meanings of the above two
numbers are the same, although they are derived using different
approaches.
where b is the channel width, d is the channel depth, s is the longitudinal slope of the river, Q is the incoming water discharge and
its hydrograph from the upper reach; Go is the incoming sediment
load and its hydrograph from the upper reach, and Do is the size
composition of incoming sediment load.
For the incoming sediment load with different grain
sizes, only the bed material load has an effect on channel formation. Once the incoming bed material load settles down, it
becomes the material composing the channel boundary and also
plays an important role in channel stability and cross-section form.
Therefore, Equation 4.14 can be transformed as follows:
4.3.7.2 TRANSVERSAL STABILITY OF RIVER CHANNELS
(1)
The Altounin Number (1962). The transversal stability
of a channel is related to the stability of the banks. The main
factors affecting transversal stability are flow direction, erodibility
of band soil and elevation difference between floodplain and main
channel. It is highly complicated and has not been fully studied,
but some indirect relationships have been obtained.
Ψ = Q0.5/S0.2b
(4.12)
where Q is the dominant discharge in m3 h–1; b is the channel
width under the dominant discharge in m, and S is the slope under
the dominant discharge in m km–1. The larger the parameter Ψ,
the more stable the channel.
Table 4.4
Φ and K
River
Reach and river pattern
Yangtze River
Jingjiang, meandering
Wuhan, branched
Nanjing, branched
Yellow River
Upstream of Gaocun,
wandering river;
From Gaocun to
Taochengpu,
transitional reach
Brahmaputra
River
Noonkaw-Aricha,
wandering-branched
Φ
K
2.9–4.1
6.7–7.8
7
0.27–0.33
0.39–0.52
0.35
0.31–0.47
0.42–0.54
0.18–0.21
0.17
b = f1 (Q, G, D); d = f2 (Q, G, D); s = f3 (Q, G, D)
where G is the incoming bed material load and its hydrograph, and
D is the boundary conditions, including the composition of
channel bed and banks.
Because morphological relationships depict the relationship between rivers suited to the conditions of incoming runoff
and sediment load and the channel boundary, they have become
the basis of the hydraulic computation of alluvial rivers, prediction
of fluvial processes and river training, etc. and have had the most
significance in river engineering.
4.4.1
Dominant discharge
The dominant discharge is such a discharge that its channelforming effects are equivalent to the comprehensive actions
Table 4.5
Ψ and C
0.65–1.81
Ψ
C
Jingjiang,
meandering river
0.87–1.56
0.67–0.77
Upstream of Gaocun,
wandering river;
From Gaocun to
Taochengpu
0.18–0.45
0.09–0.17
0.48–0.75
0.17–0.20
River
Reach and river pattern
Yangtze River
Yellow River
2.8–7.9
(4.15)
CHAPTER 4 — FLUVIAL PROCESSES
DETERMINATION OF DOMINANT DISCHARGE
(1) Makaviev’s method (1955) — The effect of
discharge on channel formation depends on its sediment-carrying
capacity and time duration. The sediment-carrying capacity can be
expressed by the product of Qm and S, where Q is the discharge; m
is an exponential; and S is the longitudinal slope. Let the frequency
of occurrence of the discharge be p. The discharge corresponding
to the maximum QmSp has the maximum effect on channel formation and can be adopted as the dominant discharge. Procedures for
determining the dominant discharge are listed as follows:
(i) Divide the long-term measured hydrograph at a cross-section
on the studied reach into a number of discharge grades.
(ii) Calculate the occurrence frequency of each discharge grade.
(iii) Draw the discharge-slope relationship and determine the
mean slope corresponding to each discharge grade.
(iv) Compute the product of QmSp for each graded discharge.
Draw the relationship between flow discharge and sediment
discharge on a logarithmic paper, where the exponential m is
the slope of the curve line. Generally, for plain rivers, m is
equal to 2.
(v) Draw the QmSp-Q relationship.
(vi) Find the maximum QmSp. The discharge corresponding to
the maximum QmSp is the dominant discharge.
As shown in Figure 4.4, there are two peak values of
QmSp. The discharge corresponding to the first peak of QmSp is
analogous to long-term average maximum flood discharge with an
occurrence frequency of 0.5–6.5 and 3 per cent on average, and its
water level corresponds to the bankfull water level. It is called the
first dominant discharge. The discharge corresponding to the
second peak of QmSp is slightly higher than the long-term average
discharge with an occurrence frequency of 17.5–44.5 per cent and
30 per cent on average, and its water level corresponds to the
elevation of the point bar. It is called the second dominant
discharge.
Discharge (m3 s–1)
4.4.1.1
€
Figure 4.4 — Relationship between QmSp and Q.
Generally, the first dominant discharge is used as the
dominant discharge to determine the river morphology of the
channel for moderate discharges. The second dominant discharge
moulds the channel of low discharge and is applied to the regulation of navigation course.
(2) Chien’s method (1987) — The channel-forming
effect of a discharge depends on its corresponding sediment
discharge and time duration. As shown in Figure 4.5, when
drawing the curves of sediment discharge (curve A), frequency
(curve B) and the product of sediment discharge and frequency
(curve C) for various grades of discharge, the discharge corresponding to the maximum value on curve C is the dominant
discharge.
Benson and Thomas (1966) calculated the dominant
discharge with the same method based on the data from nine rivers
in the United States. The results indicate that the occurrence
frequency of the dominant discharge is 7.6–19.5 per cent, with the
average of 12.4 per cent. In their calculation, the sediment
discharge is only for the suspended load.
(3) Chikurimora’s method (1969) — This method used
the weighted sediment discharge to determine the dominant
discharge, i.e.,
n
Qd =
∑ QsiQi
i=1
n
∑ Qsi
(4.16)
i=1
where Qd is the dominant discharge in m3 s–1, Qi is the discharge
of i grade in m3 s–1, Qsi is the sediment discharge corresponding
to the discharge of i grade in t s–1, and n is the number of divided
grades.
4.4.1.2 BANKFULL DISCHARGE
The field and experimental data indicate that the velocity in river
channels increases with the rising of water levels. The effect on
channel formation is greatest when the water level is at the elevation of floodplains. The flow disperses and the effect on channel
formation is weakened when the water level rises further. Andrews
(1980) also concluded that bankfull discharge corresponds to the
discharge when sediment transport is the strongest. It is thus
reasonable that bankfull discharge can be used as the dominant
discharge. When determining the bankfull discharge for a river
reach, the reach should be of sufficient length, and some crosssections and their corresponding water levels should be measured
in the reach so as to avoid shortcomings caused by using data from
only one or two cross-sections. Bankfull discharge is determined
directly according to the water level corresponding to the elevations of the floodplains along the reach. This method is similar to
(A) Sediment discharge
(B) Frequency
(C) A× B
produced by discharge according to a long-term discharge hydrograph. Dominant discharge has the greatest influence on the
molding of river channels. Although the highest flood has great
potential, it cannot play a maximum role in channel formation
because of its short time period. Low water flow has a long duration, but it cannot play the maximum role either because it has a
small volume of discharge. Therefore, dominant discharge is a
rather large discharge, instead of the maximum flood discharge.
61
Figure 4.5 — Determination of dominant discharge.
62
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
that proposed by Leopold (1964). For determining the bankfull
stage, Riley (1972) also suggested measuring the width-depth ratio
of the cross-section at different water levels. The width-depth ratio
decreased with an increasing water level, and reincreased with an
increasing water level when the flow was over the flood plains.
Riley concluded that the water level corresponding to the turning
point of the relationship between the width-depth ratio and water
level was the bankfull stage.
It should be pointed out that under some conditions, for
example when the channel cross-section is not regular, the natural
levees along the flood plains are higher than the flood plains, or
the mountainous or upland rivers have no flood plains, etc., the
accurate determination of the bankfull stage is somewhat difficult.
where Qb is the bankfull discharge in m3 s–1, A is the wetted area
of cross-section at bankfull stage in m2, and S is the slope of the
water surface.
(4) Relationship between bankfull discharge and annual
average discharge. Based on the data published by Shumm (1968)
and Carlton (1965), Chang (1979) established the relationship
between bankfull discharge and annual average discharge, as
shown in Figure 4.6.
4.4.1.3
(2)
EMPIRICAL EXPRESSION FOR BANKFULL DISCHARGE
(1) Expression of the Institute of Hydraulic Research of
the YRCC (1978):
4.4.1.4
INTERVALS
(1)
(3)
Qd = 7.7Qf
0.85
+ 90Qf
1/3
(4.17)
where Qd is the dominant discharge (bankfull discharge) in m3 s–1,
and Qf is the long-term average discharge in flood seasons in
m3 s–1.
(2) Hey’s expression (1975):
Qb = 1.06A0.8
(4.18)
(4)
where Qb is the bankfull discharge in m3 s–1, and A is the area of
the watershed (km2).
(3) Williams’ expression (1978):
(5)
Qb = 4.0A1.21S0.28
(4.19)
BANKFULL DISCHARGE ESTIMATED BY RECURRENCE
Leopold (1964) found that the recurrence interval of bankfull
discharge was 1.5 years, based on the data from 13 stations
in the eastern United States.
Nixon (1959) concluded that bankfull discharge had an
average frequency of 0.6 per cent based on data from rivers
in England and Wales.
Pickup (1976) found that there were two bankfull discharges
based on data from the intermittent river in the Cumberland
River basin in Australia. One corresponded to a flood with a
recurrence interval of 20 years and played a role in the
formation of river banks, size and shape of the main channel.
Another corresponded to the floods occurring 3 to 5 times a
year, which determined channel width and the slope of low
water.
Emmett (1975) found that the recurrence intervals of bankfull discharge was 1.5 years.
Chien (1987) concluded that the use of a discharge with a
recurrence frequency was reasonable. He made a simplification for the computation of the dominant discharge. A flood
with an interval of 1.5 years might be roughly applied as the
dominant discharge if there were not enough data.
Table 4.6
Coefficient β for different material
Calcareous limerock
Limestone
Dolomite
Apos and stone
β
0.017
0.010
0.008
0.003
Bankfull discharge (cfs)
Distance from gorge outlet (km)
Material
Annual average discharge (cfs)
Figure 4.6 — Relationship between bankfull and annual average
discharge (after Chang, 1979).
Water surface differences between
gorge outlet and stations (m)
Figure 4.7 — Longitudinal profile of the Lower Yellow River,
CHAPTER 4 — FLUVIAL PROCESSES
4.4.2
Longitudinal profiles
The longitudinal profile is the result of long-term actions between
flow and channel bed. The shape of the longitudinal profile
depends on the incoming runoff and sediment load from its watershed and the geology of the channel, while the elevation of the
longitudinal profile is controlled by the downstream base of
erosion.
Seen in detail, a longitudinal profile of a natural river is
a smooth curve, and can be categorized into three types: sunken,
protruding and straight. For most plain rivers, the longitudinal
profiles are sunken, but for some mountainous rivers in the front
reach, they are protruding. The longitudinal profiles for both
mountainous and plain rivers are fluctuated and are saw-toothed
because of different geology, channel widths, and the existence of
pools, crossings and sand bars.
(3)
4.4.2.1 GEOMETRIC EXPRESSIONS OF LONGITUDINAL PROFILES
The longitudinal profile can be expressed by exponential or
semi-logarithmic curves when there are no break points along
the river.
(1)
Sternberg’s expression (Chien, et al., 1987). If there are
no different materials coming from the tributaries, the bed materials of an alluvial river become finer and finer because they suffer
from wear and tear when moving along the river. The relationship
between the weight of bed material and the distance it moves can
be expressed as follows:
4.4.2.2
e–β1L
W = Wo
The Yivanov Expression (1951)
h = H (l/L)n
F = 2.45L0.60
(4.26)
where F is the difference of water surface from the outlet of the
river gorge to a gauge station in m, and L is the distance from the
outlet to the station in km (Figure 4.7).
EMPIRICAL RELATIONSHIPS BETWEEN LONGITUDINAL
SLOPE AND WATERSHED FACTORS
(1)
Relationships between longitudinal slope and the
composition of the channel bed. From the viewpoint of sediment
transport equilibrium, the coarser the sediment in the channel, the
steeper the channel slope. The opposite is also true. When bed
sediment becomes finer, the channel slope becomes smoother.
(a) For the channel slope of main streams and tributaries of the
Middle and Lower Yellow River (Chien, 1987):
(4.27)
(4.20)
(b)
For the slope of the Jingjiang Reach of the Yangtze River
(YRWRC, 1959):
S = 25D502.38
(c)
(4.28)
The slope of the rivers in the Central Asian part of the former
USSR can be expressed as follows (Altwunin, 1957):
(4.21)
S = 0.85D501.10
where Do is the grain size of bed material at the beginning crosssection, D is the grain size of bed material after travelling a
distance L, and β is the wear of the coefficient particle as shown in
Table 4.6 (Sedimentation Committee, 1992).
(2)
Shulits’ expression (1941). The channel slope is directly
proportional to the grain size of bed material, so:
S = Soe–α1L
(4.25)
where h is the elevation at a certain position, H is the elevation at
the origin of a river, L is the horizontal distance from the origin to
the estuary, l is the distance from the position to the estuary, and n
is the morphological exponent.
(4)
The Chien Expression (1965). The longitudinal profile
of the Lower Yellow River can be expressed as follows:
S = 41D501.3
where Wo is the weight of bed material at the beginning crosssection, W is the weight of bed material after passing a distance L,
and β is the wear coefficient of the particle. Because the weight of
a sediment particle is directly proportional to its diameter,
Equation 4.20 can be transformed into:
D = Doe–β1L
63
(4.22)
where So is the channel longitudinal slope at the beginning crosssection, S is the channel longitudinal slope at a position with a
distance of L from the beginning cross-section, L is the distance,
and α is the change of the coefficient slope, which is related to the
materials of channel bed and banks. Based on the data from the
reach downstream of Otowi on the Rio Grande River, the following
expressions are obtained (Sedimentation Committee, 1992).
(4.29)
where S is the longitudinal slope in 1/10 000, and D 50 is the
median size of bed material in mm.
(2)
Relationsihps between longitudinal slope and discharge
or watershed area. Since sediment discharge is proportional to a
high power of flow discharge, the relationship between slope and
discharge reflects to some extent the relationship between slope
and the incoming runoff and sediment load from the river basin. If
there are not enough measured data, the watershed area can be
used instead of the discharge.
(a) The slope for rivers in Siberia, Russia (Makkaveev, 1959):
S=
250
Q
(4.30)
0.43
(b)
The slope for rivers in China (Li, 1965):
D50 = 0.47e–0.0059L
S = 0.022e–0.0092L
(4.23)
S=
(4.24)
where D50 is the median size of bed material in mm, and L is the
distance in m.
20.9
Q
(c)
(4.31)
0.27
The slope for rivers in the eastern United States (Hack,
1957):
64
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
2 4
 D  0.6
S = 60  50 
 A 
 
(d)
2
S = 1.15η (
(4.32)
The slope for the mainstream and tributaries of the Yellow
River (Chien, 1965):
 D  0.34
S = 296  50 
 A 
 
 0.59
1 / 2

 D 
50




n
Q
w
C
+ 0.385 D50
−0.52
1.10
0.11
D50 Gb
(4.34)
(4.35)
(4.36)
where S is the longitudinal slope, Qn is the bankfull discharge in
m3 s–1, D50 is the median size of bed sediment in mm, and Gb is
the sediment discharge of bed load under bankfull discharge in
m3 s–1.
Hey (1982) extended his data to 66 stable reaches on the
Rivers Wye, Severn and Tweed, and obtained the following
expression:
−5.3
S = 0.679Qn
0.97 0.13
D50 Gb
2
+
2
2
2
0.43 ∂ Vos C α VobQ 2 / 9
(
)
8/9
2 2
∂x
k β
g
(4.38)
2
kα VobQ
( 2
)
∂x β gVos C
1/ 3
4
2
where n is the Manning coefficient; w is the mean fall velocity in
cm s–1, C is the sediment concentration in kg m–3, Q is the dominant discharge in m 3 s –1, and D 50 is the median size of bed
sediment in mm.
(c) Longitudinal slope for some rivers in England
Hey (1982) obtained the following expression based on
the data from 25 stable crossings on the River Wye (a gravel-bed
river):
S = 1.02Qn
+ 0.807
∂
2 /9
S = 1.15η {
3.5 1.28 1.28
0.357
2
K α Vo δ Q
2
)
(4.33)
where Qn is the bankfull discharge in m3 s–1, C is the sediment
concentration of bed material load corresponding to bankfull
discharge in kg m–3, and D50 is the medium size of bed sediment
in mm.
(b) Longitudinal slope of the Lower Weihe River (a main tributary of the Yellow River) (North-west Institute, 1962)
S = 2620
4 2
4
After simplification,
where Q is the dominant discharge (Equation 4.30) or bankfull
discharge (Equation 4.31) in m3 s–1, A is the watershed area in
km2, D50 is the median size of bed sediment in mm, and S is the
slope in 1/10 000.
(3)
Relationships between longitudinal slope and factors of
sediment.
(a) The slope of rivers in China (in laboratory) (Li, 1965):

 C
S = 45.5 
Qn

4
β g VosC
(4.37)
in which the bankfull discharge was calculated using the correction Colebook formula and the sediment discharge of bed load
was calculated using the Meyer-Peter formula.
(4)
Theoretical solution of longitudinal slope for alluvial
rivers and estuaries.
Dou (1964) deduced an equation of slope from the
theory of minimum activity for alluvial rivers and estuaries.
4
g Vos C
4 2
k α
4
2
VobQ
}
2 /9
(4.39)
where S is the slope, Q is the long-term average discharge, C is the
long-term average sediment concentration, Vos, Vob are the halting
velocity of suspended load and bed material, respectively, α is the
bank
relative stability of the channel banks and channel bed, α———
αbed
(when the stability of the channel banks is close to the stability of
the channel bed, α = 1.0); β the coefficient of tidal wave (β =
∆H
1 + 0.35 H , where ∆H is the tidal difference, H is the mean depth
under moderate tidal stage) and for the general estuary and nontidal estuary β = 1, K is the parameter, K = 0.055 γsη (√g/c)6,
where γ s is the specific weight of the sediment particle; g is
gravity acceleration; c is the Chezy coefficient, η is the ratio of
bottom velocity to mean velocity, and δ is the ratio of average
sediment concentration to near-bed sediment concentration under
saturation. Values for αbank and αbed for different materials are
listed in Table 4.7.
4.4.3
Cross-sectional morphology of rivers
The cross-sectional morphology of rivers can be divided into two
categories, the morphology of cross-sections at a station, and
cross-sectional morphology along rivers. The morphology of a
cross-section at a station means the changes in sizes of the crosssection in a short reach or at a cross-section under a different
discharge. It reflects the changes in geometry of the wetted crosssection. The cross-sectional morphology along rivers implies the
changes in channel geometry of different rivers, or in the upper
and lower reaches of the same river, caused by different incoming
runoff and sediment load and conditions of channel boundaries.
The data at different cross-sections of different rivers or different
reaches of the same river are unified by the bankfull discharge or
the discharge corresponding to a certain occurrence frequency.
They reflect the changes in the channel geometry of the rivers.
The two cross-sectional morphologies cannot be obscured,
because the changes in bed sediment and slopes along the river are
substantially greater than those at a cross-section.
Table 4.7
Indexes of soil stability, αbank and αbed
Material composing banks
and channel bed
Coarse sand
Medium-coarse sand
Medium sand
Fine sand
Silty sand
Silty clay
Mild clay
Clay
Heavy clay
αbank
(grain size in mm)
(2.0-1.0)
(1.0–0.5)
(0.5–0.25)
(0.25–0.10)
(0.10–0.05)
(0.05–0.01)
αbed
2.5–2.0
2.0–1.5
1.5–1.2
1.1–0.9
1.0–0.8
1.0–0.8
1.7–1.13
2.2–1.8
2.5–2.3
CHAPTER 4 — FLUVIAL PROCESSES
4.4.3.1 HYDRAULIC GEOMETRY
Leopold (1953) concluded that natural rivers in a state of equilibrium have simple exponential relationships between the width,
depth, flow velocity and discharge, like those on the graded canals
in India and Pakistan. These relationships are called the hydraulic
geometry of rivers, and are expressed as follows:
b = α1Q β1; d = α2Q β2; v = α3Q β3
(4.40)
According to the continuity law of flow:
β1 + β2 + β3 = 1
65
for different rivers, as shown in Table 4.8. The bankfull discharge
should be used in these expressions, but the exponentials are
different for different frequencies of dominant discharges
(Table 4.10) (Chien, et al., 1987).
(2)
Ratio of width to depth of cross-sections. Ratios of
width to depth are used to express the shapes of cross-sections,
either the narrow-deep type or the wide-shallow type. Based on
the plain rivers in the former USSR, the National Institute of
Hydrology Research suggested the following expression (Xie,
1987):
b
(4.41)
=ξ
(4.43)
d
and
α1α2α3 = 1
(4.42)
The above expressions are simplified and special examples. In
fact, the coefficients α1, α2, α3 and exponentials β1, β2, β3 are variables in the relationships both for cross-sections and along the
rivers because of the influences of other factors, in addition to
those of the discharge. The exponential in the relationships of river
morphology are listed in Table 4.8 (Chien, et al., 1987).
Chien (1987) suggested that the exponential for the
hydraulic geometry of rivers, β1, β2, β3, can be adopted as 0.14,
0.43 and 0.43 respectively, on average, as shown in Table 4.9.
4.4.3.2 HYDRAULIC GEOMETRY ALONG RIVERS
(1)
Changes of exponential. The exponentials in Leopold’s
expressions for the hydraulic geometry along rivers are changeable
where b and d are the average width and depth, respectively, of
cross-sections in a reach corresponding to the bankfull discharge
Table 4.9
Average β1, β2 and β3
β1
β2
β3
0.16
0.43
0.42
Average value of data at 158 stations
Data from 10 stations on the Rhine
River in Europe
0.12
0.13
0.45
0.41
0.43
0.43
On average
0.14
0.43
0.43
Sources of data
Average value of data at 206
cross-sections on the River Ryton in
the United Kingdom
Table 4.8
Exponential in relationships of hydraulic geometry for rivers throughout the world
Country
China
€
United States
United Kingdom
Canada
River
6 small wandering rivers in north China
Meandering in the Lower Yellow River
Bend reach
Straight reach
Jingjiang Reach of the Yangtze River
(average of 3 cross-sections)
Rivers in middle and west regions
Intermittent rivers in semi-arid regions
16 rivers in Central Pennsylvania Stale
Brandywine Creek in eastern United States
Data from 158 stations
White River in Washington State
Small river on beaches affected by tide
27 rivers in England and Wales
23 gravel rivers in England and Wales
17 rivers in low land in England
17 rivers in high land
3 small rivers in southern England
Cross-sectional relationship
Longitudinal relationship
β1 β2 β3
β1 β2 β3 Q
0.48 0.35
0.16 0.30
0.28 0.18
0.08 0.46 0.46
0.26 0.40 0.34
0.29 0.36 0.34
0.4 0.41 0.55
0.12 0.45 0.43
0.38 0.33 0.27
0.09 0.13 0.78
0.40
0.30
0.36
0.45
0.10 Qm
0.20
0.09 Q2.33
0.13 Qn
0.76 0.20 0.04
discharge when
velocity is maximum
0.49
0.45
0.53
0.52
0.27
0.40
0.40
0.32
0.24 Qn
0.15 Qn
0.07 Qn
0.16 Qn
0.13 0.42 0.44
Rivers with gravel-composed banks
and channel bed
20 gravel rivers
12 sand rivers
10 stations on the Rhine River in Europe
0.50
0.50
0.55
0.42
0.50 0.415 0.085 Qn
0.527 0.333 0.140 Q2
0.53 0.32 0.15 Qn
0.13 0.41 0.43
NOTE: Qn is the bankfull discharge; Qm the annual average discharge; and Qa is the discharge corresponding to the occurrence interval of a years.
66
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
in m, and ζ is the geometrical coefficient. For a gravel channel bed
ζ = 1.4, for a coarse sand bed ζ = 2.75, and for a fine sand bed
ζ = 5.5. For the rivers in China, ζ is small for meandering rivers
and larger for wandering (braided) rivers, as shown in Table 4.11.
This expression is widely used in China because of its simple
structure.
where b and d are the width and depth, respectively, and ib and iw
are the percentage of silt and clay in both the banks and channel
bed. Later, Schumm (1968) further obtained the following expression based on the data from the rivers in plains of the United
States and the Murumbidgee River and paleochannels in Australia.
Table 4.10
Different exponentials in Leopold’s expressions for different
dominant discharges
Dominant
discharge
(m3 s–1)
β1
β2
β3
Q50%
Q15%
Q2%
Qn
0.34
0.38
0.45
0.42
0.45
0.42
0.43
0.45
0.32
0.32
0.17
0.05
Manati River
Basin
Q70%
Q50%
Q30%
0.46
0.44
0.46
0.27
0.30
0.32
0.27
0.35
0.25
Bolindin River
Basin
Q50%
Q15%
Q2
0.46
0.54
0.61
0.16
0.23
0.31
0.38
0.23
0.08
Location
River
United States
Brandywine
Puerto Rico
United
Kingdom
ζ
Reach
Yangtze River
Meandering Lower Jingiang River
Meandering Upper Jingiang River
Branched, downstream of Chenlingji
2.55–2.70
2.67–3.27
3.42–3.63
Hanjiang River Meandering
Yellow River
Wandering, upstream of Gaocun
Transitional, downstream of Gaocun
2.0
19.0–32.0
8.6–12.4
(3)
b
(4.44)
where b and d are the average width and depth of a cross-section
corresponding to the bankfull discharge respectively, m is the
exponential and η is the morphological coefficient. Values for m
and η are listed in Table 4.12.
RELATIONSHIPS BETWEEN FACTORS OF WATERSHED AND
=
255
M
1.08
(4.45)
2 iw d
ib b + M
ib b + 2 idw d
b + 2d
(4.49)
−0.66
0.58
(4.50)
where Q2.33 is the discharge corresponding to an occurrence of
2.33 years. The stronger the erosion-resistance of bank materials,
the smaller the channel width.
Bray (1982) used the data from 70 gravel rivers in
Canada and found the following expression:
b
0.2
= αQ2
(4.51)
d
where Q2 is the discharge corresponding to an occurrence interval
of 2 years in m3 s–1. Coefficients α relating to bank materials are
shown in Table 4.13.
Mountain reaches
Mountain foot reaches
Middle reaches of rivers
Lower reaches of rivers
Silt bank
Sand bank
(4.46)
η
M
10–16
9–10
5–9
0.8–1.0
3–4
8–10
0.5–0.8
Table 4.13
Relationship between α and bank material
α
Bank material
(1)
Effects of sediment composition. Based on the data from
90 rivers with areas of 4.4–147 000 km2 and an annual average
discharge of 0.57–146 m 3 s –1, Schumm (1960), obtained the
following expression:
M =
−0.64
b = 33.1Q2.33 M w
HYDRAULIC GEOMETRY ALONG RIVERS
M =
= 34.6 M
Reach
=η
d
0.76
In the formation of alluvial rivers, the size of a cross-section is
determined mainly by the discharge, and the shape of a crosssection, wide or deep, is determined by the composition of bed and
bank materials.
(2)
Effects of materials composing the channel boundary.
Ferguson (1973) used the percentage of silt-clay in bank materials
Mw as a parameter and found the following expression based on
Schumm’s (1968) data.
m
b
(4.48)
Table 4.12
m and η in Altwunin’s expression
d
4.4.3.3
d = 0.51Qm0.29M0.342
d
Altounin’s expression (Xie, 1987):
b
(4.47)
where Qm, is the annual average discharge (all units are in m and
s), and
NOTE: Q6% is the discharge corresponding to an occurrence frequency of 6 per
cent; and Qn the bankfull discharge.
Table 4.11
Geometrical coefficient for rivers in China (Sedimentation
Committee, 1992)
b = 43.7Qm0.38M–0.39
Sand and medium gravel (d > 64 mm)
19.3
Sand and gravel (d < 64 mm)
20.1
Gravel covered by silt
15.2
Silt and clay
14.1
Chien (1963) found the ratio of stable width to depth
under the dominant discharge as follows:
b
d
2.5
= 4 λuc
(4.52)
CHAPTER 4 — FLUVIAL PROCESSES
where λuc is the ratio of threshold velocities of bed and bank
materials when water depth equals 1 m.
(3)
Effects of incoming sediment load. Considering the
relative erodibility of bed and bank materials for more than 60
rivers in China and abroad, Yu (1982) established the following
expressions.
0.5
b = 3.5Qm (
m
)
m
0.40
d
C
D50
d = 0.26Qm (
b
0.28
)
−0.13 −0.10
d 50
−0.18
D50
m
0.10
= 13.5Qm (
)
0.46
D50
C
−0.11 −0.08
d 50
−0.02 −0.02
C
d 50
(4.53)
(4.54)
D50
d
D50
= A1 [
Q
]
2
Z1
(4.57)
ζ
k
g
g
= A2 [
2
]
(4.58)
D50 gD50 S
0.3 / m 0.6
ζ
Under the equilibrium state of sediment transportation, the
hydraulic geometry along rivers can be obtained by the analytic
solutions of the following equations:
Equation of flow continuity
Q = bdu
Equation of flow resistance
U=
1 2 / 3 1/ 2
d
S
n
(4.59)
(4.60)
0.3
0.1 / m
(C
0.73 0.4 0.2
S
W
0.1
n
0.2 / m
D50
d
γs − γ
2
D50
γ
C
(4.63)
)
W
0.3 0.1
Q
0.73 / m
)
(4.64)
)
0.73
W
γs − γ
2
D50
γ
2
γs − γ
γ
1.075
γs − γ
−1.062
gD50
γ
)
γs − γ
D50
γ
)
γs − γ
2
D50
γ
(4.66)
−1.075
(4.67)
gD50
G
(
1.296
gD50
G
(
2
gD50
)
G
(
2
D50
)
Qn
D50
−0.296
gD50
Qn
6
= 3.56 × 10 (
D50
(
Qn
6
)
1.062
(4.68)
gD50
where b is the width of the water surface, d is the average depth of
the main channel, Qn is the bankfull discharge, G is the amount of
bed load transported under the bankfull discharge, and D50 is the
median size of the sediment composition along the channel
boundary.
Table 4.14
A1, A2 and Z1, Z2
ANALYTIC SOLUTION OF HYDRAULIC GEOMETRY ALONG
RIVERS
(4.62)
)
0.3
= 3.09 × 10 (
S = 1.37 × 10 (
Z2
C
W
0.2
(4.65)
)
0.2
0.73 / m
k
Q
where C is the sediment carrying capacity, W is the falling velocity of suspended load, K and m are the coefficient and exponential
in the equation of sediment-carrying capacity of flow, n is the
Manning coefficient, ζ is the morphological coefficient in
Equation 4.43, and Q is the discharge.
(2)
Parker’s expressions. According to the shear stress distribution on the cross-sectional boundary of gravel rivers and bed
load-carrying capacity of flow, Parker obtained the follow solutions (1978).
4
Q
Q
0.2 / m
Q
(
g ζ
k
s=
C
0.1 / m
0.1 0.6
u=
0.6
(
0.2
D50 gD50 S
where Q is the dominant discharge in m3 s–1, b and d are the
average width and depth corresponding to the dominant discharge
in a reach in m, S is the longitudinal slope of the channel; D50 is
the median size of bed material in m, A1 and A2 are the empirical
coefficients, and Z1 and Z2 are the empirical exponentials. A1, A2
and Z1, Z2 are listed in Table 4.14.
4.4.3.4
0.2 / m 0.8
g
b
where Q is the dominant discharge, S is the channel slope, and A is
the coefficient of the stable width of the channel, which is related
to river patterns and ranges from 0.75 to 1.70.
Velikanov’s non-dimensional expression (1958):
b
k
(4.55)
(4.56)
(4.61)
Since there are four unknowns, a supplementary equation must be
added. The solutions reflect the relationship between the factor of
watershed and hydraulic geometry. However, the coefficient and
exponents in these solutions should be calibrated with measured
data, and some revision is needed according to the calibration.
(1)
Xie’s expressions (1980). Xie introduced Equation 4.43
as the supplementary equation, and obtained the following relationships.
b=
where b and d are the average width and depth under long-term
average discharge, Qm is the long-term average discharge, D50 is
the median size of bed material in mm, d50 is the median size of
suspended load in mm, C is the long-term sediment concentration
in kg m–3, and m is the stability index equal to the side-slope coefficient of the bank lying between the historical low stage and the
long-term annual water stage. The data range for long-term
average discharge is 3.6–28 000 m3 s–1; for long-term average
sediment concentration, 0.08–179 kg m–3; the median size of
suspended load is 0.017–0.077 mm; the median size of bed sediment is 0.025–13.5 mm.
Taking the channel slope reflecting the incoming sediment load, Altwunin (1957) obtained the following expression for
rivers in Central Asia, in the USSR.
u3 m
)
gdw
C=k(
Equation of sediment carrying
d=
b = AQ0.5S–0.2
67
River
A1
A2
Z1
Z2
Jingjiang Reach of the Yangtze River 1.16 0.16 0.39 0.38
Wandering reach for rivers in
northern China, and small rivers
in models
15.6 0.27 0.39 0.33
Rivers in the former USSR
5.60 0.29 0.40 0.35
68
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
(3)
Dou’s hypothesis of minimum activity. Dou (1964)
supposed that river channels would mould their cross-sections to
the state wherein channel activity is at a minimum, and established
the following expression as the minimum activity parameter.
Kn =
Q2
[(
Qm
U
b
2
λαUob
) + 0.15 ]
d
(4.69)
C=k
4.4.3.6
(1)
HYDRAULIC GEOMETRY FOR CANALS
Lacey’s expressions (Chang, 1988)
Flow resistance:
where Q2 is the discharge with an occurrence interval of two
years, Qm is the long-term average discharge, λα is the ratio of
stability indices of bank and bed materials, αband/αbed, and αband
and α bed are the stability indices for bank and bed materials
respectively (in Table 4.7), U is the average velocity in the crosssection, Uob is the stop velocity of bed sediment, and b and d are
the width and depth of the channel. The sediment-carrying capacity of flow is expressed as:
U
Q* = Q/ (γs–1 gd550)1/2, Q is the bankfull discharge, γs is the
specific weight of bed sediment, d* is the non-dimensional depth,
–
–
d* = d /d50, and d is the average depth.
–
U = 1.15(f d )1/2
U=
1.346
gdUos
3
b = 1.33(
8
8
kλα uob
)
1/ 9
(4.71)
R
f = 1.6D501/2
(4.81)
Nd = 0.0225f1/4
(4.82)
b = 2.67Q1/2
(4.83)
Hydraulic geometry:
where b is the width of water surface in ft, and Q is the discharge,
in cfs.
f
S=
Canal slope:
5/ 3
1830Q
2
d = 0.81(
8
kλαUobQm
)
1/ 3
(4.72)
gUos c
4
b
d = 1.65(
4 4
2
g Uos c Qm
4 14 14
k λα Uob
)
1/ 9
(4.73)
b = 1.8Q0.5
(4.74)
–
d = 0.116Q0.4ks–0.12
(4.75)
where b is the width of water surface in ft, Q is the bankfull
–
discharge in ft3 s–1, d is the average depth in ft, and ks is the
Nikuradse roughness of sand particles in ft.
(2)
Parker’s expressions (1979). Parker supposed that the
shear stress of the bankfull discharge exceeds the critical shear
stress by 20 per cent, and obtained the following expressions.
Fb = U2/d
b* =
*
(4.85)
where U is the average velocity in ft s–1, and d is the flow depth in
ft.
–
Side slope factor:
Fs = U3/b
(4.86)
–
–
where b is the average width, b = A/d in ft, A is the discharge area
in ft2, and d is the flow depth in ft.
Flow resistance:
U
2
= 3.63(1 +
gdS
C
2330
)(
Ub
ν
)
1/ 4
(4.87)
where υ is the coefficient of kinetic viscosity, and C is the concentration of suspended sediment in ppm.
Empirical value of bed and side slope factors are:
Fs = 1.9d1/2
4.4Q0.5
(4.84)
1/ 6
When Q and D50 are given, the width, depth and slope can be
obtained. As regards the ranges of application of Lacey’s method,
the median size of bed sediment is 0.15–0.4 mm, and discharge is
5–5 000 ft3 s–1. The canal bed is composed of sand and the side
slope is composed of cohesive material.
(2)
Blench’s expressions (Chang, 1988).
Channel bed factor:
4.4.3.5 HYDRAULIC GEOMETRY OF GRAVEL RIVERS
(1)
Kellerhall’s expressions (1967). Kellerhall’s empirical
expressions for quasi-equilibrium gravel rivers are as follows:
(4.80)
S
–
where U is the average velocity in ft s–1, d is the average depth in
–
ft; d = A/b, A is the discharge area in ft2, b is the width of water
surface in ft, R is the hydraulic radius in ft, S is the canal slope, f is
the Lacey silting coefficient, and Nd is the absolute roughness.
where U is the average velocity in the cross-section, d is the
average depth of flow, and Uos is the velocity of suspended sediment. Dou (1964) obtained the following solutions.
guos CQm
1/ 4 1/ 2 1/ 2
Na
3
(4.70)
d
(4.79)
(4.88)
(4.76)
d* = 0.253(Q*)0.415
(4.77)
S = 0.223(Q*)–0.41
(4.78)
where b * is the non-dimensional width of water surface,
b* = b/d50; b is the width of water surface, d50 is the median size
of bed sediment, Q* is the non-dimensional bankfull discharge,
Fs = 0.1 for light cohesive side slope; Fs = 0.2 for
medium cohesive side slope; Fs = 0.3 for high cohesive side slope.
From Equations (4.85), (4.86) and (4.87):
b=(
F bQ 1 / 2
)
Fs
d =(
FsQ
Fb2
)1 / 3
(4.89)
(4.90)
CHAPTER 4 — FLUVIAL PROCESSES
S = ( Fb )
5
1
1
6 ( F ) 2 γ 12
s

1 
c
/3.63(1 +
) × gQ 6 
2330


(4.91)
If discharge, sediment concentration, grain size and the
viscosity of the slope material are given, the size of a stable canal
can be determined. This method is suitable for a sandy canal with
the side slope composed of cohesive material.
(3) Simons and Albertson’s expressions.
Formula of quasi-equilibrium width:
P = K1Q0.5
(4.92)
–
b = 0.9p = 0.9k1Q0.5
(4.93)
–
b = 0.92b – 2.0
(4.94)
69
4.5
FLUVIAL PROCESSES OF MEANDERING
RIVERS
4.5.1
Plane morphology of meandering rivers
Meandering rivers consist of a series of bends of alternate curvatures connected by straight crossing reaches. The terms used to
describe stable meanders are defined in Figure 4.8.
Essential elements of meandering rivers include:
meandering wave length (L m ); meandering belt width (T m )
(Hm); curvature radius (R); width of straight reach (crossing)
(B); length of curve line (s); Central angle (θ); and length of
crossing (L).
Formula of canal depth:
R = K2Q0.36
Tm or Hm: meandering belt width
R: curvature radius
B: width of straight reach
S: length of curve line
Q: central angle
(4.95)
d = 1.21R (R < 7ft)
(4.96)
Figure 4.8 — Morphological elements of meandering rivers.
d – 2 + 0.93R
(R≥7ft)
(4.97)
4.5.2
Formula of flow resistance:
U = K3 (R2S)m
U
Lm = kQm
2
gdS
(4.98)
Relationships between meander wavelength and
discharge
Based on the data from natural rivers and small rivers in laboratories,
the wavelength and discharge have the following relationship:
= k4 (
Ub
γ
)
0.37
(4.99)
–
where P is the wetted perimeter, b is the average width, d is the
canal depth, K1 is the coefficient related to canal types, R is the
hydraulic radius, K2 is the coefficient related to canal types, m is
an exponential, K3 and K4 are the coefficients, and S is the longitudinal slope. All units are in the English system.
The above morphological formulae were estimated
based on the data for a sandy canal with medium and fine bed
sediment and for a cohesive canal with the bed sediment finer than
sandy, coarse sand gravel canals. For a type 4 canal, the medium
size of bed sediment is 20–82 mm.
Simons and Albertson divided canals into 5 types: (a)
sandy bed and side sandy slope; (b) sandy bed and cohesive side
slope; (c) cohesive bed and cohesive side slope; (d) coarse particles without viscosity; (e) the same as (b), but with a high
sediment transport and a sediment concentration of 2 000–
8 000 ppm. The coefficients are listed in Table 4.15.
Table 4.15
Coefficients for various types of canals
Coefficient
(4.100)
where Lm is the meander wave length, Q is the discharge coefficient, and k and the exponential m vary according to the results of
the different authors listed in Table 4.16.
Table 4.16
Coefficient k and exponent m
Author
Source of data
Chien
(1965)
Rivers in India, the
United States, China
and from model
Dury
(1964)
Sinuous valleys of
some rivers in
the world
K
m
50
0.5
Bankfull discharge
54.3
0.5
Long-term
average maximum
discharge
156
0.46
Annual average
discharge
Carlson
(1965)
Q
For some rivers in the United States (Chien, et al., 1987):
Lm = 0.935Qm0.8M–0.74
(4.101)
where Qm is the annual average discharge in m3 s–1, and M is the
content of silt-clay in the bed and bank materials.
Type of canal
K1
3.5
2.6
2.2
1.75
1.7
K2
0.52
0.44
0.37
0.23
0.34
K3
13.9
16.0
–
17.9
16.0
K4
0.33
0.54
0.87
–
–
m
0.33
0.33
–
0.29
0.29
4.5.3
Relationships between central angle and curvature
radius
(1) Lacey’s formula (Sedimentation Committee, 1992)
R=
Q
0.5
ϕ
(4.102)
where R is the curvature radius, ϕ is the central angle in radians,
and Q is the discharge.
70
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
Table 4.18
Coefficient of KL KT and KI
(2) Formula for the Lower Yellow River (YRCC, 1985)
For the reach upstream of Gaocun:
For the reach downstream of Gaocun:
R=
R=
4500
ϕ
(4.103)
185
3220
(4.104)
185
ϕ
(3) Formula for the Middle and Lower Yangtze River
(YRWRC, 1959):
−0.73
R = 330
Qmax
ϕ
(4.105)
1.15
–
where Q max is the long-term average maximum discharge in m3 s–1,
R is the curvature radius in m, and ϕ is the central angle in radians.
(4) Ouyang’s formula (1983):
R = 48.1 (QS1/2)0.83
(4.106)
where Q is the bankfull discharge, S is the slope for the bankfull
discharge, and R is the curvature radius in m.
(5) Chien’s formula (1987):
R = kQn0.5S–0.25ϕ–1.3
(4.107)
where Qn is the bankfull discharge in m3 s–1, S is the slope in
1/10 000, and ϕ is the central angle in radian. For the Yellow and
Yongding Rivers, K = 10, for the Jingjiang River (part of the
Middle Yangtze River) and Nanyunhe River, K = 3. The above
formula was confirmed by Velikanov based on the data from plain
rivers in the former USSR (1958).
4.5.4
Relationships between meander elements and width
of straight (crossing) reaches
(1) Relationships between curvature radius and width of
straight reach (crossing):
R = KRB
(4.108)
where R is the curvature radius, and B is the width of the crossing
reach. The coefficients of KR are listed in Table 4.17.
(2) Relationship between wave length and the width of
straight reach:
€
Lm = KLB
(4.109)
(3) Relationship between meander belt width and the
width of straight reaches:
Tm = KTB
(4.110)
Table 4.17
Coefficients KR
Author
River
KR
Chien, et al.,
1987
Rivers in China, the United States
and France, and in laboratories
YRCC, 1985
Lower Yellow River
YRCC, 1987
Yangtze River
3
2–6
3.5 < KR < 5–10
Sources of data
KL
KT
KI
Rivers in China, the United States and
France, and in laboratories (Chien, 1987)
12
4.3
9–15
Carlston, 1965
5.8
1–3
(4) Relationship between the length and the width of
straight reach:
I = KIB
(4.111)
Coefficients of KL, KT and KI are listed in Table 4.18.
4.5.5
Relationships between configurations and crosssectional geometry of meanders
Chitale’s empirical expression (1970). Based on the data from 42
rivers, Chitale obtained:
s = 0.917( B ) −0.065 ( D ) −0.077 S −0.052
h
Hm
(4.112)
H m / B = 36.3( B ) −0.471( D ) −0.050 S −0.453
h
h
(4.113)
where D is the average size of bed sediment, Hm is the meander bed
width, B and h are the width and depth of flow, respectively, s is the
length of the curve line, and S is the slope in 1/10 000 (all units of
length are in m).
4.5.6
Crossings
Crossing sections are located between bends of reverse curvature.
In alluvial rivers, crossing sections are approximately rectangular,
in contrast to triangular sections in bends. The water surface slope
through crossings is usually flat at high stages, resulting in deposition in the crossings. At low stages, the water surface slope over
crossings becomes relatively steep. For the relatively stable crossing of the Arkansas River prior to canalization, the maximum
depth in crossings was a function of channel width (Peterson,
1986).
4.5.7
Dynamic line of flow
The transversal distribution of velocity in the cross-section of a
bend is not uniform and there is always a maximum velocity
along the water surface of the cross-section. The dynamic line of
flow is a line along the river that connects the locations on the
water surface where the vertical average velocities are the
maximum. It is also called the main current line. The line
becomes sinuous and flows along the concave sides of bends in
low waters, and passes straight through the centre part of the
water surface in high waters.
(1) Chang’s expression (1983). Based on the measured
data from the Jingjiang Reach of the Yangtze River, the relationship between curvature radius of the flow dynamic line (R) and the
curvature radius of bend (R0) was expressed as follows:
0.73
R = 0.26 R0
(
2
1
b 0.72
3 2 0.23
)
(
Qd
S
)
d
(4.114)
where √ b / d is the average cross-sectional geometry, Q is the
discharge, and S is the slope in 1/10 000.
(2) YRWRC expression (1971). Based on the data from
the Yangtze River, the following expression was obtained:
CHAPTER 4 — FLUVIAL PROCESSES
R = 0.053R0 (
Q 0.35
)
gA
(4.115)
where R is the curvature radius of flow dynamic line in m, R0 is
the curvature radius of channel bend in m, Q is the discharge in
m 3 s –1 ; g is the gravity acceleration in m s –2 , and A is the
discharge cross-section in m2.
(3) Chang’s theoretical expression (1982):
R=
3
1
( R0
ϕSg
Q 2
)
A
(4.116)
where R is the curvature radius of flow dynamic line in m, R0 is
the curvature radius of channel bend in m; Q is the discharge in
m3 s–1, g is the acceleration of gravity in m s–2, A is the discharge
cross-sectional area in m 2 , φ is the central angle of bend in
radians, and S is the slope of flow dynamic line in 1/10 000.
71
longitudinal slope at the convex side is larger than that at the
concave side. The opposite occurs in the downstream reach of the
top (Chien, et al., 1987).
4.5.10 Transversal circulating flows
Under the action of the transversal slope, i.e., the difference in
surface elevations between concave and convex sides, a circulating
(spiral) flow will form with the surface flow towards the concave
bank and the bottom flow towards the convex bank. The structure
of the circulating flow is complicated in natural rivers. In addition
to the main circulating flow caused by the transversal slope, subcirculating flows also occur under the local action of meanders
(Figure 4.9) (Zhang, 1980).
4.5.8
Transversal slope of water surface
When water flows through the channel bend, the elevation of the
water surface on the concave side is always higher than that on the
convex side. The difference between surface elevations on both
sides and the transversal slope can be expressed as follows (Chien,
1987).
∆h =
αV
g
SZ =
∆h
b
=
2
b
(4.117)
R
αV
2
(4.118)
gR
where ∆h is the difference between water elevations on concave
and convex sides in m, b is the width of water surface at the crosssection in m, V is the average velocity in the cross-section in
m s–1, R is the curvature radius of dynamic line of flow in m, and
g is the acceleration velocity of gravity in m s–2.
4.5.9
Longitudinal slope of water surface
Under the influence of transversal slope of the water surface, the
transversal distribution of longitudinal slope of water surface is
not uniform. The maximum slope appears where the circulating
flow is developed. As shown in Table 4.19, the maximum longitudinal slope of water surface occurs at the top of a bend. The
longitudinal slope upstream of the top is smaller than that downstream of the bend. In the reach upstream of the top, the
Table 4.19
Longitudinal slope of water surface at Laijiapu of the
Jingjiang River
Site of cross-section
Inlet of bend
From inlet to top
of bend
Top of bend
Figure 4.9 — Transversal velocities and circulating flows in the
Laijiapu Reach of the Yangtze River.
4.5.10.1 DISTRIBUTION FOR TRANSVERSAL VELOCITY (RADIAL) OF
CIRCULATING FLOWS (ROZOVSKI, 1957, 1965)
g
[Fi (η) − F2 (η)]
VZ = dU
kC
k 2R
For a smooth bed surface:
For a rough bed surface: Vz =
dU
2
k r
{F1(η) −
(4.119)
1/ 2
g
[F (η) + 0.8(1 + ln η]} (4.120)
kc 2
The relationship between F1(η), F2(η) and η is shown in
Figure 4.10.
when k = 0.5, C ≥ 50 (Xie, 1987),
VZ = 6U
d
( 2 η − 1)
(4.121)
R
where VZ is the transversal velocity at the position with a distance
Z above channel bed, K is the Karman constant (in smooth and
regular-shaped bends, K = 0.5, and for natural rivers,
K = 0.3–0.55), U is the vertical-average value of longitudinal
velocities along depth in m s–1, ∆ is the water depth at a vertical
Water surface slope Water surface slope
at concave side
at convex side
(1/10 000)
(1/10 000)
–0.007
0.424
0.019
0.079
0.0849
2.40
From top to outlet
of bend
0.530
0.21
Outlet of bend
0.700
€
0.797
Figure 4.10 — Relationship between F1 (η), F2 (η) and η.
72
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
line, R is the curvature radius of the position at the vertical line,
η = Z/∆, and C is the Chezy coefficient.
4.5.10.2 RELATIVE INTENSITY OF CIRCULATING FLOWS (XIE, 1987)
The transversal velocity VZ in Equations 4.119, 4.120 and 4.121 at a
point above the channel bed with a distance of Z can be considered
as the intensity of circulating flow at that point. The ratio of Vz to
the corresponding average longitudinal velocity along the vertical
U, Vz/U, is called the relative intensity of the circulating flow, from
Equation 4.121:
Vz
=6
U
d
( 2 η − 1)
Vz
U
d
Vx
1+
St = sin β ≈
= ( −10.1 – 5.23)
Vx
r
Fd
d
(4.125)
R
( 8 τ *c )
1/ 2
(1 + f )
(1 − λ) (1 + 2 f
1/ 2
1/ 2
(4.131)
)
where St is the longitudinal slope of the channel bed surface, and
Fd is the density Froude Number.
(4.124)
when η = 0.01 near the channel bed and η = 0.99 near the water
surface, then:
Vz
D
U
Fd =
g
(1 + ln η)
kc
(4.130)
where τ*c is critical shear stress.
Putting Equations 4.127, 4.128 and 4.130 into 4.126, the
longitudinal slope of the bed surface can be expressed as follows:
6( 2η − 1)
=
(4.129)
τ 1/ 2
ρ −ρ
1/ 2
U*c = ( c )
=( s
gdτ *c )
ρ
ρ
(4.123)
R
4.5.10.3 VORTEX INTENSITY OF CIRCULATING FLOW (XIE, 1987)
The ratio of the transversal velocity Vz to the longitudinal velocity
Vx at a certain point is called the vortex intensity, i.e.:
Vz
τ*c = τc /(ρs – ρ) gd,
R
= ( −0.588 – 5.88 )
(4.128)
U*c
where d = d50, the median size of bed sediment, U* (r) is the shear
velocity at position r, U* (r) = U/ (f/8)1/2, and U*c is the critical
shear velocity of moving bed sediment particles.
According to the Shields shear stress:
(4.122)
η = 0.01 near the channel bed, and η = 0.99 near the water
surface. Hence:
U* ( r )
Zb = d
(
ρs − ρ
ρ
gd )
(4.132)
1/ 2
Equation 4.133 was proved by flume experimental data.
If U and the transversal changes of sediment particles are given,
the transversal slope of the bed surface can be obtained by integral
of Equation 4.133. The average velocity for a vertical line is:
U ( r ) = [8 Sc
4.5.10.4 TRANSVERSAL SLOPE OF BED SURFACE AND DISTRIBUTION
rc gD ( r )
r
1/ 2
(4.133)
]
f
OF SEDIMENT PARTICLES
Under the action of circulating flow and channel bed, the transversal transport of sediment particles occurs and the transversal slope
of the channel bed surface is thereby formed. Because of the
complicated exchanges of sediment between the transported particles and bed sediment, the distribution of bed sediment also
becomes non-uniform. Coarse particles appear near the thalweg
line. Chang (1988) introduced some advanced results on the transversal slope of bed surface.
(1) The Falcon-Ascanio-Kennedy expression (1983).
Based on an equilibrium of the radial component of flow acting
force and the component of float weight of sediment particles on
transversal slope, the following expression is obtained:
τor = Zb (1 – λ) (ρ – ρ) g sin β
1+m
( 2 + m)m
ρ
D
r
U
1
D
1/ 2
−
1
1/ 2
=(
Dc
r
1+ f
2f
1
1/ 2
1/ 2
2
(4.127)
where D is the depth of flow, r is the radius of curvature, U is the vertical average of longitudinal velocity, and m is the parameter, m = 1/f1/2.
−
1/ 2
1
1/ 2
)
( 8 τ *c )
fg
ρs − ρ
ρ
1/ 2
1−λ
rc
8 Sc τ c g
[
(4.126)
where τor is the radial component of boundary shear stress, Zb is
the thickness of the bed surface layer, λ is air voids of bed surface
layer, ρs, ρ are specific weights of sediment and water, β is the
transversal slope (dip angle) of the channel bed surface, and g is
acceleration of gravity.
τ or =
where U(r) is the average velocity for a vertical line corresponding
to the radius of y, Sc is the longitudinal slope at the central line
with the radius of rc; f is friction in the Darcy-Weisbach formula,
and D(r) is the depth corresponding to the radius of r.
Putting Equation 4.133 into Equation 4.131 and integrating, then:
1/ 2
]
(4.134)
d
It is a slight protruding line.
(2) The Englund-Bridge expression. Considering the
equilibrium of acting force caused by spiral currents, bottom
currents, gravity and friction on sediment particles on the transversal slope of a bed surface and in the longitudinal direction,
Englund (1974) obtained:
tan δ =
tan β
tan θ
(4.135)
where β is the transversal slope of bed surface, θ is the angle of
repose, tan θ is the coefficient of dynamic friction, and δ is the
CHAPTER 4 — FLUVIAL PROCESSES
intersection angle between the direction of the bottom current and
the longitudinal flow direction.
Bridge (1977), based on the equilibrium of transversal
tractive force and the component of gravitational force acting on a
particle on the transversal slope of the bed surface, further
obtained the following expression:
tan β =
d=
3τ 0 tan δ
)
1/ 6
d
U
=(
τc
(4.138)
r
3 ρDSc τ c
(4.139)
2 ( ρs − ρ ) r tan φ
D
)
5/ 3
rc
(4.140)
r
(3) The Odgaard expression (1981, 1982, 1984).
Odgaard’s method is a revision of the Falcon-Ascanio-Kennedy
expression:
2
1 + m'
1/ 2
2 r [( S − 1) gdcr ]
(4.141)
m' ( 2 + m' )
where α = a/V, a is the projective area of a spheroid after standardization, V is the volume of the spheroid, S = ρs/ρ specific
weight of the sediment; dcr is the diameter of sediment particles in
the critical state of moving, and m' is the reciprocal of the velocity
exponential of grain roughness.
From the Shields critical shear stress:
m' = K
)
1/ 2
(4.144)
r
(
rc
)
3/ 2
(4.145)
=(
D
)
7 / 18
(
rc
Dc
)
1/ 4
(4.146)
r
In application of Odgaard’s expression, firstly m' should
be calculated by Equation 4.142, and then the transversal slope of
the bed surface can be calculated by Equation 4.141, if average
depth velocity and grain size on the slope are given.
4.5.11 Sediment transport in meandering rivers
4.5.11.1 TRANSPORT OF SUSPENDED LOAD
In general, the distribution of suspended load is not uniform along
the depth. The sediment concentration is higher and the grain size
is coarser near the channel bed. In a bend reach, because of the
influences of spiral flow, water with high concentrations and
coarse particles is concentrated along the convex bank, and that
with low sediment concentrations and fine sediment particles is in
the concave bank. The distribution of sediment concentration
through the depth near the concave side is also more uniform. In a
straight (crossing) reach, the distribution of sediment concentration along depth is uniform, and the transversal distribution of
vertical average concentration corresponds to the transversal
distribution of vertical average velocity.
The transversal sediment discharge caused by circulating
(spiral) flow can be shown in the expression by Xie (1987):
gsn = gs
d
6
1 − ηa
R
J1
Jn
(4.147)
where gsn is the transversal sediment discharge per unit width; d is
the flow depth; R is the radius of the curvature, ηa = a/h, a is the
thickness of the bed surface layer, and g—
s is the average longitudinal sediment discharge per unit width.
J1 =
J1 =
∫
1
ηa
∫
z
1
ηa
(
1−η
) dη,
η
( 2η – 1)(
z
1−η
) dη,
η
ηa = 0.001
ηa = 0.01
(4.148)
(4.149)
z is the exponential in the sediment concentration distribution.
U
(4.142)
1/ 2
[( S − 1) gdcr τ *c ]
U
From Equation 4.133
rc
r
where S is the longitudinal slope corresponding to
U
)(
Dc
Dc
Uc
radius r.
3d D
D
Odgaard supposed that the transversal bed surface was in
a straight line and that the Shields critical shear stress τ*c was
proportional to –2/3 power of d:
(4.137)
If the water depth D is known, the corresponding grain
size of the sediment, d, on the transversal slope of bed surface can
be obtained. Here, d is the grain size of sediment particles at any
point on the transversal slope of the bed surface, r is the radius,
and Sc is the longitudinal slope corresponding to the radius rc on
the central line.
sin β =
(
Putting Equation 4.145 into Equation 4.144
2 ( ρs − ρ ) g tan φ
S = Sc
dr
d'
Uc
3τ 0
τ 0 = ρgDSc
=(
d1
where τ0 is the shear stress on the longitudinal bed surface, and d
is the grain size of sediment particles on the transversal slope of
the bed surface.
For fully developed flow:
d=
U
(4.136)
2 dg ( ρs − ρ )
73
Uc
=
m'
(
D
mc ' Dc
)(
rc
r
)
(4.143)
From Strickler’s formula, the Manning’s roughness coefficient n is proportional to 1/6 power of grain size d, in m;
N = d1/6/21.1.
4.5.11.2 BED LOAD TRANSPORTATION
The transport of bed load in meandering rivers is characterized by
the following two phenomena (Xie, 1987):
(1) According to experimental data, the sediment
particles eroded from the concave bank of a bend are carried by
flow and partly deposited at the crossing and convex bank of the
next bend. The remaining particles are further carried and deposited
at the downstream crossings and convex banks of downstream
74
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
bends. However, when the circulating flow is strong, the sediment
particles eroded from the concave bank are carried directly to the
opposite convex bank and settle there. The former is called sameside transporting of sediment and the latter, different-side
transporting. For meandering rivers, same-side transporting of
sediment is more common than different-side transporting.
(2) Bedload particles often move in a transporting belt
along the river instead of spreading all over the channel bed. The
transporting belt is situated near the point bars of convex banks.
The transversal transport of bed load caused by spiral
flow is controlled by the transversal slope of the bed surface.
Ikeda (1982) conducted a wind tunnel experiment with sand particles of 0.26 and 0.42 mm and estimated:
q* '
tan β
0.0085[
τ*
(
τ*
τ *c τ *c
0.5
− 1)]
(4.150)
where q*' is the dimensionless transversal bed load discharge per
unit width, τ* is the dimensionless shear stress or Shields shear
stress, and τ*c is the critical Shields stress.
q* ' =
qb '
3 1/ 2
[( S − 1) gd ]
(4.151)
where qb' is the transversal bed load discharge, and S = ρs/ρ, ρs, ρ
is the specific weight of sand and flow.
Parker (1984) considered the effects of the transversal
slope of the bed surface and the spiral flow, and estimated:
qb '
qb
=
q* '
q*
= tan δ −
1 + ( C L / C D ) tan φ τ *c 1 / 2
(
)
tan β (4.152)
τ*
tan φ
where qb, q* are the longitudinal bed load discharge; CL is the
coefficient of lifting force; CD is the coefficient of tractive force; δ
is the angle between bottom velocity and longitudinal velocity,
and Φ is the angle of repose.
4.5.12 Characteristics of fluvial processes
4.5.12.1 COLLAPSE OF CONCAVE BANKS AND GROWTH OF
CONVEX BANKS
Generally, meandering rivers are in the equilibrium state of sediment transport. Under the action of spiral flow, the sediment
deposited at convex banks is mainly from erosion of the concave
side. As a result, the channel has a continuous migration over the
years. Figure 4.11 shows the transversal migration of the crosssection at the apex of the Laijiapu Bend in the Yangtze River. The
river channel migrated rightward a distance of one km in ten
years. Some examples of the rate of bank collapse for rivers
throughout the world are listed in Table 4.20 (Chien, et al., 1987).
4.5.12.2 MIGRATION OF MEANDERINGS
The shear stresses acting on the bank and channel bed reach a
maximum at the position downstream of the apex of the bend, and
the eroded sediment particles deposit at the convex bank, causing
the point bar to develop. With the collapse of the concave bank
and the growth of the convex bank, the channel bend, as a whole,
gradually migrates downstream. During the migration process, the
outside of the bend is changed, but the centre part of the crossing
may remain basically unchanged. Therefore, the adjacent bends
move around a fixed point, and an S-shaped meander may form
(Figure 4.12).
Figure 4.11 — Migration of the apex of the cross-section at the
Laijiapu Bend.
(a) Accumulative erosion at concave bank and deposition at convex
bank
(b) Changes of cross-sections at top of bend
1. Accumulative erosion at concave bank
2. Accumulative deposition at convex bank
3. Accumulative difference of deposition and erosion
4.5.12.3 CUTOFFS
As the S-shaped bend develops, the apex of the two adjacent
bends located on the same side come closer, and the difference of
water surface at both ends of the neck becomes larger. Once the
overbank flood occurs, the neck may be scoured. A new channel
may be formed, widened and deepened, and the old bendway may
become separated from the river by deposition, surviving as an
oxbow lake. This phenomenon is called natural cutoff.
Subsequently, the channel upstream of the cutoff is eroded
because of the steep slope, and the channel downstream of the
cutoff is deposited because of the lower slope. If the pilot channel
is not protected from erosion, a new meander (bend) is formed
again (Figure 4.13).
4.6
FLUVIAL PROCESSES OF WANDERING RIVERS
The Lower Yellow River is a notorious wandering river. Its special
fluvial processes are used to describe the outstanding features of
wandering rivers.
Figure 4.12 — Changes in S-shaped bends.
CHAPTER 4 — FLUVIAL PROCESSES
75
broken and disorganized channel beds. For example, in the
Huayuankou Reach of the Lower Yellow River, which is a typical
wandering reach, the channel slope is 0.0002–0.00025, the water
depth is only 1 to 3 m, and the velocity is higher than 3 m s–1.
Special water surface phenomena, corresponding to bed forms
such as dunes and anti-dunes etc., often occur because the Froude
numbers of its flow are far greater than those in ordinary alluvial
rivers. The Brahmaputra River in Bangladesh is a wandering
branched river. Although its channel slope is smoother than that of
the Lower Yellow River, its velocity is also high because of its
large volume of discharge. Similar flow surface phenomena also
occur in that river (Zhou, 1998, 1995).
4.6.1.2 CHARACTERISTICS OF SEDIMENT TRANSPORT
In China, all the wandering rivers carry huge amounts of sediment load. For example, the long-term average sediment
concentration is 27.3 kg m –3 at Huayuankou Station on the
Lower Yellow River, and 44.2 kg m–3 at Sanjiadian Station on
the Yongding River. Sediment concentration and sediment
discharge vary substantially at the same flow discharge. The
Figure 4.13 — Changes in the Nianziwan Bend on the Yangtze River
after cutoff.
4.6.1
Flow and sediment transport
4.6.1.1 CHARACTERISTICS OF RIVER FLOW
Wandering rivers have steep slopes, small water depths and high
flow velocities in wide and shallow channels with fragmented,
Table 4.20
Rate of river bank collapse (Chien, et al., 1987)
Country
River
Yangtze
River
Area of
watershed
(km2)
Width of
river (m)
Annual
discharge
(m3 s–1)
Jingjiang Reach
Jiujiang estuary
China
Yellow
River
United Kingdom
Railway bridge
Tongbadou
Tongbadou-Gaocun
Gaocun Sunkou
Rheidol River
Endrick River
Tyfi River
United States
179
98
633
25
6.9
6 042
66.2
1020–1400
91.5
16
Canada
Pembina River
Beatton River
16 000
64
370
Australia
Torrens River
78
5–10
Poland
Wisloka River
Former USSR
Comprehensive
statistics
Obi River
Klaralven River
Czechoslovakia
Hernad River
Bangladesh
Brahmaputra River
Remarks
max. 88.4
av. 30.0
Max. 200
Min. 2.5
Av. 48.7
470
1949–1967
Meandering
Branched
Wandering
Transition
1.75
0.5
2.65
1951–1971
1986–1957
1905–1971
23
14.9–40.5
0.36
0.67
1/3 width of
flood plain
1.7–7.0
1722–1971
1963–1970
1807–1958
1937–1968
1879–1954
6.6
More than
100 years
1880–1970
3.35
0.48
1910–1956
1250 years
0.58
1960–1963
22.5
8–11
1970–1972
1897–1958
1 434
max. 100
av. 10–15
0–15
0.23
0.32
1.6
1800–1850
1850–1950
1950–1956
Des Moines River
Sweden
Date of
survey
409
178
Mississippi River
Ohio River
White River
Downstream of
Missouri River
Little Missouri River
Rate of bank
collapse
(m/a)
19.2
225
650
Meandering
Meandering
1897–1958
5 420–
11 820
120
5 400
50–60
10–30
5–10
1937–1972
934 990
6 000–
13 000
1 898
6–275
1952–1963
Meandering
76
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
sediment-carrying capacity for bed material load is determined
by both the flow intensity and incoming sediment concentration. Meanwhile, when the sediment concentration of incoming
runoff is rather high, the sediment-carrying capacity for bed
material load is also high along the river. The more incoming
sediment there is, the more sediment is sluiced. If the incoming
sediment concentration is taken as a parameter, the relationships between sediment discharge, incoming sediment
concentration and flow discharge at the stations on the Lower
Yellow River may be expressed as follows (Sedimentation
Committee, 1992).
α β
Qs = kQ S0
(4.153)
where Qs is the sediment discharge for bed material load in t s–1,
Q is the flow discharge in m3 s–1, S0 is the sediment concentration
for bed material load at the upper neighbour station in kg m–3, K
is the coefficient of sediment transport, and α and β are the exponents. For the stations on the Lower Yellow River, α = 1.1–1.3,
β = 0.7–0.9, and K is determined by degradation or aggradation at
an earlier stage.
4.6.2
Morphological features
4.6.2.1 STATIC FEATURES
The static features of wandering rivers include the following:
(1) There are dense mid-bars, branches and scattered flows in
the channel. (2) The channel configurations are more smooth
and straight, with a sinuous coefficient (total length of
branches)/(length of channel) of 1–1.3, which is smaller than
that of meandering rivers (1.5–2.5). (3) The channel beds are
wide and shallow. The maximum width of the wandering reach
of the Lower Yellow River is more than 10 km, and b d–1 is
20–40, which is 10 times as large as that of the meandering
reach of the Yangtze River.
4.6.2.2 DYNAMIC FEATURES
The dynamic behaviour of wandering rivers may be described as
follows: (1) The mid-bars move quickly and the river bed can be
easily eroded and deposited. (2) The positions of main currents
change constantly. Sometimes, the position of the main current
can change completely during a flood. (3) The range of main
current shifting is large and the shifting rate is high. For example,
the main current has migrated 6 km in 24 hours in the Lower
Yellow River. (4) There are two types of migration of the main
channel — gradual shifting and sudden shifting in channel evolution. Gradual shifting often occurs in flood-rising periods, and
sudden shifting occurs in flood-falling periods.
4.6.2.3 NODE POINTS
In the wandering reach of the Lower Yellow River, the river
configuration is chequered longitudinally with wide and narrow
channels. The wide channel contains dispersed flows, dense midbars, disordered branches, a scattered platform, and a strong
shifting of the main current. The narrow channel contains relatively concentrated flows, less sand bars, and a weak migration of
the main channel. The narrow channel is called the node point
(Chien and Zhou, 1965).
The node points play an important role in controlling
the wandering of the main current and the changes of
configuration.
There are two types of node points. The first is called the
grade 1 node point, which has a fixed position and two support
bases on both sides of the channel, and plays a role in controlling
the configuration above medium water level. The second is called
the grade 2 node point, which has an unfixed position and one
support base on one side of the channel. It can control the configuration below the median water level.
The conditions for forming the grade 1 node point on the
Lower Yellow River are: (1) Man-made controlling works on both
sides of the river channel (Figure 4.14 (b)); (2) Cliff or vulnerable
spots on one side of the channel, and a clay boundary on another
side (Figure 4.14 (a), (c), (d) and (e)).
The grade 2 node points are shown in Figure 4.15. Their
support bases are often embankments or high banks on one side of
the channel, and the other side is a low bank or side bar. The position of grade 2 node points may migrate along the reach when the
discharge changes.
(a) Convex cliff of Mongshan
(b) Man-made structure control
(c) Vulnerable spot and unerodible bank (d) Vulnerable spot revetment (e) Convex vulnerable spot
Figure 4.14 — Grade 1 node point on the Lower Yellow River.
(a) 26 March 1959, Q = 774 m3 s–1
(b) 3 April 1959, Q = 2870 m3 s–1
(c) 15 August 1959, Q = 3800 m3 s–1
Figure 4.15 —Grade 2 node point on the Lower Yellow River.
CHAPTER 4 — FLUVIAL PROCESSES
Grade 1 node points have the following features in plain
morphology (Chien, et al., 1965):
B2 = 3.82B1– 1.45
(4.154)
B2 = 0.34L – 0.31
(4.155)
where B2 is the shifting range of the main channel in a wide reach,
B1 is the shifting range of the main channel in a narrow reach, and
L is the length of the wide reach.
77
where ∆G is the intensity of channel degradation and aggradation
during a flood in t day–1, “–” marks degradation, and “+” marks
aggradation, S/Q is the coefficient of incoming sediment load in
kg.s m–6, and S and Q are the average sediment concentrations in
kg m–3 and average discharge in m3 s–1 during the flood, respectively. In the Lower Yellow River, if S/Q ≥ 0.015, both the main
channel and the flood plains are in deposition, and if S/Q < 0.015,
the main channel suffers from erosion and the flood plains are in
deposition.
4.6.3.2
DEGRADATION AND AGGRADATION FOR
WANDERING RIVERS WITH RELATIVE LOW
4.6.3
4.6.3.1
Channel degradation and aggradation
CHARACTERISTICS OF DEGRADATION AND AGGRADATION
FOR WANDERING RIVERS WITH HIGH SEDIMENT
CONCENTRATION
The Lower Yellow River is a remarkable example of a river with a
large amount of sediment load. Its features of channel degradation
and aggradation may be described as follows (Chien, et al., 1965,
1987).
(1) The Lower Yellow River is characterized by serious
aggradation with an annual amount of siltation of 0.4 × 109t. The
channel bed has risen by 7 to 10 cm/yr in past years and long-term
accumulation has resulted in the river becoming a suspended river
having flood plains 3 to 5 m above the ground outside the
embankments. Ninety per cent of the siltation is in the wandering
reach.
(2) The channel aggradation of the Lower Yellow River
can be classified into two types, namely streamwise deposition
and retrogressive deposition. The main cause of streamwise deposition is the huge amount of incoming sediment load from the
watershed and the insufficient sediment carrying capacity of the
flow. The deposition develops from upstream to downstream
reaches resulting in a decrease in the sediment concentration and
grain sizes of suspended load along the river. Retrogressive deposition is caused by the raised local datum of an estuary, caused by
estuarine deposition. The range of retrogressive deposition is 200
to 300 km from the river mouth of the Lower Yellow River, while
all the deposition occurring in the wandering reach is streamwise
deposition (Zhou, 1982).
(3) Deposition occurs mainly in flood seasons, which
account for 70 per cent of annual deposition. During the flood
season, about 90 per cent of deposition is caused by floods. Most
deposits are silted on the flood plains, and the main channel is in a
state of erosion during the floods. According to measured data
from six overbank floods in the period from 1950 to 1960, the
total amount of deposition, including the deposition on flood
plains and the erosion in the main channel, was 1.65 × 109t.
Under the conditions of medium and low flows, deposition always occurs in the main channel. The depositions on flood
plains during floods and in the main channel in medium and low
flows are restricted, which results in the parallel raising of the
flood plains and the main channel.
(4) During a flood, the channel bed is eroded in the
rising stage and aggraded in the falling stage. The intensity of
erosion and aggradation in floods may be expressed as follows
(Sedimentation Committee, 1992):
2
0.75
∆G = 137Q [ S − 0.33( S )
]
Q
Q
(4.156)
SEDIMENT CONCENTRATION
As mentioned above, the Brahmaputra River in Bangladesh is a
wandering-branched river with a long-term average sediment
concentration of 0.81 kg m–3. The features of degradation and
aggradation for the river can be summarized as follows (Zhou,
1998).
(1) The degradation and aggradation of the river is
mainly caused by the transport of bed load and the coarse particles
of suspended load near the channel bed. The river is nearly in
equilibrium, with an average deposited thickness of 0.01 m in the
past one hundred years.
(2) The main form of channel degradation and aggradation is the growth and decline of the main channel and branches.
No obvious raising of the surface of islands and side bars is found.
There is no retrogressive deposition because the estuary has no
extension.
(3) Erosion and deposition are affected by sudden
events in the upper reaches. For example, following the great
earthquake in the 1950s in Upper Assam, India, the Yalutsangpo
River and the Upper Brahmaputra River caused earth and debris to
slip into the river, which rose 3 m at Dibrugarh in five years. From
1950 to 1957, the channel bed rose by 0.5 to 2.4 m in a reach of
168 km of the river in India. Deposits in the upper reaches have
been carried into the Lower Brahmaputra River in Bangladesh
since the late 1970s, and this has resulted in a gradual aggradation
of the downstream channel.
4.6.4
Degradation and aggradation in hyperconcentrated
floods
4.6.4.1 FEATURES OF HYPERCONCENTRATED FLOODS IN THE
LOWER YELLOW RIVER
The hyperconcentrated floods coming from the Loess Plateau in
the Middle Yellow River basin have a peak discharge of 4 000 to
8 000 m 3 s –1 and a sediment concentration higher than
400 kg m–3, the highest being 911 kg m–3 after regulation by the
Sanmenxia Reservoir. On average, the size distribution of
suspended load of the hyperconcentrated floods is as follows.
Sediment size smaller than 0.025 mm accounts for 50 per cent,
0.025–0.05 mm, 24 per cent and coarser than 0.05 mm, 26 per
cent. The average median size has a relationship with the
maximum sediment concentration in a flood, as expressed below
(Zhou, 1998):
D50 = 0.000027 × S + 0.0139
(4.157)
where D50 is the average median size of suspended load in a flood
in mm, and S is the maximum sediment concentration in a flood in
kg m–3.
78
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
4.6.4.2 FLOW PATTERNS AND TRANSPORT MODES
The flow pattern of hyperconcentrated flow can be classified into
the laminar and turbulent flows. Large amounts of fine suspended
sediment restrain the development of turbulence. When the sediment concentration, especially for the fine sediment, reaches a
certain level, the turbulent flow is easily trensferred into the
laminar. The transport modes of the hyperconcentrated flow can
also be classified into the pseudo-homogeneous and heterogeneous (two phases) flows based on the vertical distribution of the
concentration according to the field data. The effective Reynold’s
numbers of the hyperconcentrated floods in the Lower Yellow
River are higher than the critical Reynold’s number, and the vertical distribution of sediment concentrations is not uniform. All the
hyperconcentrated floods in the Lower Yellow River thus belong
to the turbulent and heterogeneous two-phase flow (Zhou, 1982,
1995, 1998, Zhao, et al., 1998).
4.6.4.3
FEATURES OF DEGRADATION AND AGGRADATION
(1) Because the high viscosity of hyperconcentrated
flow causes a decrease in the falling velocity of sediment particles, the sediment carried by the hyperconcentrated flow may be
transported over a relatively long distance. However, high levels
of aggradation occur simultaneously when the flow passes
through the Lower Yellow River. According to measured data, the
deposition caused the hyperconcentrated floods on average
accounts for 55 per cent of the incoming sediment load and 75
per cent of the total deposition in the Lower Yellow River, of
which 86 per cent is deposited in the wandering reach of the river
(Zhou, 1995).
(2) During aggradation, the sediment for various grain
sizes settles and results in the decrease of their sediment
concentration in the wandering reach. However, coarse sediment
particles settle easily, and their deposition accounts for a larger
portion of the incoming sediment, while the deposition of fine
sediment particles accounts for a lesser portion of the incoming
sediment. Therefore, the suspended sediment carried by hyperconcentrated floods becomes finer and finer along the river
(Zhou, 1998).
(3) The wandering reach of the Lower Yellow River has
a wide and shallow channel. The serious deposition of the hyperconcentrated flood mainly occurs on the low flood plains beside
the main channel, causing the main channel to narrow. On the
other hand, the flood flow is forced to be concentrated in the main
channel, and leads to erosion. Under certain conditions the main
channel can be sustantially cut down to form the so-called high
flood plains and deep main channel. During the hyperconcentrated
floods in 1977, on the upper part of the wandering reach the width
of the main channel decreased from 3 000 m to 400 m, and the
maximum elevation difference between the flood plains and the
main channel reached 6 m at Huayuankou Station due to sharp
cutting. The preconditions for forming such high flood plains and
deep main channels include (Zhou, 1983, 1998) the following: (1)
the flood peak discharge should be over 5 000 to 6 000 m3 s–1; (2)
the sediment concentration should be higher than 400 kg m–3; and
(3) the flood peak discharge and the maximum concentration
should occur nearly at the same time. These three preconditions
should be satisfied simultaneously. Furthermore, even though such
a cross-section is shaped, it is unstable and may easily be eroded
by the wandering flow, and recover the original wide and shallow
cross-section.
4.6.5
Shrinking of river channel
Since the 1980s, under the combined influences of climate
change, increases in water supply in rural and urban areas, and the
completion of large reservoirs in the upper reach, etc., incoming
runoff and sediment load in the Lower Yellow River have
decreased by 34 and 48 per cent, respectively. No flow in the
downstream channel occurs for three to four months in the dry
season every year. Although the total aggradation has decreased in
the Lower Yellow River, 85 per cent of the deposition has accumulated in the main channel. As a comparison, the deposition in the
main channel was 23 per cent in the 1950s. As a result, the width
of the main channel in the wandering reach decreased from
1 000–1 500 m to 800–1 000 m, with the minimum width being
only 600 m. The flood-conveying capacity of the main channel has
also dramatically decreased, which causes the flood control conditions to worsen. If the runoff increases, the question of whether
the channel can be enlarged to the width of the 1950s will remain
a problem.
4.7
FLUVIAL PROCESSES OF ANABRANCHED
RIVERS
The branches are characterized by stable islands. The river
channel is divided by the islands into two or more stable branches.
There are 41 branched reaches with a total length of 817 km on
the Middle and Lower Yangtze River from Chenglingji to
Jiangying, over a stretch of 1 120 km.
Tieban Island
Nanyang Island
(a) Straight
Tianxingzhou branched reach
Moerzhou branched reach
(b) Slightly sinuous
(c) Goose-head
Figure 4.16 — Types of anabranched rivers.
CHAPTER 4 — FLUVIAL PROCESSES
4.7.1
Morphological characteristics of anabranched rivers
4.7.1.1 CLASSIFICATION
According to their shape, anabranched rivers can be classified into
three subtypes (Xie, 1987).
(1) Straight anabranched rivers. Each branch is relatively straight. The sinuous index is 1.0 to 1.2 and the branches are
symmetrical (Figure 4.16 (a)).
(2) Slightly sinuous anabranched rivers. The outlines of
these anabranched rivers are slightly sinuous; but at least one
branch should have a sinuous index of 1.2 to 1.5. Most rivers have
two simple branches, but some have three multi-branches
(Figure 4.16 (b)).
(3) Goose-head-shaped anabranched rivers. At least one
branch has a sinuous index larger than 1.5. Most rivers have two
or more islands to divide the channel into a multi-branched one
with three or more branches (Figure 4.16 (c)).
4.7.1.2 MORPHOLOGICAL INDICES
The plain morphology of anabranched rivers can be expressed as
follows (Sedimentation Committee, 1992):
(1) Coefficient of branches K1
K1 =
Total length of branches
length of central line of channel
Q2 = (1 – m) Q2
3
22
m
Maximum width (including width of islands)
Width in narrow reach upstream of branches
(4.160)
b1 + b2 = b0 [(
6
S0 311 611
S 3
) m
+ ( 0 ) 11 (1 − m ) 11 ]
S1
S2
ηm =
Qm
Qm + Qn
1
Q0 = 1 ζd0 3 S0 2
n
(4.172)
where m and n represent the main subbranch and branch, respectively.
1
ηm =
An
(
dn
)
2/3
(
Lm
)
1/ 2
(
Ln
nm
)
(4.173)
nn
where Am and An are the discharge area of the main subbranch and
branch, respectively, Lm and Ln are the length of the main branch
and subbranch, respectively, and nm and nn are the roughness of
the main subbranch and branch, respectively. The ratio of sediment diversion can be expressed as follows:
ζm =
Qm Sm
1
=
Qm Sm +Qn Sn 1 + Qn Sn
Q S
(4.174)
m m
(4.161)
where S is the average sediment concentration in kg m–3.
If Sm /Sn = Ks,
ξm =
(4.162)
4.7.2
Morphology of cross-sections
If there are two branches, let b0, d0, and S0 be the width, depth and
slope of the single channel before bifurcation, b1, d1, S1 and b2,
d2, S2 represent the width, depth and slope of the two branches,
and suppose that the roughness and morphological relationship
remain unchanged, then (Chien, et al., 1987),
b0
b1
b3
=
=
=ζ
d0
d1
d3
(4.171)
4.7.3
Ratio of discharge and sediment diversion
Taking two branches as an example, the ratio of discharge diversion of the main branch can be expressed as follows (Xie, 1987).
(5) Length-width ratio of island K5
Length of island
K5 =
Maximum width of island
(4.170)
( B1 + B2 ) > b0
(4) Length-width ratio K4
Length of branched reach
K4 =
Maximum width of branched reach
(4.169)
d0
3
S0 3 22
)
(1 − m ) 11 d0
S2
Am d m
(3) Widened ratio K3
(4.168)
3
11
since S1 > S0, S2 > S0, m < 1; hence d1 < d0 and d2 < d0
Putting Equations 4.169 and 4.170 into Equation 4.163,
1+
The K2 of a branched river must be larger than 1.5.
11
S0
)
S1
d2 = (
(4.158)
2 × total length of islands and mid –bars in branched reach (4.159)
Length of central line of channel
K3 =
d1 = (
Thus,
(2) Index of branches K2
K2 =
79
(4.163)
nm
1 − nm
Ks
4.7.4
+ nm
(4.175)
Fluvial processes
4.7.4.1 MAIN FEATURES
The main feature of the fluvial processes of anabranched rivers is
the growth and decline of the main channel and branches. The
main channel might be transformed into the branch, and the
branch also might be transformed into the main channel because
of changes in water and sediment diversion. During transformation, the original main channel is silted and raised, while the
original branch is scoured, and descends.
(4.164)
4.7.4.2
CHANNEL DEFORMATION FOR DIFFERENT ANABRANCHED
RIVERS
11
1
11
1
Q1 = 1 ζd1 3 S1 2
n
Q2 = 1 ζd 2 3 S2 2
n
Let
m = Q1/Q0
(4.165)
(4.166)
(4.167)
(1) The fluvial processes of straight anabranched rivers
are the alternate distribution of pools and side bars and their
parallel shifting downstream. If the flow conditions and the
entrance of the branches are changed, the main channel is transformed into a branch and the branch may be transformed into the
main channel.
80
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
(2) For a slightly sinuous anabranched river, the main
channel is often located on the concave side. When the concave
bank is eroded and shifts to a hard boundary, the main channel
becomes stable, and the branches are also stable. Such a stable
situation might last more than 100 years.
(3) Goose-head-shaped anabranched rivers mostly are
formed from slightly-sinuous anabranched rivers. For example, in
the Luxikou Reach of the Yangtze River, the left branch continuously developed through erosion in the floods of 1926, 1931 and
1933, and a typical goose-head-shaped branch was formed in
1934. An unstable goose-head-shaped and multi-anabranched
reach was finally formed after the island was cut by flood flows
(Figure 4.17).
4.8
(1) Alternate side bars. Alternate side bars cause the
main current line to be sinuous. The size of side bars depends on
the size of river channels (Xie, 1987).
b = 0.57B
(4.176)
L = 2.8B
(4.177)
where b is the width of the side bar, B is the width of the channel,
and L is the length of the side bar. All the parameters are under the
bankfull discharge.
(2) Side bars alternate with pools along the river. The
relationship between the distance of pools and river width can be
expressed as follows (Chien, 1987).
FLUVIAL PROCESSES OF STRAIGHT RIVERS
Lp = 6B
4.8.1
Morphological features
In alluvial rivers, straight rivers have straight outlines with a relatively short length, such as the straight reach between two bends
of meandering rivers or the single straight river between two
branched reaches. Straight rivers have the following main features.
(4.178)
where Lp is the distance between the pools and B is the width of
the river. The expression coincides with that of the meander. This
implies that straight rivers, in essence, have the generality of
meandering rivers.
(3) Riffles and pools occur alternately along the
thalweg. In low water seasons, sprays can be found on the surface
downstream near the riffles.
4.8.2
Features of flow and sediment transport
Straight rivers have pools and crossings with a sinuous main
current. Straight rivers also have circulating flows, but the flow
intensity is weaker than that of meandering rivers. The sediment
transport rate of bed load on crossings is lower than that in pools.
Obvious sorting of sediment particles can be found. The coarse
particles are concentrated on crossings and the sediment composition in pools is fine. Sediment sorting also exists vertically on the
crossings. The coarse sediment is located near the surface, while
fine sediment is situated in the deep layers.
4.8.3
Features of fluvial processes
(1) The migration of alternate side bars downstream and
the corresponding shifting of pools are the major characteristics of
the fluvial processes of straight rivers. Therefore, the river, including the side bars, pools and crossings, as a whole moves some
distance downstream after a certain time period.
(2) The river channel is widened periodically. When
the side bars move down, the erosive banks on both sides are
covered by the side bars. Correspondingly, the formerly covered
banks are exposed and re-eroded by the flow. Thus, the bank
lines recede, causing the channel to be gradually widened.
Then, the wide side bars are cut off by the flow and become
mid-bars or islands. Once one branch is blocked, the island
connects with the bank and the channel becomes narrowed once
again.
Luxikou
Figure 4.17 — Luxikou anabranched reach on the Yangtze River.
4.9
STABILIZATION AND RECTIFICATION OF
RIVER CHANNELS
The training of alluvial rivers can be classified into long distance
regulation and local regulation. Long distance regulation is aimed
at flood control and navigation, while local regulation aims to
prevent banks from collapsing, to stabilize the intake of water
diversion and enforce the channels upstream and downstream of
bridges, etc.
CHAPTER 4 — FLUVIAL PROCESSES
4.9.1
Parameters of river training planning
The main parameters of river training include design discharge
and planning of channel width and channel alignment.
4.9.1.1
DETERMINATION OF DESIGN DISCHARGE
(1) Design discharge of flood channels (Xie, 1987). The
main purposes of training flood channels are to raise the floodcarrying capacity of the channel, prevent important embankments
from collapsing, and guarantee flood control safety. Design flood
discharge is determined by the recurrence intervals of floods. In
China, the recurrence of river training works in the most important
region is 1 to 0.33 per cent, in important regions it is 2 per cent
and in general regions it is 10 to 5 per cent. But most rivers have
recurrence intervals of 5 per cent. The recurrence interval of
design flood for rivers varies from country to country depending
on the economy of the country.
(2) Design discharge of low flow channels. The purposes
of training low flow channels are to ensure the conditions of
navigation and water diversion and to stabilize the location of
diversion intakes. Two methods can be used to determine design
discharge:
(a) The discharge is determined using the water level, which is
in accordance with a guarantee modulus from the long-term
daily water levels. The guarantee modulus of navigation in
China is 90–95 per cent;
(b) The low discharge corresponding to the historical lowest
water level or the long-term average low water level is taken
for the design discharge for the low flow channel.
(3) Design discharge of moderate flow channel. The
floods of alluvial rivers are conveyed mainly by the moderate flow
channel that is moulded by the dominant discharge. If the moderate channel of a river is controlled, the training of its flood channel
and low flow channel can be easily resolved. The determination of
dominant discharge is illustrated in section 4.4.1.
4.9.1.2 DETERMINATION OF CHANNEL WIDTH
The channel width of river training is the surface width of the
straight channel (crossing) corresponding to the bankfull
discharge. Two methods can be used to determine the training
width of the channel:
(a) The morphological relationships in section 4.4.3 can be used
to determine the channel width under the bankfull discharge.
Coefficients and exponents in these relationships should be
determined with field data from the trained river;
(b) Statistical method. Analysing the data from typical rivers
with the same river pattern and a channel width corresponding to the bankfull discharge may be used for the trained
river.
4.9.1.3 ALIGNMENT
In order to minimize damage caused by the stream on stabilization
and rectification structures, the river channel should be shaped in
an alignment consisting of a series of easy bends with the flow
directed from one bend to the next one downstream. In the Lower
Yellow River, the principles of river regulation are mainly for
flood control, but proper consideration is given to the protection of
floodplains, as well as diversion for irrigation and improvements
for navigation. The aim of river regulation is to stabilize the
channel for moderate floods through effective measures, because
moderate floods often threaten vulnerable sections. Therefore, the
81
dominant discharge is adopted as the design flood of the moderate
channel regulation.
The alignment of river training of the Lower Yellow
River is determined by the following relationships according to
field data of the Lower Yellow River (Xu, 1983).
R = 3250/Φ2.2
(4.179)
R = (2 – 4) B
(4.180)
L = (2 – 5) B
(4.181)
where R is the radius of the bend, L is the length of straight
stretch, and B is the channel width in the straight stretch (all in m).
There are two types of alignment in the Lower Yellow River: (a)
alignment with successive bends; and (b) alignment with sharp
bends and long straight stretches (Figure 4.18).
4.9.2
4.9.2.1
Structures of river training works
STRUCTURES OF TRAINING WORKS FOR MODERATE AND
LOW FLOW CHANNELS
Structures of training works on moderate and low flow channels
mainly include long and short groins and revetment.
(1) Groins. Groins extend out from the bank into the
flow. Long or short groins are used to cut off side channels and
chutes, concentrate a braided river into a single channel, concentrate a channel to increase depth, realign a river reach, prevent
bank erosion and protect structures along banks and near bridges
and utility crossings. Groins are aligned either at an angle or
perpendicular to the flow. Experience indicates that groins aligned
either at right angles to the bank or in a slightly downstream direction are more effective than groins angled upstream from the bank
line.
The length of the groin depends on its location (in a
crossing, at a bend, across an old channel, etc.). The length of long
groins on the Lower Yellow River is 100 to 200 m, the longest
being 3 km, and short groins are 10 to 20 m long.
Groin spacing is usually 1.5 to 6 times the groin length,
but 1.5 to 2.0 times the groin length gives the best defined channel
for navigation. Groin spacing is equal to groin length on the
Yellow River. Figure 4.19 shows the dike system on the rivers.
On the rivers in the United States, spur dikes, pile dikes,
pile dikes filled with stone, dikes and fencing dikes are widely
used, but earth-rock groin structures are used on the Yellow,
Yangtze and other rivers.
Vulnerable spot at
Penglou
Constraint at
Mazhangzhuang
Vulnerable spot at
Yingfang
Figure 4.18 — Typical alignment of river training of the Lower Yellow
River.
82
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
(2) Revetments. A revetment is used to stabilize concave
banks or to protect eroding bank lines of flood plains. In the
United States, various types of revetment are used, such as standard revetments with mattresses on the stream banks, standard
trench-fill revetments on stream banks, pile revetments with stone
fill and stone-fill revetments, etc., but in China, revetments use
earth-stone structures and standard revetments with a mattress, etc.
According to the needs of channel stabilization and
rectification, groins and revetments can be used simultaneously
for protecting banks or flood plains from erosion.
4.9.2.2 STRUCTURES OF TRAINING WORKS FOR FLOOD CHANNELS
The structures of training works on flood channels include
embankments and bank protecting works. Embankments should
be parallel with the direction of flood flow and the moderate flow
should be taken into consideration. In China, the flood peak
discharge with a recurrence interval of 0.33 to 1 per cent should be
taken as the design discharge for the most important embankments. The recurrence interval of important embankments is 2 per
cent, and for general embankments this figure is 10 to 5 per cent.
The top elevation and space of embankments are determined by
the hydraulic computation of water surface profile corresponding
to the design recurrence intervals. The design recurrence intervals
of embankments vary from country to country because they are
related to national flood defence policies.
The top width of the embankment depends on the path
of filtration and traffic requirements in flood seasons. The side
slope of the embankment is related to the soil properties of the
embankment, rates of rising and falling of water levels during
floods, the duration of floods, wind waves and filtration, etc. If the
embankment is composed of loam or sand loam and is higher than
5 m, and the flood duration is long, a side slope coefficient of 2.5
to 3.0 can be adopted. The Jingjiang embankment on the Yangtze
River is 10 m high and the flood duration is 1 to 3 months. The
side slope coefficient of its upstream slope is 2.5 to 3, and that of
the downstream slope is 3 to 6.3.
The bank protecting works in the flood channels are the
same as those in moderate and low flow channels.
4.9.2.3 DREDGING
Dredging is widely used in the improvement and maintenance of
navigation conditions in rivers and harbours. In recent years,
dredging has also been used in desilting reservoir sedimentation,
strengthening embankments, and forming and improving farmland.
The major problems associated with the disposal of
dredged material are: (i) ensuring the availability of a sufficient
disposal area for initial and future maintenance dredging within a
reasonable (economically feasible) distance of the dredging operations; and (ii) the potential adverse environmental effects
associated with the disposal of dredged material, including an
increase in turbidity, the resuspension of contaminated sediment
and a decrease in dissolved oxygen.
Costs and the potential environmental impacts are fundamental considerations when evaluating alternative dredging and
disposal methods and disposal sites, and many factors must be
considered in developing dredging operations, including:
(a) Determining the quality of the material to be dredged
initially and the frequency and quantity of future maintenance dredging;
(b) Sampling to determine the physical and chemical properties
of the material to be dredged to ensure that an appropriate
type of dredge is used, to assess dredged production rates so
that time and cost estimates are realistic, and to identify any
pollutants in the material to be dredged;
(c) Selecting an appropriate dredge type and size, disposal
method and disposal area to ensure environment protection;
(d) Identifying adequate disposal areas for both initial and future
maintenance dredging considering the physical properties of
the dredged material;
(e) Long-term management of disposal sites for a maximum
storage volume and beneficial use after the sites are filled
(Peterson, 1986).
4.9.3
4.9.3.1
River training of meandering rivers
MEASURES OF RIVER TRAINING FOR STABILIZING RIVER
CHANNELS
Figure 4.19 — Dike systems (after Peterson, 1986).
In order to stabilize the channels, bank protecting works are used
to prevent the successive collapse of banks on the concave side.
There are three types of bank protecting works which are
commonly used.
(1) Smooth bank protecting works. The anti-erosion materials or
matters are bedded directly on the banks or channel beds.
(2) Groins. Groins or spur-dikes are used to direct the flow.
(3) Combined use of smooth bank protecting works and groins
and spur-dikes. The combined works are often used to
protect the banks on a long reach.
Meanwhile, ecologically acceptable designs, e.g.
preserving or recreating meander bends and the range of geomorphological and flow environments for habitat improvement
purposes etc., should be taken into consideration.
CHAPTER 4 — FLUVIAL PROCESSES
4.9.3.2
CUTOFF
(1) Natural cutoff can develop on a meandering river,
as the neck between two neighbouring bends becomes thinner.
Once overbank flow occurs, the neck can be cut off and a new
channel connecting the two bends may form. Natural cutoffs
occurred 33 times between 1700 and 1870 on the Mississippi
River (Xie, 1987).
(2) Artificial cutoff. A limited pilot channel of a relatively small cross-sectional area is excavated to connect the long,
looping bends, enabling the excavated channel to be developed
and enlarged to full channel dimensions by the flow.
The length of the designed pilot channel is determined
by the cutoff ratio (length of original river channel / length of pilot
channel). The optimized cutoff ratio is 3 to 9, while the ratio of the
excavated cross-sectional area and the original cross-sectional area
is 1/5 to 1/30, and the side slope of the excavated channel is 1:2 to
1:3, according to data from rivers in China (Xie, 1987).
The cross-sections of the pilot channel should be made
as deep as possible in order to increase the channel’s flow velocity
and erosive capacity.
On navigation rivers, the pilot channel should be
designed according to the navigation standard. The excavated
width and depth should meet the needs of navigation so as to
ensure that navigation is unimpeded after excavation.
The entrance location of the pilot channel should be
designed in accordance with the configuration of meanders and
the geological structure of channel bends. The newly developed
channel should be protected by bank protecting works to avoid
the renewed development of long, looping bends. On the
Mississippi River a dike system had to be built in 1975 in order to
maintain the navigation depth.
4.9.4
River training of wandering rivers
The regulation of wandering rivers is mainly aimed at controlling the main current, rectifying the plan configuration,
transforming the wide-shallow and scattered channels into a
smooth, stable and single channel in order to increase flood
conveying capacity, and improving the conditions of flood
control. For wandering rivers with a high sediment concentration, river control is much more complicated. Soil and water
conservation works in upstream eroded areas and reservoirs on
the main stems and tributaries should be constructed so as to
adjust the conditions of incoming runoff and sediment load, in
addition to the training works on the rivers.
On the Lower Yellow River, the river training works
consist of the embankment, works at vulnerable spots and
constraint works. At vulnerable spots there are spur dikes and
groin systems, and bank protecting works are constructed along
the surface of the embankment where the flow often attacks in
order to protect the embankment and the banks. The constraint
works consisting of long groins with short spur-dike systems and
revetments are constructed along the embankment and flood plains
to protect the banks and flood plains, to control the main current
and to form a stable channel of moderate flow.
The length of the groin is two thirds of the practical
length. The angle between the flow and the groin is 30 to 45°. The
ratio of groin space to groin length is 0.8 to 1.04. After regulation,
the wandering range in the wandering reach of the Lower Yellow
River has decreased from 2 200 to 1 600 m, and the area of the
flood plains has increased by 9 000 ha. In the transitional reach,
83
the maximum wandering range has been reduced from 5 400 to
1 400 m, with the average range from 1 800 to 5 600 m. The water
depth at bankfull discharge has increased from 1.47–2.37 m to
2.05–3.73 m. All the changes in the wandering of the main
current, the shape of the cross-section and the curvature indicate
that the river channel has the tendency to be transformed into a
meandering river in the transitional reach (Xu, 1983).
4.9.5
River training of anabranched rivers
The aim of river training in anabranched rivers is to stabilize the
ratio of flow diversion, or to improve the conditions of water and
sediment transport in the main and fork channels.
4.9.5.1 MEASURES FOR STABILIZING FLOW DIVERSION RATIO
For the stabilization of flow diversion ratio, the plan configuration
of the anabranched river should be stabilized. Therefore, the
control works at the upstream node point, fish-mouth works at the
head of the island and bank protecting works at the entrance of the
branched reach should be constructed to fix the inlet flow and
island of the river.
4.9.5.2 WORKS OF FORK-CHANNEL BLOCKADE
On multi-branched reaches, measures for blocking fork-channels
and strengthening the main channel should be adopted to meet the
requirements relating to navigation and water supply for industry
and agriculture. A chute dike can be used to block the fork channel
on medium and minor rivers. On rivers with high sediment
concentrations, fence dikes and other infiltrated dikes are used,
which can easily be blocked by turbid water.
4.9.6
River training of straight rivers
The aim of river training on straight rivers is to fix the alternate
point bars so as to stabilize the straight channels. Dikes with an
upstream direction and submerged dikes can be used to reinforce
the point bars. Low dike systems have been used on the Rhine
River with favourable results.
4.9.7
Regulation of shoal reaches
The purpose of the regulation of shoal reaches is to improve
navigation conditions. The main training measures include: (1) the
construction of river training works to constrain the water flow, fix
the upstream and downstream point bars, block the subsidiary
branches, stabilize the bank lines and maintain the size of the
navigation course, and (2) dredging the navigation course to
decrease deposition and maintain the scale of the navigation
course.
4.9.7.1
PARAMETERS FOR DESIGNING NAVIGATION COURSES
(1) Assurance rate of navigation. The assurance rate of
navigation should be determined according to the state standard.
For example, the navigation standard for natural rivers in China is
listed in Table 4.21.
(2) Size of navigation courses.
(a) Minimum water depth.
dmin = t + ∆d
(4.182)
where dmin is the minimum water depth in the navigation course, t
is the maximum draught of allowance ships, and ∆d is the additional depth.
84
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
Table 4.21
Navigation standard for natural rivers
Grade of navigation course
Minimum depth at shoals (m)
Assurance rate of navigation (%)
(b)
1
2
3
4
5
6
7
73.2
2.5–3.0
1.8–2.3
1.5–1.8
1.2–1.5
1.0–1.2
0.8–1.0
98–99
93–97
90–95
85–95
80–93
80–90
75–90
Minimum width.
For double line course
bmin = 2 (b + b1)
(4.183)
where bmin is the minimum width of the navigation course; b is
the maximum width of allowance fleets, and b1 is the distance
between ships or between ship and bank. The width of the navigation course should be 5 times larger than the ship width.
(c) Curvature radius.
Rmin = (3 – 6) c
(4.184)
where R min is the minimum curvature radius, and c is the
maximum length of a fleet.
The length of a straight reach should be two times that
of the maximum length of a fleet. The determination of the control
line of the navigation course should refer to the data from rivers
that have similar conditions of hydrology, geology and navigation.
The conditions of navigable rivers include: (i) single
channel and no bifurcations in low water periods; (ii) smooth bank
lines, uniform curvatures, and appropriate length of crossing
reaches; (iii) no obvious difference between the depth in pools and
on the shoals of crossing reaches; (iv) uniform water surface
slope; (v) symmetrical cross-sections approximate to parabola on
crossing reaches; (vi) no criss-cross or short criss-cross between
upstream and downstream pools. A physical model is significant
for studying the regulation of important shoals and crossing
reaches.
REFERENCES
Andrews, E.D., 1980: Effective and bankfull discharge of streams
in the Yampa River basin, Colorado and Wyoming. Journal
of Hydrology, Volume 46, pp. 311-330.
Altwunin, C.T. and Y.A. Bozonov, 1957: Protective Constructions
of Rivers. Water Conservancy Press (in Chinese).
Altwunin, C.T., 1962: Regulated Rivers. Finance Press.
Benson, M.A. and D.M. Thomas, 1966: A dominant discharge.
Bulletin of the International Association of Scientific
Hydrology, Volume 11, Number 2, pp. 76-80.
Bray, D.T., 1982: Regime Equations for Gravel-bed Rivers. (ed.)
R.D. Hey, et al., pp. 517-552.
Bridge, J.S., 1977: Flow, bed topography, grain size and sedimentary structure in open channel bends: a three-dimensional
model. Earth Surface Processes and Landforms, pp. 401416.
Carlston, C.W., 1965: The relation of free meander geometry to
stream discharge and its geomorphic implications. American
Journal of Science, Volume 263, pp. 860-885.
Chai Tingsheng, 1963: Study on geometrical relations of the
Middle and Lower Yangtze River. Compilation of Research
Reports, Nanjing Hydraulic Research Institute.
China-Bangladesh Joint Expert Team, 1991: Study Report on
Flood Control and River Training Project on the
Brahmaputra River in Bangladesh. Volumes 1 and 2.
Chang, H.H., 1979: Geometry of rivers in regime. Journal of the
Hydraulics Division, ASCE, 105 (Hy6).
Chang, H.H., 1988: Fluvial Processes in River Engineering. John
Wiley and Sons.
Chang Jingdou and Sun Hancheng, 1983: Influences of changes of
bend flow conditions on formation of river patterns of Upper
and Lower Jingjiang River. Journal of Sediment Research,
Volume 1.
Chang Ruijing, 1961: Dynamics of River Flows. China Industrial
Press.
Chang Zhitang, Ling Wanquan and Shen Yongjian, 1982: Studies
on Dynamic Axis Line of Flow in Natural Meandering
Rivers. Compilation of papers of Changjiang River Science
Institute.
Chien Ning, 1958: Discussion on stability index of alluvial rivers.
Journal of Geography, Volume 24, Number 2, pp.128-144.
Chien Ning and Dai Dingzhong, 1980: The problems of river sedimentation and the present status of its research in China.
Proceedings of the International Symposium on River
Sedimentation, Volume 1, Guanghua Press.
Chien Ning, Zhang Ren and Zhou Zhide, 1987: Fluvial Processes.
Science Press (in Chinese).
Chien Ning and Zhou Wenhao, 1965: Fluvial Processes of the
Lower Yellow River. Science Press (in Chinese).
Chikurimora, 1969: Calculation of dominated discharge.
Proceedings of the Tenth Congress, IAHR, Volume 5.
Chitale, S.V., 1970: River channel patterns. Journal of the
Hyraulics Division, ASCE 96 (Hy1), pp. 201-222.
Dou Guoren, 1964: Channel morphology of alluvial rivers and
tidal estuaries. Journal of Hydraulic Engineering, pp. 1-13
(in Chinese).
Dury, G.H., 1964: Principles of under fit streams, Professional
Paper Number 352-A, USGS.
Emmett, W.W., 1975: The channels and waters of the Upper Salmon
River, Idaho. Professional Paper Number 870-1, USGS.
Englund, F., 1974: Flow and topography in channel bends.
Journal of the Hyraulics Division, ASCE, 100 (Hy11),
pp. 1631-1648.
Falcon-Ascanio, M. and J.F. Kennedy, 1983: Flow in alluvial river
curves. Journal of Fluid Mechanics, Volume 133, pp. 1-16.
Fang Zongdai, 1964: Classification of river patterns and its applications to river training. Journal of Hydraulic Engineering
(in Chinese).
Ferguson, R.I., 1973: Channel Pattern and Sediment Type Area.
Volume 5, pp. 38-41.
Friedman, G.M. and J.E. Sander, 1978: Principles of
Sedimentology. John Wiley and Sons.
Hack, J.T., 1957: Studies on longitudinal stream profiles in Virginia
and Maryland. Professional Paper Number 294-B, USGS.
CHAPTER 4 — FLUVIAL PROCESSES
Hey, R.D., 1975: Design discharge for natural channels. Science,
Technology and Environment Management.
Hey, R.D., et al., (ed.), 1982: Gravel-bed Rivers. John Wiley and
Sons, pp. 5-13 and pp. 553-580.
Huang Sheng and Zhong Xiujuan, 1986: The classification and
process of estuary in China. Proceedings of the Third
International Symposium on River Sedimentation, University
of Mississippi, pp. 501-506.
Ikeda, S., 1982: Bed load transport on side slope. Journal of
Hydraulic Engineering, ASCE, 108 (Hy11), pp. 1369-1373.
Institute of Hydraulic Research of the YRCC, 1978: Preliminary
analysis of effective volume of reservoirs on rivers carrying
huge sediment load. Collections of Reservoir Sediment.
Kellerhal, R., 1967: Stable channels with gravel-paved beds.
Journal of the Waterways and Harbor Division, ASCE, 93
pp. 63-64.
Lane, E.W., 1957: A study of the shape of channels formed by
natural streams flowing in erodible material.
Sediment Series Number 0, Missouri River Division, U.S.
Army Corps of Engineers.
Leopold, L.B. and T. Maddock, 1953: The hydraulic geometry of
stream channels and some physiographic implications.
Professional Paper Number 252, USGS.
Leopold, L.B. and M.G. Wolman, 1957: River channel patterns:
braid, meandering and straight. Professional Paper Number
282-B, USGS, pp. 45-62.
Leopold, L.B, M.G. Wolman and J.P. Miller, 1964: Fluvial
Processes in Geomorphology. W.H. Freeman.
Li Baoru, 1965: Calculation of longitudinal slopes and longitudinal profiles. Journal of Yellow River (in Chinese).
Ling Chenkun, 1963: Categorization of river patterns. Journal of
Nanjing University. Nanjing University Press.
Makkaveev, N.E., 1955: River Channel and Erosion in its
Watershed. Academy of the USSR Press.
Makkaveev, N.E, 1959: Morphological characteristics of river
systems. Fluvial Processes of Plain Rivers (in Russian).
Nixon, M.A., 1959: A study of the bankfull stages in rivers in
England and Wales. Proceedings of the Institute of Civil
Engineers, Paper Number 6322.
North-west Institute of Hydraulic Research, 1962: Analyses and
studies on river bed morphology at stations on the Middle
and Lower Weihe River. Compilations of River Regulations
of the Middle and Lower Weihe River.
Odgaard, A.J., 1981: Transversal bed slope in alluvial channel
bends. Journal of the Hydraulics Division, ASCE, 107
(Hy12), pp. 1677-1694.
Odgaard, A.J., 1982: Bed characteristics in alluvial channel bends.
Journal of the Hydraulics Division, ASCE, 108 (Hy11),
pp. 1268-1281.
Odgaard, A.J., 1984: Grain distribution of river bed armor layers.
Journal of the Hydraulics Division, ASCE, 110 (Hy9),
pp. 1267-1272.
Ouyang Lutai, 1983: Preliminary discussion on evolution and
stability of meanders. Journal of Sediment Research.
Parker, G., 1978: Self-formed straight rivers with equilibrium
banks and mobile beds (Part 2. The gravel river). Journal of
Fluid Mechanics, Vo1. 89, Part 1, pp. 127-146.
Parker, G., 1979: Hydraulic geometry of active gravel rivers.
Journal of the Hydraulics Division, ASCE, 105 (Hy9)
pp. 1185-1201.
85
Parker, G., 1984: Discussion of lateral bed load transport on side
slope, by S. Ikeda. Journal of Hydraulic Engineering, ASCE,
110 (2), pp. 197-199.
Peterson, M.S., 1986: River Engineering. Prentice-Hall,
Englewood Cliffs.
Pickup, G. and R.F. Warner, 1976: Effects of hydrologic regime on
magnitude and frequency of dominant discharge. Journal of
Hydrology, Volume 29, Number 1/2, pp. 51-76.
Riley, S.J., 1972: A comparison of morphometric measures of
bankfull. Journal of Hydrology, Volume 17, pp. 23-31.
Rohkin, F.M., 1948: River Mechanics, Hydraulic Problems of
Natural Rivers.
Romashen, V.V., 1977: Categories of Fluvial Processes.
Translations of Geomorphology (21): Rivers. Institute of
Geography, the Academy of Science of China, pp. 1-12.
Rosinski, K.U., 1950: River Channels. Academy of Science of the
USSR (in Russian).
Rozovski, Y.A., 1965: Circulating flow and its relations to water
surface. Proceedings of Fluvial Processes, China Science
Press (translated, in Chinese).
Schumm, S.A., 1960: The shape of alluvial channels in
relation to sediment type. Professional Paper Number 352-B,
USGS.
Schumm, S.A., 1968: River adjustment of altered hydrologic
regimen-Murrumbidgee River and plane channels, Australia.
Professional Paper Number 598, USGS.
Schumm, S.A., 1971: Fluvial Geomorphology: Channel
Adjustment and River Metamorphosis in River Mechanics.
(ed.) H. W. Shen, Volume 1.
Schumm, S.A. and H.R. Khan, 1972: Experimental study of
channel patterns. Bulletin, Volume 83, Number 6, Geological
Society of America, pp. 1755-1770.
Scientific Research Institute of the Yangtze River, 1959:
Characteristics of Jingjiang River. Journal of Sediment
Research.
Sedimentation Committee, CHES, 1992: Handbook of River
Sedimentation. China Environmental Press.
Shulits, S., 1941: Rational equation of river bed profile transaction. AGU, Volume 22, Number 1941.
Simons, D.B., 1979, River and Channel Morphology, Modeling
of River. (ed.) Hsieh Wen Shen, John Wiley and Sons.
The Yellow River Conservancy Committee, 1985: Training Yellow
River Press.
Velikanov, M.A., 1958: Fluvial Processes. Moscow Press.
Williams, G.P., 1978: Bankfull discharge of rivers. Water
Resources Research, Volume 14, Number 6, pp. 1141-1154.
Xie Jianheng, 1980: River Sedimentation Engineering. Water
Conservancy Press, pp. 235-311 (in Chinese).
Xie Jianheng, Ding Junsong and Wang Yunhui, 1987: Fluvial
Processes and River Training. Water Conservancy and
Electripower Press.
Xu Fuling, Guo Tiying, Hu Yisan, et al., 1983: Training means for
the Lower Yellow Reach of the Yellow River and their effect.
Proceedings of the Second International Symposium of River
Sedimentation, Nanjing.
Yangtze River Water Resources Committee (YRWRC), 1959:
Study on characteristics of the Jingjiang River. Journal of
Sedimentation Research, Volume 4.
Yangtze River Water Resources Committee (YRWRC), 1971:
River Training of Meandering River Cut-off. p. 97.
86
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
Yellow River Conservancy Committee (YRCC), 1981: River regulation. Yellow River Press.
Ying Xueliang, 1965: Causes for the formation of a meandering
river and experiment. Journal of Geology, Volume 31,
Number 4, pp. 287-303.
Yivanov, Y.V., 1951: Quantitative characteristics for longitudinal
profiles of rivers. Journal of Geographical Information,
Volume 82, Number 7 (in Russian).
Yu Jun, 1982: Studies and applications of geometrical formulae of
plain rivers. Renminchangjiang, Volume 3.
Zhang Ruijing, and Xie Baoling, 1980: Studies on fluvial
processes of meandering rivers. Proceedings of the
International Symposium on River Sedimentation, Volume 1,
Beijing, pp. 427-436.
Zhao Yan, Zhou Wenhao, Fei Xiangjun, et al., 1998: Basic laws of
fluvial processes of the Lower Yellow River. Yellow River
Press.
Zhou Wenhao and Chen Jianguo, 1998: River morphology and
channel stabilization of the Brahmaputra River in
Bangladesh. International Journal of Sediment Research,
Volume 13, Number 4, pp. 44-58.
Zhou Wenhao and Chen Jianguo, 1998: Behaviours of sediment
transport for different grain sizes in hyperconcentrated
floods of the Lower Yellow River. Proceedings of the
Seventh International Symposium on River Sedimentation,
Hong Kong.
Zhou Wenhao and Fan Zhao, 1982: Changes of longitudinal
profile of the Lower Yellow River. Journal of Sediment
Research, Volume 4 (in Chinese).
Zhou Wenhao and Fon Zhao, 1983: Influences of hyperconcentrated floods on degradation and aggradation of the Lower
Yellow River. Proceedings of the China Institute of Water
Resources and Hydropower Research, Volume 11.
Zhou Wenhao, Zeng Qinghua, et al., 1982: Characteristics of
fluvial processes for hyperconcentrated flow in the Yellow
River. Proceedings of the Second International Symposium
of River Sedimentation, pp. 618-626.
Zhou Wenhao, Zeng Qinghua, et al., 1995: Sediment transport
behaviours for different size groups in the Lower Yellow
River. International Sediment Research, Volume 10,
Number 3, pp. 51-68.
Zhou Wenhao and Zhao Huaxia, 1995: Characteristics for channel
deformation of the Brahmaputra River in Bangladesh.
Proceedings of Advances in Hydro-sciences and
Engineering, Volume II, March, pp. 1967-1974.
CHAPTER 5
RESERVOIR SEDIMENTATION AND IMPACT ON RIVER PROCESSES
5.1
INTRODUCTION
Reservoirs are built for many purposes, including flood
control, water supply (for agriculture, industry, and urban usage),
power generation, navigation and recreation, etc. As rivers carry
sediment load, whether in large or small amounts, reservoir
sedimentation occurs simultaneously with the impounding of
water. Meanwhile, the river channel downstream of the reservoir
experiences modifications induced by the changes in flow and
sediment regimes. Those changes upstream and downstream of
dams lead to intensive changes in environment, ecology and river
morphology, affecting engineering projects along the river, etc. It
is necessary to predict such changes. Making full use of the
benefits of reservoirs and developing appropriate measures to
mitigate the side effects of dam construction are a necessity for the
sustainable development of reservoirs.
5.1.1
Dam construction
According to the International Committee on Large Dams, large
dams (higher than 15 m) numbered 5 000 in 1950, and by 1985
the number had increased to more than 36 000, half of which were
in China. In recent years, however, the rate of construction has
decreased because very few good dam sites remain to be exploited
in a way that is both economically and environmentally sustainable in developed countries.
Furthermore, many reservoirs have been filling with
sediment, which depletes their storage capacity, and many have
exceeded their life expectancy. Some of them are to be
decommissioned.
Although there has been a decline in dam construction
in recent years, 292 dams higher than 60 m were still under
construction in 1994, including 68 in China, 48 in Japan and 37 in
Turkey.
The total storage capacity of reservoirs in the world has
been estimated by various sources. One estimation is 4 000 to
6 000 billion m3, and another is 5 per cent of the total runoff in the
world (38 830 billion m3), i.e. 2 000 billion m3. The percentages
of runoff regulated by reservoirs in each continent of the world are
listed in Table 5.1 (Beaumont, 1978). With the exception of South
America and Oceania, the percentages for the rest of the continents range from 14 to 21 per cent.
In China, dam construction has boomed since 1949,
when the People’s Republic of China was founded. As of 1995
there were 86 000 dams, 18 000 of which were large dams. There
are 358 large reservoirs (larger than 100 million m3), with a total
storage capacity of 300 billion m3, which accounts for more than
two thirds of the total reservoir storage capacity of the country.
5.1.2
Rate of loss of storage capacity
The rate of loss of storage capacity depends on the sediment yield of the river on which a reservoir is built, the
morphologic factors of the reservoir and the operational scheme of
a project, etc. In various regions, the rates of loss of reservoir
capacity are quite different.
Globally, the overall annual loss rate of reservoir storage
capacity is estimated at 1 to 2 per cent of the total storage capacity.
In China in 1989, 232 large and medium-sized reservoirs had a total
loss of 11.5 billion m3, accounting for 14.3 per cent of the total
capacity of 80.4 billion m3. Tables 5.2 (Qian, 1994) and 5.3 (Qian,
et al., 1987) list some loss rates of storage capacity in China. The
differences among the various reservoirs are quite significant.
At the end of the 1950s, an investigation was conducted
on the situation of sedimentation in 1 100 reservoirs in the United
States. Data from 66 representative reservoirs are listed in
Table 5.4 (Gottschalk, 1964).
Table 5.3
Total capacity loss of reservoirs in China
Reservoir
Qingtongxia
Yanguoxia
Gongzui
Sanmenxia
Guanting
Naodehai
Fengjiashan
Danjiangkou
River
Dam height Design
(m)
capacity
(106 m3)
Yellow
Yellow
Dadu
Yellow
Yongding
Liuhe
Qianhe
Hanjiang
42.7
57
88
106
45
41.5
73
110
605
220
310
3 760
2 270
196
389
16 050
Percentage
of loss (%)
93.0
74.6
71.0
39.0
24.3
19.5
5.9
3.9
Table 5.1
Percentage of regulated runoff (%)
Africa
North America
Europe
Asia*
Oceania
South America
20.6
15.1
14.0
6.1
4.1
21.0
* Not including China
Table 5.2
Annual loss rate of storage capacity in some provinces in China (%)
Shaanxi
3.02
Shanxi
Gansu
Inner Mongolia
Ningxia
Hebei
Shandong
Hubei
2.9
2.4
2.1
2.0
1.1
0.44
0.20
88
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
Table 5.4
Reservoir sedimentation in various regions in the United States
Region
North-east
South-east
Middle West
Middle south
North Great Plains
South-west
North-west
Total
Number of
reservoirs
Loss of storage
storage
capacity (%)
Annual loss
rate
3
10
11
12
9
15
6
66
24.7
18.6
14.0
8.8
9.6
15.7
7.0
15.6
0.82
0.81
0.85
0.51
1.28
0.53
0.30
0.71
The rate of loss of reservoir capacity is closely related to
the rate of erosion of the watershed above the reservoir. Table 5.5
provides evidence of such a situation (YRCC, 1993).
Table 5.5
Loss rate of reservoir capacity in a 30-year period in the
Yellow River basin
Rate of erosion
(t km–2.a)
20 000–30 000
15 000–20 000
10 000–15 000
5 000–10 000
2 000–5 000
1 000–2 000
500–1 000
200–500
100–200
<100
Rate of loss of
total capacity (%)
Annual rate of
loss (%)
52.6
51.2
41.1
43.1
41.1
20.1
15.4
14.0
11.7
3.8
1.75
1.71
1.37
1.44
1.37
0.67
0.51
0.47
0.39
0.13
5.1.3
Sustainable development of reservoirs
The concept of sustainable water management presumes socially
acceptable, ecologically sound, economically justifiable and technically feasible projects. It has strong ethical connotations, such as
environment protection, respect of future generations and equity
within our generation. The concept of sustainable development
has to be made operational: whatever the definition used for the
term ‘sustainable’, to make the definition operational, one must
list all the consequences of each possible decision, assess their
likelihood, and plan the optimum value system that will be used in
the future to evaluate these consequences.
Applied to reservoirs, the operational concept of
sustainable management presumes the extension of the useful life
of reservoirs to a reasonable maximum. In order to attain this goal,
appropriate decisions should be made at each phase of the
reservoir’s life cycle, including the planning, design,
implementation and operational stages. Once the end approaches,
the storage reservoir should be decommissioned with the least
possible harm to the affected society. In short, prolonging their
lifespan is a key issue for the sustainable development of
reservoirs.
How to preserve the long-term capacity of reservoirs is
the most important issue. Three basic methods of sediment control
for reservoirs are as follows: (1) decreasing the amount of sediment
that enters a reservoir by reducing sediment erosion from the
watershed upstream of the reservoir or by intercepting the sediment
before it enters the reservoir; (2) sluicing sediment-laden flows to
decrease the amount of sediment that deposits in the reservoir; and
(3) removing the deposited sediment by flushing, dredging, and/or
syphoning, etc.
5.1.4
Prediction of reservoir sedimentation
Nowadays, the prediction of reservoir sedimentation is mainly
based on mathematical modelling, although empirical methods are
still in use.
The major drawback of sedimentation models remains
the uncertainty of sediment transport computations and of the estimation of river channel resistance. These are the basic research
topics of lasting priority in sedimentation engineering.
5.1.5
Issues related to reservoir sedimentation
The construction of a dam in a river valley causes changes in the
flow regime, which consequently leads to a significant change in
sediment regime and a transformation of fluvial processes. Three
river reaches should be studied in this respect, namely the reservoir itself, the upstream reach and the reach below the dam.
The impacts of reservoir sedimentation manifest themselves in many areas, such as the environment, ecology, the safety
of the project, the economy, and society in general. These impacts
are discussed in Chapter 1.
5.2
PROCESSES OF DEPOSITION IN RESERVOIRS
5.2.1
Movement of sediment in reservoirs
Sediment movement mainly depends on water flow. In a reservoir,
there are two main patterns of flow motion, namely backwater
flow and quasi-uniform flow. Under the conditions of backwater
flow, the water depth increases longitudinally, and the flow velocity decreases accordingly. Sediment transport may have two
patterns. The first pattern is sediment transport under open channel
flow, where sediment particles diffuse to the whole section. As the
flow velocity decreases longitudinally, deposition takes place; this
is called backwater deposition. The second pattern is sediment
transport by density current, which is formed by a heavy sediment
load with fine particles, which dives into the bottom of the reservoir and moves along the channel bed toward the dam. The
sediment transport under quasi-uniform flow is similar to that of
natural rivers. When the incoming sediment load is different from
the sediment transport capacity of the flow, longitudinal deposition or erosion will occur. In summary, the sediment transport
patterns in reservoirs may be classified as follows:
(1) Sediment transport under quasi-uniform open channel flow;
(2) Sediment transport under backwater flow:
(i)
Sediment transport under open channel flow;
(ii) Sediment transport by density current.
5.2.2
Basic characteristics of reservoir deposits
(Qian, et al., 1987)
5.2.2.1 LONGITUDINAL PROFILES
There are three different shapes of longitudinal profiles of deposits
in reservoirs, namely delta, wedge and narrow band. The geometric shape of reservoir deposits depends on: (i) the composition and
diameter of the incoming sediment load; (ii) the amount of incoming load relative to the storage capacity; and (iii) the geometry and
operational mode of the reservoir.
CHAPTER 5 — RESERVIOIR SEDIMENTATION AND IMPACT ON RIVER PROCESSES
A delta forms in most impounding (storage) reservoirs in
which the ratio of the storage capacity, V, to the incoming annual
runoff, W, is large; the pool level is frequently kept at high elevations, and the incoming sediment load is comparatively coarse
(Figure 5.1).
A wedge forms in gorge-type reservoirs in which V/W is
small, incoming sediment is comparatively fine, and the pool level
frequently fluctuates. Sediment will soon reach the dam site, as
shown in Figure 5.2.
A narrow band may form in some of gorge-type reservoirs in which V/W is large, the incoming sediment is
comparatively fine and the pool level fluctuates frequently. This
shape of deposit is caused by the large fluctuations in pool level
(Figure 5.3).
Several rules of thumb have been developed to differentiate between the various shapes of deposits in reservoirs.
Jiao:
Delta
V/Ws > 2,
∆H/Ho < 0.15
Wedge
V/Ws < 2,
∆H/ Ho > 0.15
Figure 5.1 — Deltatic deposits in Guanting Reservoir, China.
Figure 5.2 — Wedge-shaped deposit in Bajiazui Reservoir, China.
250
Bed elevation (m)
240
230
220
210
200
190
135
115
95
75
55
Distance from dam (km)
35
15
Figure 5.3 — Narrow band deposit in Fengman Reservoir, China.
0
89
where V is the average storage capacity in a time interval, ∆T, in
m3, Ws is the incoming sediment load in ∆T, in m3, ∆H is the
amplitude of pool level in ∆T, in m3, and Ho is the average water
depth above the discharging outlet in ∆T (Jiao, 1980).
Another rule of thumb:
Delta
SV/Q > 108
∆H/Ho < 0.1
Band
0.25 × 108 < SV/Q < 108
0.1< ∆H/Ho < 1
Wedge SV/Q < 0.25 × 108
∆H/Ho > 0.1
where S is the sediment concentration in kg m–3, and Q is the
discharge in m3 s–1.
Luo used only one parameter, Ws/γs'V
Delta
0.78–1.75
Band
1.1–3.98
Wedge
4.38–5.2
where γs is the unit weight of deposits in t m–3 (Luo, 1977).
5.2.2.2 DELTA
(1)
Longitudinal profile. The longitudinal profile of a delta
can be divided into several reaches: tail reach, top-set reach, foreset
reach, and bottom-set reach.
(i) Tail reach: This is a transition reach between the natural
stream and the delta proper. The flow, after entering the
backwater zone created by the construction of the dam,
begins to deposit part of its sediment load. The bed becomes
progressively flatter and finer in composition along the river
course. Following the rise of the top-set in the reach immediately below, the tail reach will extend upstream at a slow
rate. The tail reach is usually of limited length for most
reservoirs, especially those built on mountain streams. The
characteristics of tail reach may be summarized as follows:
the reach has super-saturated sediment transport, a selective
deposition of sediment particles, a broad and shallow crosssection, and a wandering river reach.
(ii) Top-set reach: The top-set reach of the delta represents a
reach essentially in equilibrium. The selective process of the
bed material in the direction of flow is no longer perceptible.
Almost all the incoming load is able to move through this
reach and deposit on the foreset of the delta, making the
delta advance. This advancing of the delta causes the backwater to rise further. This, in turn, disrupts the temporary
balance maintained in the preceding stage and brings about
further deposition. The top-set bed will gradually rise as a
consequence of the advance of the delta. However, when
rising, the bed profile remains essentially parallel. Flow and
sediment transport in a state of equilibrium are the main
characteristics of the top-set reach.
(iii) Foreset reach: The water depth abruptly increases
downstream from the pivot point of the delta, and once
again selective settling of sediment particles occurs. The
bed in this reach is formed under circumstances similar to
those of the free settling of particles in a settling basin. The
slope of the foreset is slightly less than the angle of repose
of the sediment particles in still water. If density current is
formed and moves along the bottom of the reservoir, the
foreset will be modified and maintained with a much
smaller slope.
The main characteristics of the foreset reach are the
rapid increase in water depth, the drastic decrease in flow
velocity, the selective deposition of sediment particles, and
the advance of the delta.
90
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
(iv) Bottom-set reach: Materials brought to the bottom-set reach
are fine, usually those carried by the density current. The bed
slope is quite flat. Fine deposits and flat slopes are the main
features of the bottom-set reach.
(2) Quasi-equilibrium state. The establishment of the
quasi-equilibrium state in a reservoir starts at the top-set. The relationship between the top-set slope, J, and the original slope, Jo, is
shown in Figure 5.4, in which data from 45 reservoirs are
included. Three straight lines represent:
Line 1
J = Jo
Line 2
J = 0.5 Jo
Line 3
J = 0.2 Jo
Most of the points are close to Line 2. This means that
the equilibrium slope of the top-set is smaller than the original bed
slope. The changes in slope imply that in addition to the river
gradient, the composition of bed material may also play an active
role in accomplishing the readjustment of the new river channel.
In a reservoir, part of the incoming wash load may turn into bed
material, and consequently the deposit will be finer than the original bed material. Figure 5.5 shows the relationship between D/Do
and J/Jo, where Do is the D50 of the original river bed, and D is the
D50 of the deposit of the top-set. The finer the deposit of the topset, the flatter the equilibrium slope of the top-set.
The quasi-equilibrium state is reached under the adjustment of all factors related to the formation of a river channel,
among which the bed slope and the composition of bed material
may be of most prominence.
There are many empirical expressions to determine the
equilibrium slope of the top-set.
By definition, an equilibrium slope is in dynamic equilibrium, i.e. there is no obvious deformation over a comparatively
long period. On the slope, the flow is uniform. For suspended load
and bed load, the equilibrium slopes are different.
The governing factors of an equilibrium slope are: (1)
dominant discharge; (2) river roughness; (3) river bed composition; (4) sediment transport capacity: either suspended load or bed
load; (5) river channel morphology: hydraulic geometry. In addition, some engineers believe that a raising of the base level may
have some effect.
Analytical or empirical approaches can be used to estimate the equilibrium slope. For an analytical approach, four
conditions must be fulfilled when an equilibrium state is reached:
(1) uniform flow; (2) flow continuity; (3) sediment transport in
saturation; (4) channel morphology in shape. There are four equations for solving four unknowns.
(i) For suspended load:
V =
1
2 / 3 1/ 2
(5.1)
J
n
Q = BhV
(5.2)
3
V
*
ρ = K(
)
m
(5.3)
ghW
B
=C
B=A
or
h
Q 0.5
(5.4)
J 0.2
2
10–1
n C
J=
0.4 *0.73 / m
ρ
K
0.73 / m
B=A
10–2
Top-set slope (J)
R
Q
J
J=
10–3
W
Q
0.73 0.73
g
0.2
(5.5)
0.5
0.2
n 20 / 11 A 5 / 11ρ *25 / 33m ( gW ) 35 / 33
(5.6)
(5.7)
K 25 / 33m Q10 / 44
(ii) For bed load, one should use a bed load transport
formula, e.g. Meyer-Peter and Muller formulae, or others.
10–4
10–4
10–3
10–2
Original bed slope (Jo)
10–1
10
(
Ks
γ
)γhJ = 0.047(γ s − γ ) Db + 0.125( )1 / 3
Kr
g
Figure 5.4 — Relationship between the topset slope and original slope.
(
1.0
γ s − γ 2 / 3 Qb 2 / 3
)
( )
γ
B
(5.8)
where Qb is the bed load discharge in t s–1, and γs and γ are the
specific weights of sediment and water, respectively in t m–3, and
Db is the diameter of bed load in m;
D/Do
€
Ks =
0.5
Kr =
1
n
(5.9)
26
1/ 6
D90
(5.10)
Adopting B as a constant, one obtains:
0
[(
0
0.5
1.0
J/Jo
Figure 5.5 — Relationship between D/Do and J/Jo.
J=
γ −γ
0.125 γ s − γ Qb 2 / 3
)
+ 0.047 s
Db ]10
γs B
γ
γg1 / 2
K
Q
( s )15 / 7 n 6 / 7 ( b ) 6 / 7
Kr
B
(5.11)
CHAPTER 5 — RESERVIOIR SEDIMENTATION AND IMPACT ON RIVER PROCESSES
For an empirical approach, there are various expressions
based on field data.
(i) For suspended load:
(a) IWHR (11 reservoirs)
J = 1.28 × 10 −4 (
ρω
q 0.6
) 0.305
(5.12)
where q is the unit discharge in the flood season in m3 s–1, ρ is the
mean sediment concentration in the flood season in kg m–3, and ω
is the mean settling velocity of suspended load in cm s–1.
(b) Li (based on rivers and models) (Shaanxi Institute and
Tsinghua University, 1979)
ρ
J = 0.00455[( ) 0.5 D50 ]0.59
Q
(5.13)
where Q is the bankfull discharge in m3 s–1, ρ is the mean concentration of bed material load in flood season in kg m–3, and D50 is
the D50 of bed material, in mm.
(c) Establishing a relationship between J and Jo, original
bed slope, for example:
J / Jo = 19.5(
d 50 0.1 1 0.15
) (
)
D50
HV
(5.14)
where d50 is the d50 of incoming sediment load in mm, D50 is the
D50 of original bed material in mm, H is the raising of the base
level in m, and V is reservoir capacity relevant to H in m3.
(ii) For bed load:
J
= 0.79( HQJo ) −0.17
Jo
(5.15)
where Q is the mean annual discharge in m3 s–1.
5.2.2.3 LATERAL DISTRIBUTION OF DEPOSITS
The lateral distribution of deposits depends on the location of the
cross-section, the operational mode of the reservoir and the sediment concentration of the river, etc.
For reservoirs built on sediment-laden rivers, the parallel
raising of the near-dam cross-sections in impounding reservoirs is
typical, as in the Guanting and Sanmenxia (1960–1964)
Reservoirs.
In reservoirs with drawdown in flood seasons, high flood
plains and deep main channels occur. The size of the main channel
depends on the discharging capacity of the outlets, as in
Sanmenxia (1964–1973), Naodehai and Heisonglin Reservoirs.
The flood plains rise in elevation and no surface erosion occurs;
only banks might collapse.
In the fluctuating backwater region in reservoirs built on
clear rivers, deposition mainly takes place in the main channel, but
in the permanent backwater region, parallel raising of the channel
bed may take place.
5.2.2.4 SPATIAL DISTRIBUTION OF DEPOSITS
Understanding spatial distribution is useful for determining the
depletion of each part of the storage capacity, which is the basis
for planning the future operation of reservoirs.
Nowadays, analytical methods are commonly applied to
solve this problem by using computer sediment models. However,
there are still many empirical methods in usage. One of them is
the empirical area-reduction method, developed by Borland and
Miller (Borland and Miller, 1960) based on field data from 30
reservoirs in the United States. In Figure 5.6 there are four curves
representing four types of reservoir morphology with various
distributions of sediment.
Type I
Lake
m = 3.5–4.5
Type II
Flood plain-foothill
m = 2.5–3.5
Type III
Hill
m = 1.5–2.5
Type IV
Gorge
m = 1.0–1.5
where m is the exponent in the expression V = Nhm, h is the water
depth at the dam site, and V is the storage capacity at h.
Based on Table 5.6, the weighted class of a reservoir is
selected. Where a choice of two types is given, sediment particle
size is used to determine which to choose according to Table 5.7.
For the user’s convenience, a working diagram is
plotted, as shown in Figure 5.7. The steps of the empirical area
reduction method are as follows: (1) Determine sediment inflow;
(2) Select the design curve; (3) Compute new zero-capacity at the
dam site: use the basic expression F = (Vs – Vo)/HAo to compute
po using Figure 5.9, where Vs is the total sediment deposition, Vo
is the reservoir capacity at each elevation h, H-original is the depth
Table 5.6
Selection of the weighted class of a reservoir
Reservoir operation
Operational
class
Shape
class
Weighted
class
I
I
II
III
I
I or II
II
II
I
II
III
I or II
II
II or III
III
I
II
III
II
II or III
III
IV
All
IV
Sediment submerged
Moderate drawdown
Percentage of reservoir depth
91
Considerable drawdown
Normally empty
Table 5.7
Effect of sediment particle size
Predominant particle size
Percentage of sediment deposited
Figure 5.6— Relative distribution of deposits in reservoirs.
Sand or coarser
Silt
Clay
Type
I
II
III
92
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
a
(iii) Compute the area at each elevation occupied by sediment
(aAo);
(iv) Compute the sediment volume for each stage increment
above the new zero-capacity elevation: Vs = 0.5 (A1 + A2) H
(v) Compute the revised area and capacity curves.
This method is suitable for large impounding reservoirs.
h
—
H
Figure 5.7 — Relative depth versus relative area of deposition.
of the reservoir normal pool, Ao is the reservoir area at a given
elevation, p = (h – hmin)/H, hmin is the original bottom elevation,
and ho = poH + hmin; (4) Distribute the sediment:
(i) Compute a, relative sediment area at each relative depth, p:
Type I
a = 5.047p1.85(1-p)0.36
Type II
a = 2.487p0.57(1-p)0.41
Type III
a = 16.97p1.15(1-p)2.32
Type IV
a = 1.486p–0.25(1-p)1.34
(ii) Compute Ao/ao, area correction factor (relevant to po);
5.2.2.5 HEADWARD EXTENSION OF BACKWATER DEPOSITION
The location of the terminal of the backwater region is not fixed; it
shifts to and fro. However, the long-term trend is headward extension.
In pace with the advancing and rising of a delta, the
backwater will extend upstream, which, in turn, causes deposition
to propagate upstream. In certain circumstances, the headward
extension of backwater deposition may develop to a grand scale,
hampering the drainage and flood control of riparian lands.
As regards the location of the terminal of the backwater
region in Sanmenxia Reservoir, for 18 years the location of the
terminal in the main channel shifted upstream discontinuously, as
in 1964, 1966, 1970, and 1977 it was pushed downstream by
floods. Only the terminal on the flood plain extended upward
continuously.
In Figure 5.8, an empirical relationship is established to
determine the extent of headward extension of backwater deposition.
5.2.2.6 PHYSICAL CHARACTERISTICS OF DEPOSITS
(1)
Longitudinal distribution of deposit diameter. The
incoming coarse sediment almost all deposits in the tail reach; at
the entrance of the top-set the bed material rapidly becomes finer.
On the top-set the bed material is almost uniform, and is much
finer than the original bed material. At the entrance of the foreset
bed, the material becomes finer once again. At the bottom set, the
deposit of density current is mainly within the range of 0.002 to
0.003 mm. A turning point exists in most of the curves, which lies
at the location of 60 to 80 per cent of the length of the backwater
from the dam site.
(2) Unit weight of deposits. The unit weight of deposits
is mainly determined by the initial unit weight, the operational
mode of the reservoir, and the consolidation rate of the deposits.
(i) Initial unit weight of deposits. Figure 5.9 shows the initial
unit weight of deposits of different particle sizes. Han, et al.
presented an expression for the initial unit weight of
deposits, as follows (Han, et al., 1981).
For D ≤ 1 mm:
D
)3
D + 4δ
(5.16)
Unit weeight (1 m–1)
γ in = 1.41(
Diameter (mm)
Figure 5.8 — Relationship between ∆H and S/QJ.
Figure 5.9 — Relationship between initial unit weight of deposit and
sediment size.
CHAPTER 5 — RESERVIOIR SEDIMENTATION AND IMPACT ON RIVER PROCESSES
Table 5.9
Value of k
For D > 1 mm:
γ in = 1.89 − 0.47 exp(−0.095
D − D1
)
D1
(5.17)
where δ is the thickness of the film water, δ = 4*10–4, D1 is the
critical diameter, D1 = 1 mm, and δin is in t m–3.
Lara and Pemberton analysed 1 300 samples of reservoir deposits in the United States and gave a measure of the
effect of the operational mode of reservoirs on the initial unit
weight of deposits, as listed in Table 5.8 (Lara and Pemberton,
1965).
The initial unit weight of a mixture may be calculated by
the following expression:
γsin = acPc + amPm + asPs
γ st = γ so
Operational mode
k (for metric units)
Clay
Silt
Sand
256
135
0
91
29
0
0
0
0
1
2
3
NOTE: Operational mode: 1-deposits submerged under water for long-term period
2-pool level drops in medium or large-scale
3-long term dry reservoir.
Table 5.10
Long-term unit weight
(5.18)
where ac, am and as are the initial unit weights for clay, silt and
sand, respectively (Table 5.8), and Pc, Pm and Ps are the percentages of clay, silt and sand in the mixture, respectively.
(ii) The effect of duration of deposition. Miller developed an
approximate expression for determining the average unit
weight of a deposited mixture in T years, as follows
(Miller, 1953):
T
+ 0.434 k(
L nT − 1)
T −1
93
Sediment
Size (mm)
Unit weight (t m–3)
Clay
Silt
Medium and fine sand
Coarse sand and fine gravel
Medium gravel
< 0.005
0.005–0.05
0.01–0.5
0.5–10
> 10
0.8–1.2
1.0–1.3
1.3–1.5
1.4–1.8
1.7–2.1
η= f(
(5.19)
where γin is the average unit weight after T years of reservoir operation; γso is the initial unit weight, and k is the constant related to
the operational mode of the reservoir and sediment size; its values
are given in Table 5.9.
(iii) Long-term unit weight of deposits. This can be determined
using Table 5.10.
5.3
SEDIMENT RELEASE FROM RESERVOIRS
5.3.1
Sediment release during flood detention
For reservoirs with serious deposition, it is necessary to know
how the situation will develop after a flood. During a flood,
water may be discharged from the reservoir at some low-level
outlets or spillways, but flood detention occurs when the
incoming flow discharge is larger than the outgoing flow
discharge. Under such circumstances part of the incoming
sediment load deposits in the reservoir and the rest sluices out of
the reservoir.
The discharging efficiency (= 1 – trap efficiency) in a
period of time (e.g. during a flood), η, is a function of the size of
the incoming sediment load, the duration of the sediment particles in the reservoir, the characteristics of the reservoir and the
ratio of the incoming water discharge to the outgoing water
discharge, etc.
1 1 1
, , )
VQi ω B
(5.20)
Qo2
where V is the storage capacity below the highest pool level
during a flood, Qi and Qo are the average inflow and outflow
discharges, respectively, and ω is the mean settling velocity of
the suspended load.
Based on data from several Chinese reservoirs, an
empirical diagram (Figure 5.10) is obtained for determining the
discharge efficiency.
D50 (mm)
S (kg m–3)
Figure 5.10 — Discharge efficiency.
Table 5.8
Initial unit weight of sediment
Initial unit weight (kg m–3)
Operational mode of reservoirs
Sediment always or nearly submerged
Moderate to considerable reservoir drawdown
Empty reservoir
River bed sediment
ac Clay < 0.004 mm)
am Silt (0.004–0.062 mm)
as Sand (0.062–2.0 mm)
416
561
641
961
1 120
1 140
1 150
1 170
1 550
1 550
1 550
1 550
94
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
Table 5.11
Density differences of density current
Reservoir
Density difference (kg m–3)
Guanting
20–100
Sanmenxia
10–50
Lake Mead
5–25
Sautet
0.5–1.0
Silting basin
0.5–100
Estuaries
Figure 5.11 — Schematic diagram of density current.
0–2
Salt water intrusion
For small and medium-sized reservoirs, the amount of
deposition during a flood may be determined by the following
expression, which is based on data from reservoirs built on sediment-laden rivers in north-west China. The incoming sediment is
loess with particles of 0.008 to 0.0375 mm. The duration of the
flood peak is less than 2 days, and the detention period is from 1
to 6 days.
η = ηw1.5
(5.21)
where ηw is the water release efficiency, ηw = Wo/Wi, Wo is the
outgoing flow during the detention period, and Wi is the incoming
flow during the detention period.
5.3.2
Density current venting
5.3.2.1 PHENOMENON AND FORMATION OF DENSITY CURRENT
When two fluids with a similar state but slightly different densities
move in relation to each other, a density current may form.
When a turbid sediment-laden flow enters a clear water
reservoir, a density current may form if the turbid flow has enough
velocity and fine particles. The density current moves along the
reservoir bed towards the dam. Under favourable conditions the
Salt content difference 0–3%
density current may reach the dam. If the bottom outlet is opened
in time, the density current may be vented out of the reservoir.
The turbid open channel flow dives into the bottom at
the plunge point, which is the point of separation between the
forward moving current and the induced reverse flow in the reservoir. This point can be distinguished by the collection of floating
debris on the reservoir surface. The head of the density current is
thicker than the main body, as the head provides the potential
energy necessary to overcome the inertia of the reservoir water
ahead of the current. Resistance also exists at the interface, which
induces the mixing of the density current and the surrounding
water.
The schematic diagram of density current and measured
data from Sanmenxia Reservoir are shown in Figure 5.11.
5.3.2.2 VENTING OF DENSITY CURRENT
Factors affecting the venting of density current are the
incoming flow and sediment conditions, the topography of the
reservoir, and outlet facilities (elevation, location, discharge
capacity, etc).
Table 5.12
Venting of density currents
Reservoir
Dam height
(m)
Capacity
(109 m3)
Annual
runoff
75
0.16
0.21
5.21
1.52
7.47
1.31
0.65
3.64
25
43
49
221
38.4
16.0
7.78
9.48
9.35
11.08
1.79
2.37
3.27
2.00
23
25
39
18
Nebeur (Tunisia)
65
0.30
0.18
Fengjiashan (China)
77
0.40
0.48
0.46
1.18
0.11
0.77
23
65
Guanting (China)
45
2.27
1.40
7.86
13.5
5.30
20.5
6.34
1.63
2.70
4.0
1.06
4.56
1.58
0.29
34
30
20
22
25
18
106
96.4*
43.2
1.70
1.47
0.30
0.31
18
21
Iri Emda (Algeria)
Lake Mead (United States)
Sanmenxia (China)
* Before reconstruction
Sediment load
In
Out
Annual
4.9
3.5
% of venting
59–64
95
5.3.3
Erosion in reservoirs
5.3.3.1 RETROGRESSIVE AND PROGRESSIVE EROSION
Although reservoirs are environments for sediment deposition,
erosion can still take place when conditions are favourable. In
reservoirs, two types of erosion may occur, namely retrogressive
erosion and progressive erosion.
When the pool level drops by a certain amount, erosion
may first take place at the pivot point of the delta and then develop
upstream. This is retrogressive erosion.
Progressive erosion takes place when the sedimentcarrying capacity is greater than the incoming sediment load.
Erosion develops and its intensity decreases. This is a common
phenomenon caused by the imbalance of the incoming sediment
load and the sediment-carrying capacity. An example of retrogressive erosion is shown in Figure 5.14.
Figure 5.14 — Retrogressive erosion in Sanmenxia Reservoir.
5.3.3.2 EROSION IN THE FLUCTUATING BACKWATER REGION
The fluctuating backwater region has dual characteristics: when it
is submerged, it belongs to the reservoir; when the pool level
Figure 5.12 — Release efficiency of density current with original river
bed slope.
Sanmenxia Reservoir
Guanting Reservoir
Heisonglin Reservoir
Lake Mead
Figure 5.13 — Release efficiency of density current with reservoir
length.
Q (m3 s–1)
Twenty-seven sets of field data from Guanting Reservoir
in 1956 to 1957 show that the discharge of density current is equal
to more than one half of the incoming flow discharge, and the
sediment load of the density current is about one quarter of the
incoming sediment load; the rest deposits near the plunge point
and only the fine particles form the density current.
In Table 5.11 the density differences of density current
measured in some reservoirs are listed. In Table 5.12 field data of
venting of density currents are listed.
Figures 5.12 and 5.13 show the relationship between the
release efficiency of density current and the characteristics of the
reservoir.
sediment discharge (t s–1)
CHAPTER 5 — RESERVIOIR SEDIMENTATION AND IMPACT ON RIVER PROCESSES
Figure 5.15 — Schematic diagram of retrogressive erosion of
suspended load.
drops and this region is out of the effect of backwater, it belongs
to the river. During the latter situation, erosion takes place in this
region. Two types of erosion may occur in this region:
(a) Erosion during drawdown: in dry seasons the pool level
gradually drops, and progressive erosion takes place on the
previously deposited sediment bed;
(b) Erosion during reservoir filling: from the beginning of the
flood season the river discharge gradually increases; during
the filling process erosion may occur in the fluctuating backwater region, which may push the terminal of backwater
deposits downward.
5.3.3.3 EMPIRICAL METHOD OF EROSION PREDICTION
The sediment carried by rivers is classified as suspended load and
bed load. In some rivers most of the transported sediment belongs
to suspended load, while in other rivers bed load accounts for a
major portion.
(1) Prediction of retrogressive erosion of suspended
load transportation. Figure 5.15 is the schematic diagram for
calculating the retrogressive erosion of suspended load.
During the period of ∆t, retrogressive erosion develops
from point B to point A in pace with the pool level drawdown from
Zo to Z1. The eroded volume ABC may be expressed as follows:
γ∆V = (Qso – Qsi)∆t
(5.22)
where γ is the unit weight of the deposit in t m–3, ∆V is the eroded
volume in the period of t in m–3, Qso is the sediment load at the
exit cross-section in t s–1, and Qsi is the sediment load at the
entrance cross-section in t s–1.
If Qso can be determined, the value of ∆V may be calculated. Some empirical formulae to determine Q so have been
96
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
derived on the basis of data from reservoirs in China. The most
commonly used formula is as follows:
Q50 = Ψ
Q16 J 12
B 0.6
same principle as that used to calculate retrogressive erosion.
Consequently, the equation and section diagram may also be used
to predict progressive erosion.
(5.23)
5.4
EMPIRICAL ESTIMATION OF LONG-TERM
DEPOSITION IN RESERVOIRS
5.4.1
Method of trap efficiency
The ratio of the sediment deposited in a reservoir to the total
incoming sediment is called the trap efficiency of the reservoir.
Trap efficiency is related to various parameters, such as the ratio
of reservoir storage capacity, V, to the average annual runoff, W;
the ratio of retention period to the average flow velocity in the
reservoir; and the specific storage of the reservoir, i.e. the ratio of
the reservoir storage to the river basin area above the reservoir.
The most commonly used method was developed by
Brune (Brune, 1953). In Figure 5.17, Brune determined the relationship between β and V/W based on large reservoirs in the
United States. Data from reservoirs in China and the Russian
Federation also follow the general trend of the Brune curve.
Other than the influence of V/W on trap efficiency, the
size of sediment particles, the operational mode of the reservoir
and the type of outlets also influence trap efficiency, as shown in
Figure 5.17. The average value of β may be determined by the
following expression:
β=
V
W
0.012 + 0.0102
V
W
V/W
Figure 5.17 — Trap efficiency (after Brune, 1953).
€
θ = Q0.6J12/B0.6
Figure 5.16 — The value of Ψ.
(5.24)
Churchill presented a method to estimate the trap
efficiency of a reservoir, using the sediment index of the reservoir,
which is defined as the period of retention divided by mean
velocity (Figure 5.18). The sediment index may be expressed as
V2/Q2L (s2 m–1), which is a dimensional index, where Q is the
average daily mean discharge, and L is the length of the backwater
region (Churchill, 1947).
The scattering of the points in the Churchill diagram is
less than that in the Brune diagram. This may be explained by the
Q10 (t s–1)
where Ψ is the parameter expressing the resistance of the river bed
in the unit of s0.6t m–4.2, Q is the discharge in m3 s–1, J is the
slope, and B is the channel width in m. It may be determined by
the method of hydraulic geometry.
The value of Ψ is determined by filed data from 10
reservoirs and the Yellow, Weihe and Fenhe rivers where the
erosion is progressive (Figure 5.16). In the diagram there are three
lines: Ψ = 650, representing the river bed composed of newly
deposited fine sediment (D50 < 0.1 mm); Ψ = 300, representing
the medium situation (D50 > 0.1 mm); and Ψ = 180, representing
the river bed composed of cohesive sediment.
The range of parameters of the field data are
Q = 0.1–5730 m 3 s –1 , J = (0.006–1.6)%, B = 10–1 000 m,
Qso = 0.0006–777 t s–1.
(2) Prediction of retrogressive erosion of bed load transportation. The process and principle of the calculation of the
retrogressive erosion of bed load transportation are the same as
those of suspended load transportation. The only difference lies in
adopting the sediment transport capacity of the bed load instead of
that of the suspended load.
Many bed load formulae are available, but they must be
verified by the field data from the river where the calculation will
be carried out.
(3) Prediction of progressive erosion. The basic principle for calculating progressive erosion is that the difference
between the outgoing and incoming sediment loads of a river
reach is equal to the volume scoured from the river bed. This is the
V2/(Q2L)
Figure 5.18 — Trap efficiency (after Churchill, 1948).
CHAPTER 5 — RESERVIOIR SEDIMENTATION AND IMPACT ON RIVER PROCESSES
fact that the sediment index includes more parameters than the
Brune index. Data from reservoirs in China also confirm the validity of the Churchill curve.
5.4.2
Method of rate of storage capacity loss
The rate of storage capacity loss may be expressed as:
α=
∆Ws ∆Ws Ws
W
β
=
=β s =
V
Ws V
V
φ
(5.25)
If trap efficiency is determined by the Brune curve or the
Churchill curve, then the value of α may be determined.
Where flow and sediment data are insufficient at the
planning stage of some small and medium-sized reservoirs, an
empirical expression for determining the value of α is obtained
based on 25 reservoirs mainly in North and Northwest China.
 V −0.8
α = 0.0002G 0.95  
F 
(5.26)
In Table 5.13, Vi is the storage capacity at t years of the reservoir’s
operation in m3, Vo is the initial storage capacity in m3, and Ws is
the annual sediment load in m3; Wo is the final volume of deposits
in m3, Wt is the amount (volume) of deposition at t in m3, βo is the
initial trap efficiency, βt is the trap efficiency at t, Q is the annual
incoming discharge in m3 s–1, S is the annual incoming sediment
concentration in kg m–3, Wr is the residual channel volume when
the equilibrium is established in m3, and γ 's is the unit weight of
deposit in kg m–3.
The expression presented by Tsinghua University is
described in detail as follows:
 W n
βt = βo 1 − s 
 Wo 
Process of depletion of reservoir storage capacity
(lifespan of a reservoir)
The rate of siltation in a reservoir decreases with time as the
storage capacity is reduced, until a residual river channel remains
in the reservoir. The difference between the original total storage
capacity and the remaining storage is called the storage of siltation. It is important to estimate the process of siltation in order to
estimate the benefit of a reservoir.
In the 1930s, the first expression for the estimation of
the remaining storage capacity was presented as follows:
 W t
Vt = Vo 1 − s 
 Vo 
(5.27)
where Vt is the storage capacity at t years of the reservoir’s operation in m3, Vo is the initial storage capacity in m3, and Ws is the
annual sediment load in m3.
At present, there are many empirical expressions for
estimating the process of depletion of storage capacity, as listed in
Table 5.13.
Table 5.13
Expressions for estimating reservoir life span
Author
Year
Expression
Orlt
1930
Vt = Vo (1 – Ws/Vo)t
Shamov
1950
Vt = Wo (1 – βWs/Wo)t
Gangchalov 1960
V
γ 's
(Wt + Wr ln o )
QS
VoWt
1965
t=
Tsinghua
University
1979
(1) n = 1 Wt = Wo [1 − (1 −

1 
  β (1 − n _ W t ) 1− n 
o
s 

Wt = Wo 1 − 1
 
Wo
 


(5.29)
where n ≠ 1.
Assuming Wt = ζWo, where ζ is the extent of siltation in
the reservoir, the following expression can be deduced:
1− n
T=
1 − (1 − ξ )
(1 − n)αV
(5.30)
o
βoWs
(5.31)
Wo
In practice, the value of n ranges between 0 and 1.
The value of n in some reservoirs in China is listed in
Table 5.14.
The less the sediment sluiced from the reservoir, the
smaller the value of n.
αVo =
where
Table 5.14
Value of n
Reservoir
Bajiazui
Yanfuoxia
Fenhe
Gufengshan
Hongshan
Cetian
Jioucheng
n
Note
0.95
0.90
0.75
0.75
0.65
0.45
0
Sluicing sediment
in between
in between
in between
in between
in between
Storing sediment
Because of the difficulty of determining the value of n,
the original expression may be used to determine n. On a log-log
paper, the relationship between βt and (1 – Ws/Wo) is plotted. The
slope of the line represents the value of n.
Vt/Vo = 1 – (1 – W1/Wo)t
Shineer
(5.28)
where βt is the trap efficiency at time t, βo is the initial trap efficiency, Ws is the amount of deposition at time t, Wo is the storage
of siltation, and n is an index expressing the decreasing rate of trap
efficiency.
After operation:
where G is the annual rate of erosion in the basin above a reservoir
in t km–2a, F is the drainage area above the reservoir in m2, V is
the reservoir storage capacity in m3, and α is the rate of storage
capacity loss in %.
5.4.3
97
1
β(1 − n)Wst 1− n
) ]
Wo
βW
(2) n = 1 Wt = Wo [1 − (1 − o s ) t ]
Wo
5.5
NUMERICAL MODELLING OF RESERVOIR
SEDIMENTATION
5.5.1
General
Based on the laws governing water flow and sediment transport,
numerical models of reservoir sedimentation can be established
98
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
and used to predict the future situation of reservoir sedimentation.
The processes for establishing the numerical model include three
steps of approximate schematization and four steps of feedback.
The first step of approximate schematization is to describe the
engineering problem by physical processes; the second step is to
describe the physical processes by mathematical equations, and
the third one is to obtain the numerical solution of the mathematical equations. Each feedback step is the process of verifying each
step of approximate schematization.
Sediment transport and its induced channel deformation
are the result of water flow motion. Simultaneously, the deformed
channel morphology has its effect on flow motion. Therefore, a
sediment numerical model includes two submodels of flow motion
and sediment transport. These two submodels should be solved
simultaneously, and their solutions are called coupled solutions.
When channel deformation is not so intensive, to simplify the
computation process, the two submodels can be solved step by
step, the first being that of flow motion and the second being that
of sediment transport. Such a solution is called an uncoupled solution and is common practice nowadays.
The development of numerical models is seeing a move
from one-dimensional to three-dimensional models. The natural
situation is always a three-dimensional one. At present, threedimensional numerical sediment models are still only on the
horizon, as the commonly used numerical models are either oneor two-dimensional. The selection of a suitable numerical model
depends on the characteristics of the problem. If a onedimensional model can simulate the problem, it is unnecessary to
use a two-dimensional model, since the computer time of the latter
is much longer than the former. In some special cases, a combined
model may be used. In some river reaches a one-dimensional
model is used, and in the remaining river reaches a two-dimensional model is used to meet engineering requirements.
At present, no analytical solution can be obtained for
any sediment numerical model. Numerical approaches must be
used to find the solution. There are a number of numerical
approaches, including the finite difference approach, which is the
most common.
The numerical model must be calibrated and verified by
separate sets of field data. The accuracy of the result of verification must meet engineering requirements.
5.5.2
5.5.2.1
Basic equations (for unit width)
CONTINUITY EQUATION
(1) Continuity equation of water:
∂
∂
∂Z
[vh(1 − S)] + [h(1 − S)] + (1 − p)
=0
∂x
∂t
∂t
(5.32)
(2) Continuity equation of sediment:
∂
∂
∂Z
(vhS) + ( hS) + p
=0
∂x
∂t
∂t
(5.33)
(3) Continuity equation of sediment-laden flow:
∂
∂h ∂Z
(vh) +
+
=0
∂x
∂t ∂t
(5.34)
Equation 5.34 is a combination of Equations 5.32 and 5.33.
Among these three equations, only two of them are independent.
When z = 0, i.e. a fixed bed, then Equation 5.34 becomes the
continuity equation of unsteady flow.
5.5.2.2
MOMENTUM EQUATION OF ONE-DIMENSIONAL
SEDIMENT-LADEN FLOW
(1) Forces include:
(i)
Pressure of sediment-water mixture
Where the specific weight of sediment-water mixture, γo is as
follows.
γo = γ + S (γs – γ) = γsS + (1 + S) γ
(5.35)
where γo and γ are the specific weight of sediment and water,
respectively.
(ii) Component of self-weight in x-direction
(iii) Bed resistance
Jf =
τo
γoh
(5.36)
where τo is the resistance per unit area.
(2) Change in momentum in unit time includes two parts:
momentum change with time and difference of momentum going
into and out of the unit section dx. The momentum equation of
one-dimensional sediment-laden flow is as follows:
1 ∂v v ∂v ∂h ∂Z γ s − γ h ∂S
+
+
+
+
−
g ∂t g ∂x ∂x ∂x
γ o 2 ∂x
v pγ s + (1 − p)γ ∂Z
[
]
= Jo − J f
h
gγ o
∂t
where
Jo = −
∂Zo
∂x
(5.37)
(5.38)
Compared with the momentum equation of clear flow,
items 5 and 6 in Equation 5.37 are added. Item 5 is the water pressure induced by the longitudinal variation of sediment
concentration, while item 6 is the change in momentum induced
by sediment deposition.
5.5.2.3 SUPPLEMENTARY EQUATION
There are three independent equations for computing river bed
changes, but there are four unknowns, v, h, s, and z. Therefore, one
more equation is needed. There are two methods to supplement
one equation.
(1) Saturated sediment transport. Assuming that
sediment concentration is always equal to the sediment transport
capacity of flow, a formula of sediment transport capacity can
be adopted as the supplementary equation. This assumption is
nearly true when the longitudinal variation of sediment transport
capacity is small, and the incoming sediment concentration of
the river reach is not too different from the sediment transport
capacity.
A finite difference method is commonly used to solve
the equations. In most cases, unsteady flow is simplified as steady
flow.
(2) Non-saturated sediment transport. When the sediment concentration is high and the longitudinal variation of the
hydraulic parameters is significant, the assumption of saturated
sediment transport cannot be upheld. In such circumstances, nonsaturated sediment transport must be considered in numerical
models. By integrating the diffusion equation in two-dimensional
steady flow along a vertical, one can obtain the basic equation of
longitudinal variation of mean sediment concentration in onedimensional flow, as follows:
CHAPTER 5 — RESERVIOIR SEDIMENTATION AND IMPACT ON RIVER PROCESSES
(5.39)
where qs is the unit sediment discharge, q is the unit discharge, Sg
is the bottom sediment concentration, and Sm is the mean sediment
concentration of a cross-section.
After a series of operations, one finally obtains the basic
equation of longitudinal variation of the mean sediment concentration under a steady state:
S = S* + ( So − S*o )e
−
αL
l
+ ( S*o − S* )
1
1
(1 − e
αL
αL
l )
(5.40)
where S* is the sediment transport capacity at the exit crosssection, So is the sediment concentration at the entrance, S*o is the
sediment transport capacity at the entrance, α is a coefficient of
recovery of sediment concentration, L is the length of the river
reach, and l is the horizontal distance for a particle settled within a
water depth, ho.
From Equation 5.40, one can find that the sediment
concentration at the exit cross-section is composed of three parts:
item 1 — sediment concentration in saturation at the exit crosssection; item 2 — the attenuated value of residual sediment
concentration at the entrance cross-section, (s o – s *o ), after
distance L/l; item 3 — modified sediment concentration in saturation in the river reach.
In Table 5.15 field data from a desilting channel are
used to check the necessity of the method of non-saturated
sediment transport. From the Table one can conclude that the
values of items 2 and 3 in Equation 5.40 are too large to neglect,
i.e. the consideration of non-saturated sediment transport is a
necessity.
For non-uniform sediment, the total sediment is divided
into n groups. For each group:
Si = Pi S (i = 1, 2, 3,…., n)
(5.41)
S*i = Pi S* (i = 1, 2, 3,…., n)
(5.42)
where P i is the percentage of the ith group to the total (by
weight). For each group, the basic equation of longitudinal
variation of mean sediment concentration is valid. The
computation procedure is quite lengthy. Those readers interested
in this subject are advised to read relevant literatures (e.g. Han,
1990).
5.6
RESERVOIR SEDIMENTATION MANAGEMENT
5.6.1
Universality of reservoir sedimentation
In Figure 5.19, reservoirs are classified according to φ and ψ, with
sediment concentration as the third parameter. Here, φ and ψ
denote the ratios of reservoir storage capacity to annual sediment
load and water runoff, respectively. In Figure 5.19, the points can
be classified into three groups. All the points fall close to one of
the three lines representing different types of rivers. The first
group represents the reservoirs built on clear rivers with sediment
concentrations lower than 1 kg m–3. The second group represents
the reservoirs built on rivers with medium concentrations from 1
to 10 kg m–3. The third group represents the reservoirs built on
heavily sediment-laden rivers with concentrations higher than
10 kg m–3. For the first group, reservoir sedimentation is not a
problem, while for the third group it is very serious.
The features of deposition and experience of reservoir
sedimentation management are more valuable a reference for
River of low sediment conc. (< 1 kg m–3)
–3
° River of medium sediment conc. (1-10 kg m )
+ River of high sediment conc. (> 10 kg m–3)
φ
dq s
dS
∂S
= q m =ε
− ωSg
∂y y = h
dx
dx
99
Ψ
Figure 5.19 — Relationship between Ψ and φ of reservoirs.
Table 5.15
Verification of Equation 5.40 by field data (in kg m–3)
Seconds
2
3
4
5
6
7
8
9
Item
Concentration
(1)
(2)
(3)
Calculated
26.2
23.4
13.3
9.77
6.97
7.46
9.33
24.6
5.42
–1.64
0.33
5.64
8.73
9.56
9.24
7.06
–8.22
2.24
7.57
3.09
2.50
–0.32
–1.77
–14.8
23.4
24.0
21.2
18.5
18.2
16.7
16.8
16.9
Error
Measured
30.8
27.8
23.5
19.2
17.6
16.7
14.8
14.3
(%)
–24.0
–13.7
–9.8
–3.6
+3.4
0
+13.5
+18.2
100
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
reservoirs in the same group, although the general law of reservoir
sedimentation is the same.
5.6.2
Indicators of reservoir sedimentation problems
In China, specifications for sediment design of hydropower and
water conservancy projects have been issued. In these specifications, the states of distress caused by the sediment problems of
hydraulic projects are classified into two grades according to the
degree of seriousness with which sediment affects the safety and
benefits of the project, namely serious and non-serious effects.
When one of the following situations occurs, the state of
distress is considered to be serious.
(1) φ is less than 50 to 100.
(2) Upstream extension of backwater deposits is so serious that
the safety of cities and industrial regions, etc. and the normal
operations of existing large or medium-sized hydraulic
projects are affected.
(3) A mouth bar may occur at the confluence of a tributary with
the main river, which may affect the functions of the reservoir.
(4) Deposition in the dam area may affect normal operation of
the inlet or outlet structures.
(5) Sedimentation may impede navigation on the river.
These specifications are mainly stipulated based on
practices in China during the past four decades.
5.6.3
Basic operating rules
Operating rules of reservoirs have a decisive influence on reservoir sedimentation. Three basic types of operating rules have been
adopted in China, namely impoundment, impounding the clear
and discharging the turbid water (I and D), and flood detention.
The first two types are often adopted. In Table 5.16 some basic
characteristics of reservoir operating rules are listed.
In Figure 5.19 various operating rules are also shown.
The long-term capacity of a reservoir is the remaining
storage capacity when the equilibrium state in the reservoir is
reached. The storage of a reservoir consists of two parts, namely
that over the flood plains and that of the main channel
(Figure 5.20). The storage capacity over the flood plains will be
gradually lost by deposition of sediment carried by overflows of
flood peaks and cannot be recovered. The loss of the storage
capacity over the flood plains is almost perpetual. A part of the
storage capacity in the main channel may be preserved through
rational use of the reservoir by lowering the pool level in the flood
season and storing during the rest of the year. This part of the
storage capacity is called the long-term capacity of the reservoir.
Under such an operational scheme, the pool level is kept
low to sluice the incoming sediment load during the flood seasons.
For small or medium-sized reservoirs, drawdown flushing is often
necessary to maintain the long-term capacity. During the rest of
the year, the water carries much less sediment than during the
flood seasons; storing water will not induce much deposition in
the reservoir.
One of the prerequisites for maintaining the long-term
capacity of a reservoir is to install sluicing outlets with sufficient capacity and proper bottom elevation in the reservoir. With
such facilities, sediment may be easily sluiced downstream and
a useful storage capacity will be maintained on a permanent
basis.
5.6.4
Sediment design of hydrological projects
At the feasibility study stage of large and medium-sized hydrological projects, sediment design should be carried out. Much
attention should be paid to basic data collection, the reservoir sediment regulation mode should be carefully studied, and calculation
approaches of reservoir sedimentation should be properly selected
based on the characteristics of river sediment and the project.
States of distress caused by the sediment problems of hydrological
projects are classified into two categories according to the degree
(a) Longitudinal profile
(b) Cross-section
Figure 5.20 — Sketch of terminal capacity of reservoirs.
Table 5.16
Reservoir operating rules
No.
Operating schemes
Regulation of sediment
Method of sediment sluicing
Period of sediment sluicing
A1
Impoundment sediment
totally trapped
None
None or dredging
None
A2
Impoundment sediment
partly trapped
None
Density current venting
sluicing
Beginning of flood seasons
B
Impounding the clear and
discharging the muddy water
Yearly or seasonally
Sluicing sediment during
detention, density current
Flood seasons
C
Detention
Sluicing
Sediment during detention,
reservoir emptying
Flood seasons
No. — Effect of sediment sluicing on downstream channels: A1-None; A2-No serious problems; B-Non-matching of flow and sediment may cause problems; C-same as B.
CHAPTER 5 — RESERVIOIR SEDIMENTATION AND IMPACT ON RIVER PROCESSES
(ii)
of seriousness with which sediment problems affect the safety and
benefits of the projects, namely serious and non-serious effects.
For projects in the serious category, sediment problems should be
studied specifically. If necessary, physical modelling should be
carried out.
5.6.4.1 COLLECTION AND EVALUATION OF BASIC DATA
The basic conditions of a basin in which a hydrological project
will be built need to be understood comprehensively. These conditions include physiographic and socio-economic conditions,
climate, hydrologic and river characteristics, soil erosion, human
activities and soil conservation, etc. The basic data include: topography charts, longitudinal profiles and cross-sections; water
surface profiles, bed materials, fluvial processes; and landslides,
bank failures and debris flows. These data relate to the reservoir
area and the river reach below the hydrological project. Details on
the location and elevation of cities, towns, industrial areas, mines,
and hydrological projects on the river reach affected by the design
project should be collected. Elements such as daily and monthly
flow discharges, suspended load discharges, bed load discharges,
sediment particle composition, mineral composition and water
temperature are the basic hydrologic data required for design
purposes. Hydrologic data from other relevant hydrologic stations
in the same basin are also needed. The characteristics of other
projects on the river, such as operational modes, should be
analysed to optimize the design for the project.
All collected data should be analysed and their rationality and reliability should be evaluated. For projects with serious
sediment problems, measured sediment data are a must.
5.6.4.2 SEDIMENT INPUT
The distribution and characteristics of sediment source areas in the
upstream basin of the design project need to be studied in detail.
For large projects with serious sediment problems, the reconnaissance of key sediment source areas should be carried out. The
effect of existing upstream projects on the sediment input of the
design project should be analysed.
For suspended load, the direct use of 20 years of consecutive hydrologic data from a hydrologic station with a difference
of watersheds between the station and the project of less than three
per cent is necessary. When the difference of watersheds is larger
than three per cent and less than 20 per cent, the difference should
be calculated. The yearly and monthly variations of suspended
loads are the main items to be analysed.
For bed load, based on field measurements or empirical
methods (formulae), the relationships between flow discharge and
bed load discharge and between flow discharge and bed load
diameter are analysed.
5.6.4.3 SEDIMENT DESIGN
For better reservoir management and to maximize resources, in
addition to hydraulic design, sediment design of a project should
be carried out, particularly for projects on sediment-laden rivers.
With a rational sediment design, the project will maintain useful
storage capacity for long-term usage.
(1) Requirements
(i)
For projects with serious sediment problems, the sediment regulation modes of operation, reservoir
operations and sediment release facilities should be
studied comprehensively.
101
(2)
(3)
Predictions should be made relating to reservoir sedimentation (amount, location, elevation, spatial
distribution, and depletion process) and also released
sediment discharge, concentration, and diameter.
(iii) For reservoirs with arms, the appearance of the river
mouth bar at the confluence should be studied.
(iv) For projects with a lengthy construction phase, sediment problems during the construction stage,
including the effects on diversion and project layout,
etc., should be studied.
(v) The effect of upstream projects on the design project
and the effect of the design project on upstream and
downstream projects should be analysed.
(vi) For navigable rivers, the effect of scour and deposition
in the fluctuating backwater region and the effect on
navigation of fluvial processes downstream of the
project should be studied.
Sediment regulation modes of operation
(i)
A comparison should be made of various alternatives
according to river sediment characteristics, reservoir
characteristics (shape, objectives and regulation
requirements, etc.), and environmental requirements.
(ii) For storage reservoirs, the pool level should be kept at
a certain level to sluice sediment during the whole, or
part, of the flood season; if the pool level is not
controlled, sediment can be sluiced by venting the
density current or reservoir emptying.
(iii) For low-head diversion projects, sediment regulation
should be carried out at several (a maximum of three)
discharges, or all of the sluices can be opened to flush
the sediment.
Calculation of reservoir sedimentation
(i)
Calculations can be carried out using numerical
models or empirical methods.
(ii) Calibration and verification of numerical models or
verification of empirical methods must be carried
out.
(iii) The rationality of the calculated results should be
checked. For projects with serious sediment problems,
several calculation methods may be adopted for
comparison.
(iv) As for the data series for calculation, long-term series,
representative series, wet-normal-dry years, or a representative year may be adopted. The average annual
sediment load and concentration of the adopted series
should approximate the long-term values.
(v) As regards the term of calculation, when the term of
quasi-equilibrium of deposition is longer than the term
of the lifespan of the key structures of a project, the
latter should be adopted as the term of calculation.
When the former is shorter than the latter, the former
should be adopted as the term of calculation.
(vi) When the trap efficiency is less than 10 per cent, it is
considered that the quasi-equilibrium of deposition
has been reached.
5.6.4.4
PREVENTION OF SEDIMENT PROBLEMS
(1) The dam site, power house and tail channel, etc.
should not be near a sediment-laden tributary (including abundant
bed load) or an active debris flow valley, etc.
102
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
(2) Where sediment deposition affects the normal operations of a project, sediment prevention measures should be
considered seriously and sediment release facilities should be
constructed in the project.
(3) For projects on navigable rivers, the approach channels of a ship lock should be studied and the corresponding
measures to mitigate sediment deposition should be adopted.
5.6.4.5
PREDICTION OF THE FLUVIAL PROCESSES BELOW A
PROJECT
This is mainly recommended for projects which significantly
change the natural flow and sediment regimes, as such fluvial
processes may have serious implications below dams.
5.6.4.6 PLANNING FOR SEDIMENT MEASUREMENT
For large and medium-sized projects with serious sediment problems, sediment measurement should be carried out from the very
beginning, or ideally before the impoundment of a reservoir.
5.6.5
Methods of reducing sediment input in reservoirs
A range of measures can be adopted to reduce sediment supply in
reservoirs.
5.6.5.1 SOIL CONSERVATION PRACTICE
The effectiveness of soil conservation in reducing sediment input
in a reservoir depends on the size of the watershed where the
reservoir is built. For a large watershed with poor natural conditions, soil conservation can hardly be effective over a short period
of time. Nevertheless, if the watershed is not very large, the effect
of soil conservation can be seen in a short time.
A good example of this is the Middle Yellow River basin.
The hydrologic data of the Yellow River show an obvious
reduction in surface runoff and sediment load in the 1970s, and even
Table 5.17
Annual runoff and sediment load at Sanmenxia
Item
1940s 1950s
1960s
1970s
1980s
Annual runoff
(billion m3)
47.8
52.6
59.3
51.3
55.8
Annual sediment
load (billion t)
1.73
1.98
1.98
1.82
1.51
Table 5.18
Areas of soil conservation works above Sanmenxia (million ha)
Year
Terraces
Reclaimed
farmland
1969
1979
1989
0.574
1.84
2.597
0.036
0.086
0.17
Afforestation Grassland Total
0.759
1.35
3.92
0.209
0.317
1.23
1.58
3.60
7.92
Table 5.19
Sedimentation rate in Guanting Reservoir
Period
1956–1960
1961–1970
1971–1980
Amount of
deposition (million m3)
Annual amount of
deposition (million m3)
82
73
70
8.2
7.3
a remarkable reduction in sediment load in the 1980s, as listed in
Table 5.17. Besides the climatic variations, human activity has
played an important role in such a reduction. The effect of human
activity may be classified into two categories: water resources
development and soil conservation.
The areas of soil conservation work above Sanmenxia
are listed in Table 5.18. The rapid development of soil conservation work is obvious and shows a close association with the
reduction in sediment loads.
Another example is Guanting Reservoir on the Yongding
River, which controls a catchment of 43 000 km 2. The mean
annual river flow at the dam site is 1.4 billion m3 and the annual
sediment load is 81 million tons. The reservoir storage is 2.27
billion m3. The project was commissioned in 1955. Reservoir sedimentation is very serious, but it has been quite different at
different periods (see Table 5.19).
Although the average annual precipitation and precipitation in flood seasons in the 1950s, 1960s, and 1970s were almost
the same, the incoming runoff and sediment load in Guanting
Reservoir have declined significantly since 1960 under the influence of human activities, as listed in Table 5.20.
The measures for reducing sediment load and their
respective effects are listed in Table 5.21.
Another approach is to bypass the input of sediment.
This method is mainly used for small or medium-sized
reservoirs where the topography is suitable for bypassing the
incoming sediment. An example is shown in Figure 5.21.
Unfortunately, this is not always successful as sediment can block
bypass channels and the topography may be unfavourable for such
a method.
From the example of Guanting Reservoir, it is obvious
that warping (colmatage) has a good impact in dealing with
reservoir sedimentation. At the same time, it increases the fertility of
the irrigated land.
The joint operation of reservoirs has also proven to be of
value at Guanting Reservoir (see section 5.8.5).
Table 5.20
Runoff and sediment load in Guanting Reservoir
Period
1951–
1960
1961–
1970
1971–
1980
Precipitation (mm)
Annual
runoff
Annual sediment
load
Annual
Flood
season
(million m3)
(million t)
444
338
1 723
59.69
412
313
1 258
15.08
427
373
832
10.23
Table 5.21
Reduction of sediment load in Guanting Reservoir
Cause of reduction
Annual reduction of sediment load
(million t)
Upstream reservoirs
Irrigation and warping
Soil conservation
17
19
5
Total
41
CHAPTER 5 — RESERVIOIR SEDIMENTATION AND IMPACT ON RIVER PROCESSES
Flood weir
Mgeni River
Flood gates
To Pietermaritzburg
0
1
2
3
4
5 kg
|__________________________
Scale
Figure 5.21 — Plan view of Mgeni Reservoir.
5.6.6
Overview of remedial measures
5.6.6.1 DRAWDOWN FLUSHING
When low-level outlets are opened, the pool level in a reservoir
drops. Consequently, the flow in the reservoir will be favourable
for scouring the previous deposits. Thus, the storage capacity of
the reservoir can be enlarged. This method is mainly used for
small and medium-sized reservoirs.
There are many examples of this method. Shuicaozi
Reservoir in China is one such example. Drawdown flushing was
carried out eight times during the period from 1964 to 1981. The
quantity of sediment flushed out each time was some 200 000 m3,
corresponding to about one third of the annual incoming sediment
load (see section 5.8.4). Although this is beneficial in preventing a
loss in reservoir capacity, there can be short-term impacts downstream. Timing and the associated water discharges are important
considerations.
5.6.6.2 RESERVOIR EMPTYING
Reservoir emptying is the limit of drawdown flushing. It is very
efficient in eroding sediment out of a reservoir, but it can only be
used in small reservoirs. Water consumption is the problem with
this measure.
Hengshan Reservoir in China is an example of reservoir
emptying. In Hengshan Reservoir (V = 13.3 million m 3), the
emptying operation was not carried out annually. After the first
eight years of reservoir operation from 1966 to 1973, 3.2 million
m3 of deposits accumulated in the reservoir. The reservoir emptying operation took 37 days in 1974, and a storage capacity of
0.8 million m3 was recovered. Reservoir emptying operations
were executed from 8 to 21 August 1976, 9 August to
30 September 1979, and 28 May to 16 June 1982, and a storage
capacity of some 1 million m3 was recovered each time.
5.6.6.4 SIPHON DREDGING
Siphon dredging makes use of the water head difference between
the upstream and downstream levels of a dam as a power source
for the suction of deposits from the reservoir to the downstream
area. It is an old method adopted for small reservoirs in some
countries. Since 1975 this method has been applied in some small
reservoirs in semi-arid areas in China, and the flushed mixture of
water and sediment is diverted into farmland for warping and irrigation. The diameters of the pipes used in China range from 0.3 to
0.6 m, the discharges range from 0.2 to 1.2 m 3 s –1 , and the
maximum sediment concentration of the flushed mixture ranges
from 500 to 1 200 kg m–3.
5.6.6.5 DREDGING
Dredging is a measure to remove deposits in small and mediumsized reservoirs. The advantages of this method are: it is highly
efficient (less water consumption), the normal operations of the
project are maintained, it can be executed at any place, the evacuated fine material can be used on farmland, coarse material can be
used for construction, and there is no limit for the recovery of
storage capacity. The disadvantages are high costs (US$ 2–4 m–3
worldwide), difficulties in disposing of dredged deposits, and
environmental problems.
5.6.6.6
DESIGN OF SEDIMENT SLUICING FACILITIES OF
RESERVOIRS
The location, elevation, size, and type of facilities are the design
elements. Some empirical formulae have been derived for determining the outlet capacity for flushing sediment and maintaining
the long-term capacity of a reservoir.
(1) Shaanxi Institute of Hydrotechnical Research
Based on the sediment transport capacity in some reservoirs, the adequate discharge capacity of a sluicing outlet, Qe, may
be determined by the following expression,
Qe = (
Ws
KTJ 1.2
)1 / 1.6
(5.43)
where Ws is the annual sediment load in tons, T is the duration of
sediment sluicing period in sec, J is the slope, determined as one
of the situations in Figure 5.22, and K is the coefficient (K = 3 in
most cases).
(2) Tsinghua University
Based on data from existing reservoirs, Tsinghua
University proposed:
Qe = (30–50) Qfm0.6
5.6.6.3 LATERAL EROSION
This technique is mainly used for recovering storage capacity on
flood plains. The objective is to break flood plain deposits and
flush them out by the combined actions of scouring and gravitational erosion caused by the large lateral gradient of the flood
plains. In so doing, it is necessary to build a low dam at the
upstream end of the reservoir for diverting water into diversion
canals along the perimeter of the reservoir, and the flow is
collected in trenches on the flood plains.
Guanshan Reservoir in China is an example of this technique. A 2-metre high diversion dam was built at the upstream end
of the reservoir. The diversion canal is 1 300 m long with a gradient
of 0.001. The scouring discharge was 0.5 m3 s–1. Within two
months, 0.4 million m3 of deposits were flushed out of the reservoir.
€
103
(5.44)
where Qfm is the average discharge in the flood season.
(a) Emptying
(b) Controlled operation
Figure 5.22 — Schematic diagram for determining the capacity of
outlets.
104
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
Table 5.22
Ratios of post- and pre-dam discharge
Mean annual discharge
Peak discharge
5% flood discharge
95% discharge
Range
Mean
Range
Mean
Range
Mean
Range
Mean
0.46–1.48
0.91
0.15–0.91
0.45
0.31–2.1
0.88
0–2.67
1.01
5.7
FLUVIAL PROCESSES BELOW RESERVOIRS
5.7.1
Fluvial processes below impounding reservoirs
5.7.1.1 CHANGES IN FLOW REGIME
Dam construction leads to a change in the flow regime below the
dam. As regards water flow, the main changes are the reduction in
peak discharges, an increase in the duration of medium flows, an
increase in low discharges, and a decrease in the seasonal and
annual variation of discharges. As regards sediment transport, the
main changes are the reduction of released sediment amount and
the loss of coarse sediment particles. However, the variations may
differ significantly due to the difference between the reservoir
storage capacity and operational mode. Williams and Wolman
analysed 21 reservoirs in the prairie and semiarid western region
of the United States. The mean annual discharge of the rivers
ranges between 1.5 and 930 m3 s–1. In Table 5.22, the changes
after the commissioning of dams are listed. The greatest change is
the reduction of peak discharge. The post-dam peak discharge is
about 45 per cent of the pre-dam value (Williams and Wolman,
1984).
In Sanmenxia Reservoir, the incoming peak discharge of
12 400 m3 s–1 in 1964 was reduced to 4 870 m3 s–1, or only about
40 per cent of the original. The duration of medium flow,
4 000 m 3 s –1, increased from 59 to 73 days. Meanwhile, the
seasonal and annual variation of discharge decreased.
5.7.1.2
DRASTIC REDUCTION IN SEDIMENT LOAD AND
CONCENTRATION
When most of the sediment is trapped in a reservoir, the released
water will be clear. Consequently, the sediment load and
concentration in the river reach below the reservoir will be much
lower than the pre-dam values. Table 5.23 lists the changes in
sediment concentration in several reservoirs. Below Guanting
Reservoir the sediment concentration of the river amounted to only
one-tenth of the pre-dam value.
5.7.1.3 EROSION BELOW DAMS
Erosion takes place below the reservoir where the released
water is clear. The distance of erosion may be quite long and
depends on the released flow discharge. Erosion develops gradually downstream. Table 5.24 shows the development of erosion
below Danjiangkou Reservoir.
In the Lower Yellow River erosion took place for
800 km along the reach, and below the Aswan High Dam on the
Nile the length of eroded reaches extends for about 1 000 km.
The erosion thickness depends on many factors of the
river channel, and varies in different rivers. For example, in the
Yellow River below Sanmenxia Reservoir, the thickness of
erosion after 4 years of clear water erosion was 1 m (the mean
diameter of bed material ranged from 0.06 to 0.1 mm). On the
Nile, where the mean diameter of the bed material was
0.15 mm, erosion thickness was 0.1 m after 3 years of erosion.
5.7.1.4 ARMOURING OF BED SEDIMENT
The selective process of water flow is the principal cause of
armouring of bed sediment. In addition, the imbalanced
exchange between suspended load, bed load, and bed material is
also responsible for the armouring of bed sediment.
Table 5.23
Ratio of post-dam to pre-dam sediment concentration
(2) Sanmenxia Dam
Discharge (m3 s–1)
1 000–2 000
3 000
Jinmenzha1
0.21
0.10
0.83
0.53
Shifosi2
0.24
0.12
0.10
0.78
Huayuankou3
0.36
0.18
Gaocun4
0.42
0.24
1, 2, 3, and 4 are 160 km, 190 km, 280 km and 485 km below the dams, respectively.
Finer than D50 by weight (%)
(1) Guanting Dam
Year
1956
1957
1958
1959
(a)
(b)
(c)
Table 5.24
Length of eroded reaches below Danjiangkou Reservoir
Year
Length (km)
1960
1968
1972
223
465
Remarks
Beginning of flood detention
Commissioning of the dam
Diameter (mm)
Figure 5.23 — Three types of bed armouring.
CHAPTER 5 — RESERVIOIR SEDIMENTATION AND IMPACT ON RIVER PROCESSES
105
Table 5.25
Change in channel width of the Yongding River
Reach
Channel width (m)
Length (km)
Pre-dam
Lugouqiao
Jinmenzha
Shifosi
Jinmenzha
Lugouqiao
Figure 5.24 — Longitudinal profiles below dams.
There are three types of armouring of bed sediment as
shown in Figure 5.23: (1) Gravel bed covered with sand bed, such
as the Colorado River below the Hoover Dam (Figure 5.23a); (2)
Gravel and sand bed, such as the Yongding River (Figure 5.23b);
(3) Fine sand bed, such as the Colorado River below the Imperial
Dam (Figure 5.23c).
The armouring of bed material has three effects: (1)
There is an increase in channel roughness, as at Yuma Station on
the Colorado River below the dams. When the diameter of the bed
material increased from 0.15 to 0.3 mm, the value of the Manning
roughness coefficient, n, increased from 0.013 to 0.032; (2) There
is a decrease in sediment-transport capacity; (3) There is a restriction on further degradation of the channel bed.
5.7.1.5 ADJUSTMENT OF LONGITUDINAL PROFILE
There are two different scenarios for the adjustment of the longitudinal profile after the release of clear water from impounding
reservoirs. If armouring of bed sediment is dominant, the slope
remains almost unchanged, such as the Colorado River below the
Parker Dam, or the Yongding River, or even becomes steeper,
such as the Colorado River below the Hoover Dam (1.6–42.3 km)
(see Figure 5.24). If armouring of bed material is not prominent,
the slope may remain unchanged or become flatter; the slope of
water surface at medium and low discharges of the Hanjiang
River below Danjiangkou Reservoir was 0.000 286 before dam
closure in 1960, and in 1978 it was 0.000 268, almost the same as
before.
30
30
1950
790
420
Post-dam
1956
1060
600
1957
1210
650
1958
1214
655
5.7.1.6 ADJUSTMENT OF CROSS-SECTIONAL SHAPE
Erosion in river channels may manifest itself in two ways, namely
degradation of the channel bed and channel widening. In various
rivers the development of erosion is different, depending on the
local conditions such as the basic characteristics of the river or
operational mode of the reservoir, etc.
The Yongding River below Guanting Reservoir is an
example of an increase in channel width. Table 5.25 shows the
change in channel width of the Yongding River, demonstrating the
drastic increase in channel width after the commissioning of
Guanting Reservoir.
Many rivers in the United Kingdom are examples of
another pattern of changes in channel width. Petts (1979)
analysed the variations of water depth and channel width of 14
rivers in the United Kingdom. The water depth and channel width
of most of these rivers remained unchanged near the dams;
further downstream in the meandering reach below the dams, the
water depth and channel width of two thirds of the rivers
remained unchanged, while the channel width of the rest of the
rivers decreased, and the water depth of two rivers decreased. In
general, the cross-section pattern of these rivers remained almost
unchanged, or became narrower and deeper. British rivers are not
long, their sediment particles are coarse, and river banks are
composed of either coarse particles or silty clay with good vegetation cover. After the commissioning of reservoirs, the reduced
flood discharges may be incapable of eroding river banks, leading
to the above-mentioned variations of the cross-sections. The river
channel cross-section below Danjiangkou Reservoir became
deeper and narrower in a similar manner to that of the British
rivers.
The variation of the channel cross-section below the
Sanmenxia Reservoir on the Yellow River during the period of
clear water release (1960–1964) was more complicated. Since the
bed material of the Yellow River is fine and the medium and low
discharges are comparatively large, channel degradation was
obvious in a 180 km reach; meanwhile, the collapse of flood
plains and increases in channel width were also significant in
some wide cross-sections. Since 1964, when the operational mode
changed from impoundment to flood detention and sediment
release, sediment deposition has been taking place in the channel,
leading to a serious collapse of flood plain banks and a widening
of the channel width.
For wandering rivers during the pre-dam period, a longterm balance between the loss and gain in flood plains prevails,
and the channel width thus remains almost constant. During the
post-dam period, however, such a balance is upset by the changed
flow and sediment regime. The collapse of flood plains increases
the channel width and makes the cross-section wider.
The loss of flood plains takes place mainly at the beginning of the new stage.
106
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
Q (m3 s–1)
Sediment (kg m–3)
rising leg of the flow discharge curve, the flood discharge is
reduced and the released sediment load is significantly reduced;
while during the falling leg of the discharge duration curve, a large
amount of sediment is released from the reservoir due to intensive
retrogressive erosion in the reservoir. Thus, the sediment peak lags
significantly behind the flood peak.
8
9
Figure 5.25 — Flow and sediment regimes of Guanting reservoir.
5.7.1.7 ADJUSTMENT OF CHANNEL PATTERN
Adjustment of the channel pattern below reservoirs is a long-term
phenomenon. Until now, no field data have shown obvious
evidence in this respect. However, laboratory tests show that the
general tendency of such an adjustment is a decrease in wandering
intensity and a gradual shift from wandering rivers to meandering
ones.
In the Hanjiang River below Danjiangkou Reservoir, the
wandering intensity in the braided-wandering reach has declined,
with some mid-bars combining and others connecting to the bank.
The river channel has become more regular than before. There had
been 26 river branches in a 240 km long reach just below the dam;
at present 15 branches have disappeared. The ratio of sinuosity has
increased from 1.25 to 1.50.
5.7.2
Fluvial processes below detention reservoirs
5.7.2.1 CHANGES IN FLOW AND SEDIMENT REGIMES
Non-match of the flow regime with the sediment regime is the
prominent phenomenon of the change in flow and sediment
regimes below detention reservoirs (see Figure 5.25). During the
Table 5.26
Elevation difference between channel bed and flood plains
Reach
Huayuankou
Jiahetan
Gaocun
Luokou
Natural
Impoundment
Detention
1.51
0.97
1.64
5.86
2.24
2.34
2.44
9.61
0.57
0.95
1.02
4.03
Units in m
Table 5.27
Change in bankfull discharges (m3 s–1)
Reach
Huayuankou
Jiahetan
Gaocun
Luokou
Natural
Impoundment
Detention
6 300
6 000
5 600
8 800
9 000
11 500
12 000
3 500
2 600
3 000
5 000
5.7.2.2 AGGRAVATION OF DEPOSITION BELOW DAMS
The annual amount of deposition of the Lower Yellow River
during the detention period of Sanmenxia Reservoir was 438
million tons, compared with 368 million tons under its natural
state. The elevation difference between the channel bed and the
flood plains dropped, as shown in Table 5.26, and therefore the
bankfull discharges of the main channel were also reduced, as
shown in Table 5.27.
5.8
CASE STUDIES
Six projects have been selected for case studies to show various
examples of reservoir sedimentation and related management
measures. The Liujiaxia Project is an example of how sediment
may be a factor in the selection of a dam site. The Sanmenxia
Project experienced several stages of reconstruction due to serious
reservoir sedimentation that had not been considered properly at
the original design stage. Through extensive studies, a new operational rule of impounding clear and discharging turbid waters (I
and D) was developed. It is useful in maintaining the long-term
storage capacity of projects built on sediment-laden rivers. The
Heisonglin Project, though much smaller than the Sanmenxia
Project, still faced similar sediment problems and solved these
problems with almost the same measures. These two case studies
show that I and D operational rules can be applied to hydrological
projects of different scales. The Shuicaozi Project was built on a
small river with a small amount of sediment load. However, sediment problems were serious due to inadequate facilities to exclude
sediment from the reservoir. After several measures were adopted,
including digging a new tunnel and dredging, the sediment problems were solved satisfactorily. The Guanting Project is an
example showing the effectiveness of various measures to reduce
sediment input in the reservoir. The Tarbela Project on the Indus
River is a key hydrological project in Pakistan. At the design
stage, no sediment management measures had been adopted. Since
the commissioning of the project in 1974, reservoir sedimentation
has emerged and become serious in recent years, as the Indus
Table 5.28
Characteristics of Liujiaxia and Sanmenxia Projects
Name of project
(109
m3)
Reservoir capacity
Length of reservoir (km)
Dam height (m)
Pool level fluctuation (m)
Catchment area (109 km2)
Annual runoff (109 m3)
Annual sediment load (109t)
Average concentration (kg m–3)
D50 suspended load (mm)
D50 top-set deposit (mm)
Installed capacity (MW)
Liujiaxia
Sanmenxia
5.74
56
147
41
181.8
26.3
0.087
3.31
0.025
0.02
1 225
35.4
106
688.4
45.3
1.6
35.6
0.038
0.02
900
CHAPTER 5 — RESERVIOIR SEDIMENTATION AND IMPACT ON RIVER PROCESSES
River carries a large amount of sediment. How to deal with sediment problems in the reservoir is still a pending issue for the
authorities. This case study shows how sediment management
should be considered for reservoirs, except for those built on clear
rivers.
Yanguo
xia Dam
Liujiaxia Dam
r
ive
eR
h
o
Ta
Daxi
a Riv
er
Yellow River
5.8.1
Liujiaxia Project
Liujiaxia Dam is the first large multipurpose hydrological project
on the Upper Yellow River for power generation, flood and ice jam
control, and irrigation. The first power came on line in 1969 and
the entire project was commissioned in 1974. Some pertinent data
are given in Table 5.28. The main dam is a concrete gravity dam.
One discharging tunnel (Qmax = 2200 m3 s–1) and two sluicing
tunnels (Qmax = 1 524 and 108 m3 s–1) are built in the dam.
Up until 1989, 1.41 billion m3 of sediment had been
deposited in the reservoir, accounting for 24.6 per cent of the original
storage capacity. Of this deposit, some 70 per cent was in inactive
storage, accounting for about 45 per cent of the original inactive
storage, and only about 8 per cent of the original active storage.
In flood seasons, incoming sediment was deposited first
in the gorge near the end of the reservoir (Figure 5.26). When the
pool level was drawn down during dry seasons, the deposits on the
top-set of the delta were eroded and transported, then deposited in
the inactive storage. The pivot point of the delta is still far away
from the dam. Thus, sedimentation in the main reservoir has not
caused any problems for the project so far.
Liujiaxia Reservoir has two small arms in the valleys of
the Taohe River and the Daxia River. The storage capacity of these
two tributaries is only 2 per cent and 4 per cent of the total storage
capacity, respectively. The Taohe River joins the main stream at a
point 1.5 km above the dam and carries 28.6 million tons of sediment per year, i.e. 31 per cent of the total sediment influx in
Liujiaxia Reservoir. The substantial and rapid deposition was
caused by the large amount of incoming sediment load in the relatively small Taohe River, and has led to serious problems in
Liujiaxia Reservoir.
The main problem is the formation of a mouth bar at the
confluence. By 1979, the inactive storage of the Taohe River was
full and the mouth bar had risen to the minimum pool level. The
proximity of the mouth bar resulted in a rapid increase in the
amount of sediment passing through the turbines. In June 1980,
when more flow was required to meet an abrupt increase in power
demand, the pool level in front of the dam suddenly dropped by a
large amplitude because the mouth bar impeded the flow of water
to the dam from the upstream part of the reservoir.
Figure 5.26 — Plan of Liujiaxia Reservoir.
107
Abrasion of turbine blades and the lining of the outlet
tunnels for sediment sluicing had been very serious problems. The
annual amount of sediment passing through power unit 2 reached
its peak in 1978 and 1979 with 11.6 and 11.9 million tons, respectively, when the top of the mouth bar was the highest. After
sediment sluicing in 1981,1984 and 1985, the amount was
reduced.
The abrasion of the turbine blades and the lining of
sluicing tunnels required a great amount of repair work. For
example, power unit 2 was damaged to such an extent that it had
to undergo repair for 125 days. It was found that the maximum
depth of abrasion was 50 mm, and the abraded area was as much
as 28.9 m2. Welding rod consumption was as high as 3.5 tons.
If the dam site of the Liujiaxia Project had been selected
above the confluence of the Yellow River and Taohe River, the
sediment problems experienced by the Liujiaxia Project would not
have been so serious at the initial stage of the operation of the
project. The decrease in the benefit of the project would not have
been large if remedial measures had been found to use the annual
runoff of the Taohe River (Qm = 178 m3 s–1).
At present, the major sediment problem of the Liujiaxia
Project is the existence of a mouth bar at the confluence of the
Taohe River. To lower the top surface of the mouth bar and to
reduce the amount of sediment passing through the turbines, drawdown flushing has been carried out four times. The effect of
sediment flushing was obviously positive and the top surface of the
mouth bar was lowered by 1.4 to 5.9 m. Sediment flushing also
recovered a certain amount of the storage capacity in the Taohe
River. The sediment flushing process was conducted at the end of
the dry season, when the pool level was close to the minimum. The
effect of sediment flushing depends on the schedule for the operation of the tunnels. This was decided in view of previous
experience, although it could also be decided through a model test.
5.8.2
Sanmenxia Project
The Sanmenxia Project was the first large multi-purpose water
conservancy project on the Yellow River, where the catchment
area accounts for 91.5 per cent of the total, and the runoff and
sediment load account for 89 per cent and almost 100 per cent of
the totals, respectively. The main characteristics of the project are
listed in Table 5.28.
The original planning of the Sanmenxia Project was
affected to a large degree by the opinion that a “large reservoir
storage capacity has to be gained by large inundation”. In 1958,
China decided to select 360 m as the normal pool level (NPL) in the
design phase, but at the first stage of construction 350 and 325 m
were adopted as the NPL and dead pool level (DPL), respectively;
the dam crest elevation was 353 m; the total storage capacity was
35.4 billion m3, of which 14.7 billiion m3 was reserved for the
sediment deposits; the installed capacity was 900MW. The main
objectives of the reservoir were to reduce the 1 000-year flood from
35 000 to 6 000 m3 s–1 and eliminate the flood threat in the Lower
Yellow River; to store all incoming sediment load and prevent
sediment deposits and bed levels from rising in the downstream
river channel; to manage the water resources of the Yellow River
and irrigate 1.48 million ha during the first stage and 5 million ha
during the second stage; and to improve navigation in the
downstream reaches. Accoding to this planning, the reservoir would
inundate 138 thousand ha of farmland, and 600 thousand people
would have to be resettled by the time the NPL was 350 m. The
108
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
Fenhe River
Beiluo River
Yellow River
er
Riv
ihe
e
W
Figure 5.27 — Plan of Sanmenxia Reservoir.
reservoir lifespan was expected to be 25 to 30 years. By combining
the reservoir with soil conservation works in the upstream reaches,
its lifespan could increase to between 50 and 70 years. It was
estimated that the sediment load in 1967 would decline by 50 per
cent thanks to soil conservation works and reservoirs on the
tributaries.
The general plan of the Sanmenxia Project and the reservoir storage capacity curve are shown in Figure 5.27.
The Sanmenxia Project was commissioned in September
1960. The operating rule in this period was to impound water and
trap incoming sediment load. The highest pool level was 332.58 m
(9 February 1961). During this period, 1.74 billion tons of sediment load entered the reservoir. However, only 7.1 per cent of the
total sediment load was vented out of the reservoir by density
current and 1.7 billion m3 of reservoir storage below 335 m were
occupied by sediment deposits. Tongguan is at the confluence of
the Yellow River and the Weihe River, which is the largest tributary of the Yellow River. The Yellow River has a maximum width
of 18 km at the confluence zone but contracts downstream to a
little over 1 km at Tongguan. The pass at Tongguan thus serves as
a local base level for the Weihe River and the Yellow River
upstream. The bed elevation at Tongguan had risen by 4.5 m from
September 1960 to March 1962. It induced a new problem,
namely the upstream extension of backwater deposits, which
would have very serious impacts on the Guanzhong Plain in the
Lower Weihe basin, a very important agricultural zone, and Xi’an
City, capital of Shaanxi Province. Such situations show that the
planning and design of the Sanmenxia Project in the 1950s were at
fault in the following respects. First, the project’s targets were too
high, such as the targets for power generation and navigation.
Second, much attention was paid to retaining sediment in the
reservoir to avoid aggradation in the Lower Yellow River, but the
impacts of reservoir sedimentation in the upstream area and reservoir area were neglected. Third, the opinion that “reservoir storage
capacity has to be gained by inundation” made the reservoir scale
too large, which was inconsistent with the national situation of
high population density and a shortage of farmland. Fourth, the
benefits of soil conservation were overestimated, since in 1967 the
incoming sediment load had been expected to decrease by 50 per
cent. Actually, the goal has not been reached.
In March 1962, it had to be decided to change the
operating rule from impoundment to flood detention and sediment
discharge in order to reduce the rapid sedimentation in the
reservoir.
In accordance with the operating rule of flood detention
and sediment discharging, just before the flood season the pool
level was drawn down to leave large reservoir storage for flood
control, and sediment was sluiced, with all sluicing gates fully
opened. As the capacity of the outlets was insufficient, two additional tunnels with the bottom level of 290 m at the left bank and
four power penstocks were converted into sluiceways. The
discharge capacity of the outlets was increased from 3 058 to
6 102 m3 s–1. The reconstruction works were initiated in 1965 and
gradually put into operation from June 1966 onwards. The trap
efficiency fell to 20 per cent. As the annual sediment load of the
Yellow River is so large, sediment deposition in the reservoir was
still too serious. In May 1969, it was decided that further reconstruction was needed, which included reopening eight diversion
bottom outlets at an elevation of 280 m and lowering the intakes
of five penstocks by 13 m, from an elevation of 300 m to 287 m.
The flow discharge capacity at a pool level of 315 m increased to
9 311 m 3 s –1 . The second stage of reconstruction started in
December 1969 and was completed in 1973.
Based on the lessons learned from the periods of
impoundment and flood detention, a new rational operating rule of
impounding the clear and discharging the turbid water was developed. During dry seasons the inflow with low sediment
concentration is impounded in the reservoir for spring irrigation
and power generation, and during flood seasons the reservoir pool
level is drawn down to sluice off most of the whole year’s sediment load, so as to keep a balance of deposition and erosion in the
reservoir in normal years and to reduce the aggradation under
favourable incoming flow and sediment conditions. Through the
proper regulation of flow and sediment, reservoir sedimentation in
Sanmenxia Reservoir has been controlled. The reservoir storage
capacities below 330 and 335 m have recovered to 3.1–3.2 billion
m3 and 5.9 billion m3, respectively. The bed elevation at Tongguan
has descended by 1.8 m. A narrow and deep channel and high
flood plains have been established in the reservoir, so that the
channel storage capacity can be preserved in the long term. The
trap efficiency has decreased to 0. The situation of the Sanmenxia
Project at various stages is shown in Table 5.29.
5.8.3
Heisonglin Project
The Heisonglin Project is a small hydraulic project on a small
river of Yeyu, China. The reservoir storage is 8.6 million m3,
controlling a catchment area of 370 km2. The dam is 45 m high
and a bottom outlet (2 × 1.5 m) with a discharge capacity of
10 m3 s–1 is installed at the dead pool level. The mean annual
Table 5.29
Situation of Sanmenxia Project
Stage
Time
Operating
rule
Discharge
capacity
(m3 s–1)
Trap
efficiency
(%)
1
September 1960 Impoundment
to March 1962
3 058
92.9
2
March 1962 to
October 1973
Flood
detention
6 102
20
3
October 1973
I and D
9 311
0
CHAPTER 5 — RESERVIOIR SEDIMENTATION AND IMPACT ON RIVER PROCESSES
Elevation (m)
runoff at the dam site is 14.2 million m3 (Qm = 0.45 m3 s–1) and
the mean annual sediment load is 0.70 million t. The mean annual
sediment concentration is 49.3 kg m–3, while the mean sediment
concentration in July and August is 113 kg m–3 and the maximum
concentration is 801 kg m–3. The suspended sediment is fine, with
D50 of 0.025 mm. The D50 of the original bed material is 18 mm.
The runoff in the flood season accounts for 45 per cent of the
whole year’s runoff, while the sediment load in the flood season is
98 per cent. The reservoir is a gorge-type reservoir.
The project was commissioned in 1959 and the adopted
operating rule was impoundment. During the first three years
(May 1959 to June 1962) reservoir sedimentation was very
serious, with a cumulative amount of deposition of 1.62 million
m3, namely 18.8 per cent of the total storage capacity. If such an
operating rule had been continued, the reservoir would have been
silted up in 16 years. Therefore, the operating rule of the reservoir
had to be changed in 1962.
In dry years sediment concentration is lower than
normal, so impoundment of water prevails even in the flood
season, when sediment may be vented out of the reservoir by
density currents. In wet years, discharging sediment prevails in the
flood season.
In the course of a year, at the early stage of the flood
season (1 to 20 July) when the flood peaks are often not high and
the sediment concentration is also not too high, density current
venting is the main method of discharging sediment. In the middle
of the flood season (21 July to 31 August), when floods frequently
take place with high sediment concentration, the pool level should
be drawn down to the flood control level (FCL) to facilitate
discharging sediment. In September, when sediment concentration
is not high, impoundment may start; when a flood occurs, density
current venting may be affected.
During the discharging of sediment, trap efficiency has
been as low as 10 per cent. Density current has formed easily in
Heisonglin Reservoir, and its trap efficiency is also low, at about
35 per cent.
Beginning in 1962, the overall trap efficiency of
Heisonglin Reservoir was 14.7 per cent. Consequently, the annual
rate of deposition in the reservoir slowed down to 0.1 million m3,
as compared with 0.54 million m3 in the first 3 years.
1. Pre-flushing (1980)
2. Post-flushing (1980)
3. Pre-flushing (1965)
4. Post-flushing (1965)
Distance (km)
Figure 5.28 — Longitudinal profiles before and after sediment
flushing, Shuicaozi Reservoir.
109
It should be emphasized that all the discharged sediment
from Heisonglin Reservoir was transported to an irrigation canal
downstream of the reservoir. The hyperconcentrated flow of
sediment contains organic manure and many nutrients, such as
nitrogen. The irrigated farmland has become more fertile, resulting
in an increased crop yield. Using the discharged hyperconcentrated
flow from the reservoir for warping not only mitigates serious
sedimentation in the reservoir, but also relieves deposition in the
channel downstream of the reservoir; it has three-fold benefits.
5.8.4
Shuicaozi Project
The Shuicaozi Project is located on the Yili River in Yunnan
Province, China. It is the second stage of four hydropower
stations, and functions as a seasonal storage reservoir and a diversion work conveying water from the Yili River to the Xiaojiang
River for power generation.
The dam is 36.9 m high. The NPL is 2 100 m with a
corresponding reservoir storage of 9.58 million m3, and the DPL
is 2 096 m with a 5.93 million m3 storage capacity. The effective
storage capacity is 3.65 million m3. The reservoir is 6 km long.
The top elevation of the spillway is 2 089 m, and the elevation of
the invert of the power station is 2 088 m. No bottom outlet was
installed in the dam. Sediment flushing was carried out through
the drawdown of the pool level through the spillway.
The mean annual discharge at the dam site is
16.3 m3 s–1, and the discharge for power generation is 2.9 m3 s–1.
The inflow is regulated by an upstream reservoir (Maojiacun
Reservoir). The incoming sediment load in Shuicaozi Reservoir is
mainly from the watershed between these two reservoirs. The
annual suspended load is 0.5 to 0.6 million tons, and the annual
bed load is 20 to 30 thousand tons.
The project was commissioned in 1958. From 1958 to
early 1981, 8.17 million m3 of sediment was deposited in the
reservoir, namely 85.3 per cent of the reservoir storage.
In order to recover a part of the storage capacity, drawdown flushing was carried out eight times during the period from
1964 to 1981. Figure 5.28 shows the longitudinal profile before
and after sediment flushing. A total of 1.22 million m3 of deposits
were flushed out of the reservoir. Owing to the high elevation of
the spillway, opportunities for drawing down the pool level were
limited. Flushing was divided into two stages. In the first stage,
from 1964 to 1966, when the upstream reservoir was not
impounded, flushing was carried out in the flood season and the
amount of flushing discharge was large. However, the top surface
of the deposits at the dam was still low, and the volume of sediment flushed out of the reservoir was small. In the second stage
(which started in 1974), flushing was only carried out for two to
three days during the Spring Festival when the power demand was
lower than normal. Since the top surface of the deposits was high
at the dam, the volume of sediment flushed out of the reservoir
was large, although the flushing discharge was smaller than during
the first stage. The quantity of sediment flushed out each time was
some 200 thousand m3, corresponding to about one third of the
annual incoming sediment load. To increase the quantity of the
sediment flushed out, a new tunnel for sediment flushing was
excavated in 1988. The intake of the tunnel is 22 m below the top
of the sediment deposits. The sluicing discharge is 50 to
170 m3 s–1. The maximum velocity in the tunnel is 18.2 m s–1, and
the total length of the tunnel is 325 m. After the completion of the
main part of the tunnel, more than 160 000 m3 of the reservoir’s
110
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
Table 5.30
Reduction of the rate of sedimentation in Guanting Reservoir
Period
Total amount of
deposition(106 m3)
Annual rate of
siltation (106 m3)
1956–1960
1961–1970
1971–1980
350
82
73
70
8.2
7.3
Table 5.31
Reduction of runoff and sediment load in Guanting Reservoir
Precipitation (mm)
Period
Annual
1951–1960
1961–1970
1971–1980
444
412
427
Flood
season
Annual
runoff
(106 m3)
338
313
373
1723
1258
832
Annual sediment
load (106t)
59.69
15.08
10.23
storage capacity was recovered when the inlet to the tunnel was
suddenly opened by blasting of the rock plug. A funnel, 482 m
long, 18.5 m deep and 100 m wide, was scoured out in front of the
intake. The sediment particles deposited near the dam were very
fine (D50 = 0.005 mm). The specific weight of the deposit was 1.1
to 1.3 t m3. Deposition was mainly caused by density current.
5.8.5
Guanting Reservoir
Guanting Reservoir is on the Yongding River in China, and
controls a catchment area of 43 400 km 2 . The mean annual
runoff at the dam site is 1.4 × 109 m3 and the annual sediment
load is 81 million tons. The reservoir storage is 2.27 × 109 m3.
The project was commissioned in 1955. The Yongding River is
heavily sediment-laden with an average sediment concentration
of 34.6 kg m–3. Reservoir sedimentation was so serious that by
1985, 612 million m3 of storage capacity had been silted up,
accounting for 27 per cent of the original storage capacity.
However, the siltation rates have been quite different in Guanting
Reservoir in different periods (Table 5.30). Although the average
annual precipitation and precipitation in the flood seasons of the
1950s, 1960s and 1970s were almost the same, the incoming
runoff and sediment loads in Guanting Reservoir have declined
significantly since 1960 under the influence of human activities,
as shown in Table 5.31.
It can be seen from Table 5.32 that sediment trapped in
the upstream reservoirs accounted for 41.5 per cent of the total
Table 5.32
Annual reduction of sediment load in Guanting Reservoir by
various measures
Trapping by
upstream
reservoirs
Irrigation
and
warping
Annual reduction of sediment
load (106 t)
17
19
5
41
Percentage of
total reduction
(%)
41.5
46.5
12.0
100
Causes of
reduction
Soil
Total
conservation
amount of reduction. Since 1958, 275 reservoirs with a total
storage of 1.4 billion m3 have been constructed. Until 1983, 0.34
billion m3 of sediment was deposited in 18 large and mediumsized reservoirs, of which the original total storage was
1.39 billion m3. The average annual amount of deposition was
17 million tons of sediment.
The largest reduction in sediment in Guanting Reservoir
resulted from irrigation and warping, with an annual reduction in
sediment of 19 million tons. There are 267 thousand ha of irrigated farmland upstream of Guanting Reservoir. Warping has been
applied to half of the irrigated land.
From 1950 to 1980, 6 200 km2 of eroded area in the
upper reaches of the Yongding River have been under control, i.e.
one fourth of the total eroded area. It was estimated that the
overall reduction in sediment yield amounted to 10 million tons.
However, in the meantime, the planting of astragalus membranaceous, a Chinese medicine herb, road construction, urban
development and mining led to an increase in soil erosion by
5 million tons. The net reduction by soil conservation measures
was therefore 5 million tons. From these data, it is evident that the
measures adopted to reduce sediment in Guanting Reservoir have
been effective.
5.8.6
Tarbela Dam Project
The Tarbela Dam Project is on the Indus River in Pakistan. The
catchment area of the Indus River is 969 000 km 2 , with an
annual runoff of 175 billion m3 and an annual sediment load of
470 million tons. Above the Tarbela Project the catchment area
is 169 579 km 2 , with an annual sediment load of 287 tons.
Although the catchment area above the Tarbela Project accounts
for only 17.5 per cent of the total catchment area of the Indus,
the annual sediment load above the Tarbela Project accounts for
66 per cent of the river’s total amount. The large amount of
sediment load carried by the Indus has been a great threat to the
Project due to serious sedimentation in the reservoir. Since the
dam’s first impounding in 1974, reservoir sedimentation at the
Tarbela Project has taken place rapidly. By 1990, 2.18 billion
m 3 of storage capacity had already been lost to deposition,
accounting for 15.2 per cent of the original reservoir capacity. A
delta has rapidly developed, with its pivot point at 1 300 to
1 310 ft elevation, which is close to the minimum pool level of
the reservoir. The rate of advancement of the delta was from 0
to 1 500 m per year, depending on the annual duration in which
the pool level was kept below 1 320 ft. If the present scheme of
operation continues, according to previous estimates the foreset
slope of the delta will reach the tunnel intakes in the period
between 2005 and 2008. Consequently, the present operation of
the Project will be impeded. Since the Tarbela Dam Project
plays an important role in the national economy (both for irrigation and power generation), its normal operation is vital for
Pakistan.
The problems induced by reservoir sedimentation in
Tarbela Reservoir above all include loss of storage capacity,
abrasion of turbines and hydraulic structures, and the danger of
blocked tunnels. No measures for sediment mitigation have been
planned. How to deal with the sediment problems in Tarbela
Reservoir is a question that is still under consideration. The
lesson is that for large hydraulic engineering projects, especially
those built on sediment-laden rivers, sediment mitigation
measures must be considered during the planning stage.
CHAPTER 5 — RESERVIOIR SEDIMENTATION AND IMPACT ON RIVER PROCESSES
5.9
MEASUREMENT OF EROSION AND
DEPOSITION IN THE RESERVOIR
Sedimentation surveys in reservoirs and river reaches are used to
determine the total quantity of erosion and/or deposition, as well
as the pattern and distribution of deposits. Such surveys are
usually made in order to modify the reservoir capacity curve and
provide data for studying the fluvial process upstream and downstream from a dam in response to the variability of flow caused by
various management measures in the river basin.
The range of surveys, frequency of measurements and
proper timing of a survey are determined by the requirements of
the research programme and according to the reservoir’s operational requirements for flood control and the multi-purpose
utilization of water resources. The range of reservoir sedimentation surveys should meet the requirements for a revision of the
reservoir capacity curve at normal high water levels and for an
evaluation of the upstream extension of reservoir deposits.
Repetitive surveys should be carried out whenever there is a
change in capacity exceeding ±3 to 5 per cent. The survey should
be conducted before or after the flood season and under relatively
stable flow conditions (Ministry of Water Resources, 1978).
Similar requirements are adopted for conducting surveys in river
reaches. However, repetitive surveys of the range lines set up at
reasonable densities are a cost effective and irreplaceable method
for studying sedimentation problems in a river reach. Xiong, et al.
(1983) made comparisons between the results of the amount of
sedimentation obtained by the range-line survey and that obtained
by the difference of sediment load observed at two terminal hydrometric stations, considering the input and output in the
intermediate areas. The study indicated that fairly good levels of
accuracy may be achieved with range lines set up at a reasonable
density. Also, the bias induced by the systematic error inherent in
the measurement of sediment discharge at hydrometric stations
would be too large if the amount of deposition or erosion was relatively small in comparison with the oncoming sediment load
(Xiong, et al., 1983; Lin, 1982).
Progress in surveying and mapping methods and in
instrumentation has been rather pronounced in recent decades.
Electronic distance meters such as microwaves, lasers and infrared
light devices are widely used. Aerial surveys together with underwater depth sounding are also commonly used. The Global
Positioning System (GPS) and Geographic Information System
(GIS) have caused a revolution in the field of surveying all over
the world. In this section, only the basic principles of conducting
sedimentation surveys will be discussed.
5.9.1
Methodology
Three methods are most commonly used to measure erosion and
deposition in reservoirs and river reaches, namely the range-line
method, the contour method (topographic survey), and the
composite method, which is a combination of the range-line and
contour methods. Selecting a method depends mainly on the
Table 5.33
Maximum allowable error in hydrographic surveying
Horizontal positioning
Depth measurement
Class 1
Class 2
Class 3
3m
± 0.5 ft
(15 cm)
6m
± 1.0 ft
(30 cm)
100 m
± 1.5 ft
(45 cm)
111
topography of the studied reach and the accuracy desired. Prior to
the advent of electronic measuring and computerized data collection and analysis systems, the range-line method was the preferred
method of collecting field data because it involved lower costs and
was less time-consuming. The development of current collection
systems has made the contour method the preferred method for
data collection and analysis.
Hydrographical surveys are recognized as either class 1,
2 or 3, depending on the level of accuracy required. Class 1 is the
highest accuracy standard and generally pertains to surveys in
support of site planning in advance of design efforts, pre- and
post-dredging activities, and other uses. Class 2 is a medium accuracy standard, and is generally used to determine channel
conditions in headwater and tributary arms, and in cross-section
surveys for reservoir volume computations. Class 3 is the lowest
accuracy standard, and is used principally for reconnaissance
investigations. The recommended maximum allowable errors for
each survey class are given in Table 5.33 (Ferrari and Dorough,
1996).
5.9.1.1 CONTOUR METHOD
A topographic survey covering the area of the whole studied reach
or only a portion thereof is a precise method, and is employed
when measuring the deposition or erosion in a reservoir or a river
reach, which is calculated from the difference in capacity at a
given elevation as measured from the topographic maps obtained
from two successive surveys. Surveys using the contour method
are employed as a control method for evaluating deposition in the
long term. The result provides a basis for the correction of the
capacities computed by the range method.
The scale of a topographic map for a reservoir or river
reach is determined by the desired accuracy of the computation of
erosion and deposition. For a medium-size reservoir or a short
river reach, a scale of 1:5 000 or 1:10 000 is preferred. For very
large reservoirs or long reaches, a scale of 1:10 000 or 1:25 000
should be used. If an accurate computation of the deposition is
required, the scale should not be less precise than 1:25 000.
In general, prior to impounding water in a reservoir
or conducting an experimental study of the fluvial processes
in a river reach, topographic surveys are conducted to provide
basic data for future studies. The topographic map is considered to be a fundamental map, which is revised periodically in
accordance with later repetitive surveys. Repetitive surveys
cover only an area over which a variation in land surface
takes place. The highest contour drawn in the repetitive
survey should, of course, coincide with the corresponding
contour on the original map, above which no change in landscape takes place.
The elevation of the highest contour measured in the
preliminary survey should be 4 to 5 m above normal high water
level or, preferably, above the possible maximum level reached at
design flood. The maximum probable range of bank erosion
should also be considered in deciding the range of the preliminary
survey. Once the map scale is properly determined, the whole
survey should be conducted according to the relevant specifications and standards.
5.9.1.2 RANGE-LINE METHOD
Relatively speaking, the range method is advantageous because
conducting the survey is simple and less time-consuming. If
112
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
ranges are arranged at reasonable intervals, the desired accuracy
can be obtained within the tolerance limit for allowable error. The
range method is a conventional method in general use for most
reservoir studies.
A sedimentation survey for a reservoir should extend at
least to some distance or several ranges upstream of the end of
backwater deposits. If the distance is large between the end of the
backwater deposits and the hydrometric station used as an inflow
sediment measuring station, a number of ranges should be set up
in such reaches. The river bed in this reach undergoes changes by
self-adjustment of the alluvial channel. From measurements
performed on these ranges, data may be obtained to verify the
water surface profile or to aid in the evaluation of sediment
balance. For a river reach, ranges should be arranged to cover
reasonably all the bends and transition regions, pools and riffles,
and wide and narrow parts, etc.
Ranges should be positioned approximately perpendicular to the major trend of the contour lines within which the
reservoir is operated. At a confluence, ranges should be set up in
large tributaries if the deposition in the tributary is estimated to be
appreciable.
The number of ranges considered reasonable implies a
minimum number of ranges established in a reservoir or river
reach which could reflect the essential pattern and distribution
of sedimentation, both longitudinally and transversely, without
being detrimental to the desired accuracy in the computation of
the total sedimentation. As a general rule, it is recommended
that the difference in the sedimentation computed by the rangeline method and by the contour method should be kept within a
limit of ± 5 per cent. Two methods may be used.
First method: On the preliminary topographic map with
a scale of 1:10 000, ranges spaced at an equal distance, for
example 200 m, are drawn approximately perpendicular to the
contours below the elevation of normal high water. Capacity or
volume at the normal high water level is calculated by the rangeline method and compared with the volume calculated from the
contour map. Computations are then made using fewer range lines
so as to select one out of two range lines, and then one out of three
range lines, etc. The simplification or reduction of range lines
should proceed until the relative error for the computation of
capacity or volume is still within the tolerance limit of ± 5 per
cent, using the volume computed by the contour method as a
reference.
Second method: Hakanson (1978) carried out studies
of the optimum arrangement of ranges in a lake survey.
Adopting his idea, using the data obtained from four large reservoirs in China, the optimum number of ranges may be computed
from the following equation (Sanmenxia Reservoir Experiment
Station, 1980).
A
Lr =
1
(5.45)
Lt F 3
where Lr is the distance between range lines at optimum density, A
represents the area enclosed by the highest contour line in km2, Lt
is the accumulative distance between ranges in km, and F = Lo/2
( π A) 1/2 , where L o is the length of the highest contour line
measured in km. Based on studies of reservoir data, it was found
that range line spacing according to the above equation will result
in surveys with a fair degree of accuracy. If the range intervals are
properly arranged, the accuracy of computing deposition by the
range-line method is within 5 per cent of that determined by the
contour method.
5.9.1.3 COMPOSITE METHOD
Contour or topographic and range-line methods may be combined
to gain a better understanding of the variations in ground surface
in a river reach or a reservoir. Bank failures usually take place at a
point not covered by pre-set range lines. The progress of delta
formation at the head of a reservoir may be studied by means of a
topographic map. Thus, a local topographic survey may be indispensable to supplement the sedimentation survey. In fact, there
will be essentially no difference between the results obtained by
the two methods if the number of ranges is increased sufficiently
so that contour lines or a topographic map can be drawn from the
data obtained from the range-line survey. This is particularly true
for bed surveys conducted by underwater soundings. Aerial
photographs in combination with underwater soundings taken in
the portion still covered by water may be used advantageously
while the reservoir is at its lowest level.
5.9.2
Instrumentation for positioning and depth sounding
5.9.2.1 DEPTH SOUNDING
Manual sounding poles, sounding weights and echo sounders are
commonly used for depth measurements. The appropriate
selection of instruments or devices depends on the local depth,
velocity, bed material composition and its degree of compaction. A
bell-shaped sounding weight made of cast aluminium (weighing 2
or 4 kg), or a sounding pole (aluminium sectional pole with each
section about 1.5 m in length, fastened together with threaded
dowels) may be used to take soundings where an echo sounder is
not available (Soil Conservation Service, 1973).
The type of echo sounder is selected mainly according
to its ability to distinguish the bed surface. The results of depth
measurements may differ with transducers of different power and
frequency response in the detection of the top of soft mud. Echo
sounders are usually specified by their relative accuracy. If a
depth 50 times the deposit thickness is measured by an echo
sounder with a relative accuracy of 1 per cent or more, a large
and intolerable error will be included in the sounding results.
Hence, a more precise instrument should be used. In general, an
echo sounder equipped with transducers operated at a low
frequency is preferred when measuring a river bed composed of
unconsolidated soft mud. A mini-echo sounder weighing only
several kilograms has been developed in the Bureau of Hydrology
of the Yellow River Conservancy Commission and has been
successfully used in both reconnaissance surveys and routine
work at hydrometric stations.
When an echo sounder is used for taking depth measurements in a reservoir or river reach, the instructions specified for
each instrument should be strictly obeyed. The transducer should
be properly installed on the bottom or the side of the measuring
boat. Water temperature should be measured and the instrument
adjusted accordingly. The depth recorded by an echo sounder
should be compared regularly with that measured by other reliable
fixed-point methods during the operation. The calibration can be
carried out by lowering an acoustic reflector, such as a flat metal
plate, to a known depth below the transducer, and adjusting the
instrument to produce an equivalent depth reading. Deviations in
depth recorded by the echo sounder can be used as a guide for any
necessary adjustment.
CHAPTER 5 — RESERVIOIR SEDIMENTATION AND IMPACT ON RIVER PROCESSES
5.9.2.2 POSITIONING OF SOUNDING POINTS
In a sedimentation survey carried out in small reservoirs or small
river reaches, for reasons of economy, expediency and accuracy,
the range-cable method of locating sounding points is most
commonly used. Equipment used by the United States Soil
Conservation Service includes an aluminium reel holding about
800 m of cable (galvanized aircraft cord with a diameter of
2.4 mm or plastic water-ski tow cable with a diameter of 6.3 mm),
equipped with a line meter (Soil Conservation Service, 1973).
In large reservoirs or broad river courses, sextants or
intersection by transits from two or three points are still used as a
traditional method of locating points. More advanced instruments
have been adopted, such as positioning by microwave, laser or
infrared electronic distance measuring instruments or systems.
The accuracy of a sedimentation survey, needless to say, relies on
the accurate positioning of measuring points, particularly in places
where the sediment deposition is not appreciable. Obviously, at
points where no deposition or erosion takes place, the elevation of
the bed surface should coincide with that measured in a previous
survey. This is a good check of the accuracy and reliability of the
sedimentation survey.
Aerial photographic techniques are most effective for
making base surveys before the reservoir is filled or at the start of
a research programme to study the fluvial processes in a river
reach. Capacity can be measured by photographic surveys, using
vertical air photography with permanent ground control points. All
such points should be coordinated and controlled in elevation to
the same degree of accuracy.
5.9.2.3 SURVEYING SYSTEM
To cope with the increasing demand for more complete and accurate
information on hydrological and geomorphologic processes in
reservoirs and river reaches, various types of surveying systems
have been developed and used. For instance, the Water Survey of
Canada had developed automated high-speed data collection and
processing systems (Durette, 1977) by 1973. In the early 1990s the
Changjiang Water Resources Commission (CWRC) developed
surveying systems, including software for communications between
shore stations and surveying vessels for navigation and data storage,
and processing methods for use with hardware such as electronic
theodolite, laser or microwave distance measuring devices and PCs
(Bureau of Hydrology 1990). Geomorphologic studies on river
reaches that undergo drastic changes during floods can be
conducted more comprehensively and accurately at a low cost.
5.9.2.4 POSITIONING BY THE GLOBAL POSITIONING SYSTEM
The Global Positioning System (GPS) is an all weather radiobased satellite navigation system that enable users accurately to
determine three-dimensional positions (x,y,z) worldwide.
Satellites are used as reference points for triangulating the position
of the receiver on earth. The position is calculated from the
distance measured using the time of transmission of the radio
signal. A minimum of four satellite observations is required to
mathematically solve the four unknown receiver parameters (latitude, longitude, altitude and time). A single GPS receiver is
usually not accurate enough for precise surveying and hydrographic positioning. Differential GPS (DGPS) is a collection
method to resolve the inherent errors of a single GPS receiver.
More than two receivers are used in DGPS, one of which is set up
at a known geographical benchmark. Differential GPS determines
113
the position of one receiver in reference to another and is a method
of increasing positioning accuracies by minimizing uncertainties.
It is not concerned with the absolute position of each unit, but with
the relative difference between the positions of the two units that
are simultaneously receiving signals from the same satellites. In a
sedimentation survey of Cascade Reservoir, the use of DGPS
made possible positioning accuracies of 1 to 2 m, which is acceptable in a hydrographic survey (Ferrari and Dorough, 1998).
DGPS interfaced with underwater depth sounding
systems in reservoir sedimentation surveys has also been used in
the reservoir topographic survey of Xiaolangdi Reservoir, as well
as in a river range-line survey in the Lower Yellow River. This
method has become inceasingly popular all over the world in
recent years.
5.9.2.5 MEASURING SEDIMENT THICKNESS
In cases where accurate maps of the original reservoir basin are
not available, the thickness of sediment deposits must be measured
directly to determine the original capacity and sediment volume. If
the water depth is not very deep, such as in small and mediumsized reservoirs, a spud or auger may be used. A sectional spud,
made up of 0.9 m (3ft) sections which can be assembled up to a
length of approximately 5.5 m (18 ft) with nickel-steel alloy dowel
pins has been used by the United States Soil Conservation Service.
The spuds are made of hardened case steel rods, 38 mm (1.5 in) in
diameter, into which encircling triangular grooves are machined at
intervals of 2.5 mm (0.1 in). The base of each groove is machined
to a depth of 3.2 mm (1/8 in) to form a cup in which sediment
deposits can be caught and held. The layer of new deposits can
generally be distinguished easily from the original bed material,
and through these means the thickness of sediment deposits can be
determined. In order to enhance accuracy, a combination of spud
and sounding is preferable if there is a thin layer of deposits (Soil
Conservation Service, 1973).
5.9.3
Measurement of bed material composition
If the intention is to measure the density and the particle size
distribution of the deposits, undisturbed sediment samples should
be taken at representative locations in the reservoir or river
reaches. The grain size, composition and dry density (unit weight)
of deposits are essential factors to be measured in a sedimentation
survey. Disturbed and/or undisturbed samples are obtained by
various means, and sent to the laboratory for further analysis. A
variety of equipment for taking samples was described by Vanoni,
et al. (1975). Sampling apparatuses for bed material including the
deposits in reservoirs or river reaches have been described in standards issued by ISO (1977c). If the size gradation of the bed
material as well as its unit weight are required, undisturbed
samples must be collected in the field. More often, only the
surface bed material is sampled for size analysis. In this case,
samplers similar to those used in river conditions, such as the
US-B54 or US-BMH-80 samplers developed in the United States,
or similar ones developed in other countries, may be used for
taking samples. For bed material composed mainly of coarse
materials, such as gravel and coarse sand on the flood plain or bars
of a river, various random methods may also be used (Tang, 1992).
5.9.3.1 UNDISTURBED SAMPLING
Included in the apparatus used for taking undisturbed samples,
ranging from simple to complicated equipment, are the ring-type,
114
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
axle-type, cylindrical revolving-type, gravity-core-type, piston
type, and vibration-type, etc. Each sampler has its particular range
of application. A ring-type device is usually made of stainless steel
pipe, 8 to 10 cm in height, with a sharpened knife-edge at one end.
An undisturbed sample is taken with the ring on the exposed
riverbed surface. The revolving cylindrical-type can be used in
unconsolidated soft deposits (Bajiazui Reservoir Experimental
Station, 1980). For shallow streams with fine bed material, the
US-BMH-53 sampler may be used. This sampler consists of a
stainless steel cutting cylinder, 5.1 cm in diameter and 20.3 cm in
length, with an internal retractable piston. For shallow streams
composed of a slightly compacted river bed of fine material, the
Phleger 840-A bottom corer may be used to take 3.5 cm core
samples (Durette, 1981). A gravity-core sampler can be used for
sampling in deep water such as in a reservoir (Vanoni, et al.,
1975). For detailed operating instructions, users of these devices
should refer to the relevant manuals or specifications.
The pit method is suitable for an exposed river bed or
flood plain. The procedure is to dig out a pit or hole of an appropriate size. The volume of the pit is measured by weighing the
amount of standard sand particles required to fill the pit and the
predetermined relationship between the volume and weight of the
standard sand. The unit weight of the deposit can then be
computed by weighing the sediment dug out of the pit. A cylindrical ring with a knife-edge is used frequently for sampling
deposits composed mainly of fine particles. The volume of the
cylindrical ring can be calculated by measuring its diameter and
height. After sealing the top and bottom of the sample, the sample
together with the ring can be sent to the laboratory for the determination of the unit weight as well as the moisture content
(Vanoni, et al., 1975).
5.9.3.2 RADIOISOTOPE DENSITY PROBE
Measurement of the unit weight of sediment deposits may be
carried out in situ with a radioisotope density probe, various types
of which are available. As with the nuclear gauge used in the
measurement of sediment concentration, the radioisotope probe
should be calibrated before its application in the field. The probe
can be lowered from a raft by a cable or it may be fastened to the
end of a drilling rod lowered along the outside pipe; by these
means, the unit weight of deposits in a lower layer can be
measured directly in situ (Vanoni, et al., 1975).
5.9.3.3 SELECTION OF SAMPLING POINTS
Samples of bed material or deposits are usually taken along the
range lines established for the sedimentation survey. The distance
between sampling points is usually set at random and is preferably
determined according to deposit thickness, although this may be
difficult to determine at the sampling time. The minimum number
of samples to be taken in a cross-section may be set at three for
€
main channel
widths of less than 500 m, and at five or more for
widths greater than 1 000 m. When the samples are taken on the
flood plain, the number of sampling points required depends on
the deposit width over the flood plain and the variation in bed
material sizes. Ordinarily, the sampling points are evenly distributed, or they can be distributed randomly.
As discussed in the previous section with reference to
the total sediment transport, the size and composition of the bed
material have an important influence on their transport and should
not be overlooked. The size and composition of an alluvial river
bed may change during floods, or it may change gradually whenever the oncoming flow condition varies. The armouring effect
due to the coarsening of bed material during erosion is an important aspect that deserves thorough research. Sampling bed material
provides valuable information on this subject and more samples
than suggested above should be taken to achieve a better understanding of the spatial distribution of bed material. The general
layout of sampling points can be arranged on a random basis if
there are no other particular requirements.
For river beds composed mainly of gravel or even largersized particles, the sampling work should be carried out more
carefully than on sand beds, in order to obtain representative
samples. Grid or transect sampling procedures may be selected for
surface sampling in armour effect studies, as well as studies into
the initiation of motion and flow resistance. Samples may be
collected by various samplers described elsewhere for subsurface
explorations that are related mainly to the study of bed material
transport (ISO, 1981b).
5.9.4
5.9.4.1
Data processing
COMPUTATION OF RESERVOIR CAPACITY OR AMOUNT OF
DEPOSITION OR EROSION IN RIVER REACHES
The frustum cone formula is a simple formula used generally for
this computation:
V=
(
)
1
A + A2 + A1 A2 ⋅ L
3 1
(5.46)
where V represents the volume or capacity occupied between two
sections or two contour lines, A1 and A2 are the areas of sediment
deposits or water at adjacent vertical sections or areas enclosed by
contour lines between which the volume is computed, V represents
the volume or capacity occupied between two cross-sections under
a pre-assigned elevation or between two contour lines, A1 and A2
are the areas of two adjacent cross-sections under the pre-assigned
elevation or areas enclosed by contour lines between which the
volume is computed, and L is the distance between cross-sections
or the interval between two contours. The difference of V for two
successive surveys in a reach gives the amount of deposition/
erosion in a reach.
Let ∆A denote the difference of area under a certain
elevation at the cross-section for two surveys. It is also the amount
of deposition or erosion expressed in the area at the cross-section.
If ∆A1 and ∆A2 do not differ by 40 per cent, the end area method
may also be used:
V=
1
( A + A2 ) ⋅ L
2 1
(5.47)
In order to obtain consistent data from the preliminary
and successive surveys, the results obtained by the range-line
method are correlated with those obtained by the topographic
survey. Correction factors are found for every specific reach or
portion of a reservoir and are applied to the results obtained by the
range method in later surveys.
Two examples are shown in Figure 5.29 to demonstrate
the characteristics of the reservoir capacity. The shaded area in
graph (a) represents the capacity between two elevations and that
in graph (b) represents the capacity between adjacent sections for
a specific elevation, sometimes taken as the normal high water
level. Computations, of course, can be carried out using graphical
methods (Vanoni, et al., 1975).
CHAPTER 5 — RESERVIOIR SEDIMENTATION AND IMPACT ON RIVER PROCESSES
115
Figure 5.29 — Examples of reservoir characteristics and the computation of reservoir capacity.
In the Lower Yellow River, repetitive range surveys have
been conducted for many years to monitor the sedimentation
process, and experiment stations were established to study fluvial
processes in some specific reaches. A software program
(RGTOOLS) is now being worked on to incorporate the functions
of management of the database, examination of the reasonableness
of the surveying data, computation of the reservoir capacity or
amount of sedimentation in river reaches, data processing and data
analysis (Liang, 1999).
5.9.4.2
COMPUTATION OF CAPACITY FROM TOPOGRAPHIC
SURVEYS
There are two approaches in computing the reservoir capacity
from topographic survey data: the point elevation-area method and
the conventional contour-area method. A computer program is
used for efficient and economically feasible data evaluation. In the
point elevation-area method, the surveyed area is large and has a
high data density, as shown in Table 5.34.
With the development of electronic measuring and
computerized collection and analysis systems, the contour method
of creating new reservoir topographic maps has become the
preferred method for reservoir sedimentation surveys. The United
States Bureau of Reclamation uses the ARC/INFO package to
develop reservoir topography from the collected data, the aerial
photographic survey data. ARC/INFO is a software package for
using the GIS. Contours for the reservoir area at selected elevation
intervals are computed from the compiled data using the TIN
(triangular irregular network) modelling package within
ARC/INFO. The Area-Capacity Computation Program is used to
Table 5.34
Number of point elevation data per km2 required in
topographic surveys
Quantity of point elevation data per km2
Detailed
survey
General
survey
Reconnaissance
survey
Rough bottom
2 500–3 500
1 500–2 500
800–1 500
Relatively
smooth bottom
1 500–2 500
800–1 500
400–800
Smooth bottom
800–1 500
400–800
100–400
Source: ISO, 1982b.
generate elevation versus capacity and/or surface areas for the
reservoir areas. The amount of deposition or erosion is the difference of capacities under the specified elevation computed for two
surveys (Ferrari and Dorough, 1996).
5.9.4.3 UNIT WEIGHT OF SEDIMENT DEPOSITS
The unit weight of sediment deposits should be obtained to convert
the deposit volume into weight. It is also an important parameter in
the study of sediment transport. The unit weight of sediment is defined
as the dry weight of sediment particles per unit volume of sediment
deposit. Methods for determining unit weight in situ or in laboratory
were discussed in section 5.5. In general, undisturbed samples are
obtained in the field and sent back to the laboratory for analysis. If it
is difficult to obtain an undisturbed sample in the field, a disturbed
sample may be taken instead and the unit weight may be estimated by
empirical formulae from size analysis data.
The initial unit weight may be obtained by the following
empirical procedure: divide the sample into size groups and weigh
each size group; mix each size group with water in separate calibrated vessels and wait until the particles settle; the deposit
volume may then be measured and the initial unit weight can be
computed. The result of an experiment conducted by Han, et al.,
(1981) is shown in Figure 5.30.
The determination of unit weight through size gradation
of deposits is suggested in literature. The initial unit weight in kg
m–3 can be computed as follows:
W = WcPc + WmPm + WsPs
(5.48)
where Wc, WM, Ws are the coefficients of unit weight for clay, silt
and sand, respectively, in kg m–3 and Pc, Pm, Ps are the percentages of clay, silt and sand, respectively. For different operation
modes of the reservoir, the coefficients are given in Table 5.35.
In determining the dry density of sediment deposits after
compaction, it is suggested that an additional value of unit weight
should be added to the initial value, as:
W = W0 + 0.4343k(
T
ln T − 1)
T −1
(5.49)
where T is in years and k is a constant based also on the type of
operation and size gradation of sediment similar to the expression
for the initial unit weight, as shown in Table 5.35.
116
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
Table 5.35
Unit weight as related to size gradation
Reservoir operation
kc
wm
km
ws
ks
416
561
641
961
256
135
0
1 120
1 140
1 150
1 170
91
29
0
1 550
1 550
1 550
1 550
0
0
0
Dry unit weight (t m–3)
Sediment submerged or nearly submerged
Normally moderate to considerably drawn-down
Reservoir normally empty
River bed sediment
wc (kg m–3)
Sediment size (mm)
Figure 5.30 — Variation of the unit weight with size
(after Han, et al., 1981).
REFERENCES
Bajiazui Reservoir Experimental Station, 1980: Cylindrical
Revolving Sampler for Taking Undisturbed Samples of Soft
Mud.
Beaumont, P., 1978: Man’s impact on river systems: a world-wide
review. Area, Volume 10.
Borland, W.M. and C.R. Miller, 1960: Distribution of sediment in
large reservoirs. Transactions, ASCE, Volume 125.
Brune, G.M., 1953: Trap efficiency of reservoirs. Transactions,
American Geophysical Union, Volume 34, Number 3.
Bureau of Hydrology, 1990: Development of River Surveying
Systems in the Changjiang River. Changjiang Water
Resources Commission.
Bureau of Reclamation, 1987: Design of Small Dams. Third
edition, Bureau of Reclamation, United States.
Churchill, M.A., 1947: Discussion of Analysis and Use of
Reservoir Sedimentation Data. (ed.) Gottschalk, Federal
Inter-Agency Sedimentation Conference.
Dai, Dingzhong, 1994: River Sedimentation Problems.
Chapter 12, Water resources development in China, (ed.)
Qian, Zhengying, China Water and Power Press, Beijing,
Central Board of Irrigation and Power, New Delhi.
Durette, Y.J., 1977: Hydac 100 — An automated system for hydrographic data acquisition and analysis. Technical Bulletin 105,
Water Survey of Canada.
Durette, Y.J., 1981: Preliminary Sediment Survey Equipment
Handbook. Water Survey of Canada.
Ferrari, R. and W. Dorough, 1996: Chapter 3-4 — Measuring
deposited materials, and Chapter 3-5 — Determination of
volume of deposits. International Conference on Reservoir
Sedimentation, Colorado State University, Fort Collins.
Ferrari, R. and W. Dorough, 1998: Cascade Reservoir 1995
Sedimentation Survey. Sedimentation and River Hydraulics
Group, Water Resources Services, Technical Service Center,
Denver.
Gottschalk, L.G., 1964: Reservoir sedimentation. Handbook of
Applied Hydrology, (ed.) V.T. Chow, McGraw-Hill.
Gu Wenshu, 1994: On the reduction of water and sediment yield
of the Yellow River in late years. International Journal of
Sediment Research, Volume 9, Number 1.
Hakanson, L., 1978: Optimization of underwater topographic
survey in lakes. Water Resources Research, Volume 14.
Han Qiwei, et al., 1981: Initial unit weight of reservoir deposits.
Journal of Sediment Research, Volume 1.
Han Qiwei, 1990: A new mathematical model for reservoir sedimentation and fluvial process. International Journal of
Sediment Research, Volume 5, Number 2.
Inland Waters Directorate, 1977b: SEDEX System Operations
Manual. Water Survey of Canada.
Institute of Water Conservancy and Hydroelectric Power Research
and Beijing Municipal Bureau of Water Conservancy, 1986:
Integrated Measures of Reducing Sedimentation in the
Guanting Reservoir (in Chinese).
ISO, 1977: ISO Standard 4364. Liquid flow measurement in open
channels — bed material sampling.
ISO, 1981b: Technical Report on Methods of Sampling and
Analysis of Gravel Bed Material. ISOFFXC I 13/SC6N 152.
ISO, 1982b: Draft standard ISO/DIS 6421, Methods for measurement of sediment accumulation in reservoirs.
Jiao Enze, 1980: Shapes of reservoir deposit. Selected Papers of
Yellow River Sediment Research, Number 4.
Lara, J.M. and E.L. Pemberton, 1965: Initial Unit Weight of
Deposited Sediments.
Liang Guoting, 1999: User’s Manual for RGTOOLS. Institute of
Hydraulic Research, YRCC.
Lin Binwen, et al., 1982: Major causes of the deviations in evaluation of quantity of sedimentation by range-line method and
difference of sediment load method. Journal of Sediment
Research, Beijing.
Lin Bingnan, 1992: Watershed and sediment management in
China. Proceedings of the Fifth International Symposium on
River Sedimentation, Karlsruhe.
Leeden, F. van der, et al., 1990: The Water Encyclopedia. Second
Edition, Lewis Publishers.
Long Yuqian and Li Songheng, 1995: Management of sediment in
the Sanmenxia Reservoir. Advances in Hydro-Science and
Engineering, Volume II, Beijing.
Luo Minsun, 1977: Reservoir Delta and its Calculation (in Chinese).
Miller, C.R., 1953: Determination of the Unit Weight of Sediment
for Use in Sediment Volume Computation. United States
Bureau of Reclamation.
Ministry of Water Resources, 1978: Tentative Standards for
Reservoir Sedimentation Survey. Beijing.
Petts, G.E., 1979: Complex response of river channel morphology
subsequent to reservoir construction. Progress in Physical
Geography, Volume 3, Number 3, pp. 329-362.
CHAPTER 5 — RESERVIOIR SEDIMENTATION AND IMPACT ON RIVER PROCESSES
Qian Ning, Zhang Ren and Zhou Zhide, 1987: Fluvial Processes.
Kexue Press, Beijing (in Chinese).
Qian Zhengying, 1994: Water Resources Development in China.
China Water and Power Press and Central Board of Irrigation
and Power.
Sanmenxia Reservoir Experiment Station, 1980: Optimistic
Density of Ranges for Sedimentation Survey.
Shaanxi Institute of Hydrotechnical Research and Tsinghua
University, 1979: Reservoir Sedimentation. Water Resources
and Electric Press, Beijing (in Chinese).
Shaanxi Provincial Bureau of Water Conservancy and Soil
Conservation, 1989: Techniques of Sediment Removal from
Reservoirs. Water Conservancy and Electric Power Press,
Beijing (in Chinese).
Shandong Provincial Office of Hydrology, 1980: Bed-load
Transport Estimated from Reservoir Sedimentation.
Shandong.
Soil Conservation Service, 1973: National Engineering
Handbook. Section 3, Sediment, USDA, Washington, D.C.
Stevens, J.S., 1936: The silt problem. Transactions, ASCE,
Volume 110.
Tang Yunnan, 1992: Study of Sampling Techniques of the Bed
Material on Gravel-bed Bars. Changjiang Water Resources
Commission.
Task Group of Sanmenxia Project, 1994: Proceedings of the
Operational Studies of Sanmenxia Project on the Yellow
River. Henan Renmin Press (in Chinese).
UNESCO, 1985: Methods of Computing Sedimentation in Lakes
and Reservoirs, Paris.
Vanoni, V.A., et al., 1975: Sedimentation Engineering. ASCE,
New York.
Williams, G.P. and M.G. Wolman, 1984: Downstream effects of dams
on alluvial rivers. Professional Paper 1286, USGS.
Working Group on Inventory of Reservoir Sedimentation in
Yellow River Basin, 1994: Report on Status of Reservoir
117
Sedimentation in Yellow River Basin. YRCC, Zhenzhou (in
Chinese).
Xia Maiding and Ren Zenghai, 1980: Methods of sluicing sediment from Heisonglin Reservoir and its utilization.
Proceedings of the International Symposium on River
Sedimentation, Beijing.
Xia Maiding, 1989: Lateral erosion — a storage recovery technique of silted-up reservoirs. Proceedings of the Fourth
International Symposium on River Sedimentation, China
Ocean Press.
Xia Zhenhuan, et al., 1980: The long-term capacity of a reservoir.
Proceedings of the International Symposium on River
Sedimentation, Beijing.
Xiong Guishu, et al., 1983: Analysis of errors in the sediment
measurement in the Lower Yellow river. Proceedings of the
Second International Symposium on River Sedimentation,
Nanjing.
Xu Mingquan, 1993: Strategy of reservoir sedimentation control
in China. International Journal of Sediment Research,
Volume 8, Number 2.
Yellow River Conservancy Commission (YRCC), 1993: Survey on
Reservoirs in the Yellow River Basin (in Chinese).
Zhang Hengzhou, 1983: Sediment-controlling problems in Yunnan
hydroelectric projects. Proceedings of the Second International Symposium on River Sedimentation, Nanjing.
Zhang Zhenqiu and Du Guohan, 1984: The rational operation of
drawdown flushing in the Shuicaozi Reservoir. Journal of
Sediment Research, Number 4 (in Chinese).
Zhou Zhide and Wu Deyi, 1991: Sedimentation Management of
the Tarbela Dam Project. International Research and
Training Centre on Erosion and Sedimentation.
Zhou Zhide and Yang Xiaoqing, 1995: Preservation of reservoir
storage capacity — experience of China. Proceedings of the
International Reservoir Sedimentation Workshop, San
Francisco.
CHAPTER 6
OPERATIONAL METHODS OF SEDIMENT MEASUREMENT
6.1
INTRODUCTION
6.1.1
Type of sediment load
Sediment load may be classified as suspended load or bed load
according to the mode of movement in the river. Suspended load is
the sediment that moves in suspension in water under the influence of turbulence. Bed load is the part of sediment load that
moves in almost continuous contact with the streambed by saltation and traction, that is, by bouncing, sliding and rolling on or
near the streambed by the force of water.
According to its origin, or source of supply, the total
amount of sediment transported in rivers may be divided into
two parts: wash load and bed material load. Wash load consists
of fine particles, which refers generally to sediment size finer
than 0.062 mm, and the amount depends mainly upon supply
from the source area. The discharge of bed material is
controlled by the transport capacity of the stream, which
depends upon bed composition and the relevant hydraulic parameters. Wash load moves entirely in suspension, while the bed
material load may move either as temporarily suspended load or
as bed load.
6.1.2
Network for measurement of sediment transport
Networks for stream gauging have been developed in many countries to collect data relevant to the development and protection of
water resources. In a sediment-laden river, sediment transport is an
important and significant item to be measured at a hydrometric
station within the framework of a stream gauging network.
Observations are made of suspended and bed load discharge in
streams with natural regimes as well as with regimes modified by
management activities. Stations that measure the sediment transport should function as components of the minimum stream flow
network.
For studying sediment problems in a river system, sedimentation surveys in river reaches, reservoirs and/or in estuarine
areas are also indispensable. Ranges or cross-sections spaced at
appropriate intervals are usually set up to serve also as a part of
the network for measuring sediment transport.
It is recommended in the Guide to Hydrological
Practices (WMO, 1994) that sediment discharge should be
measured at 15 to 30 per cent of stations within the minimum
network of stream gauging stations. The minimum network standard (expressed in area per station) is 1 000 to 2 500 km2 for flat
regions, 300 to 1 000 km2 for mountainous regions, and 140 to
300 km2 for small mountainous areas with very irregular precipitation. In arid regions or places where conditions are extremely
difficult, larger areas per station may be tolerated.
In general, the need to measure sediment discharge at a
hydrometric station, or to conduct a sedimentation survey in a
river reach, is determined by the importance of the sediment
problem in the development of water resources. It relates to a large
extent to the quantity of sediment transported in the river and the
temporal variation of this quantity. Measurements may be made
only in the flood season at some of the stations. For some river
reaches, experimental or auxiliary stations may be set up to carry
out detailed studies.
6.1.3
Classification of hydrometric stations for sediment
measurement
It is stipulated in the Chinese Standards for Sediment
Measurement in Rivers (Chinese Standards GB 50159-92,
1992) that basic sediment measuring stations should be
classified into three categories, as follows (items of sediment
measurement and accuracy requirements in different class stations
are different):
Class I: Stations that play an important role in controlling sediment yield from the drainage basin and are focal to the
design and operation of major hydrological projects and to river
regulation or in the study of fluvial processes are classified as
Class I stations in the stream gauging network. For Class I
stations, suspended sediment discharge and sediment concentration, size gradation of suspended sediment and bed material
should be measured the whole year round. Bed load should also be
measured at some of these stations by direct or indirect methods.
Class II: Stations at which the sediment yield is from
major tributaries that are representative in the physio-geographical
regions in the drainage basin, or stations that are supplementary to
a Class I station located on the main stem of the river, belong to
Class II. The accuracy requirement for taking measurements is
lower than that required for Class I stations. Particle size gradation
must be measured at some of these stations. Sometimes, measurements may be conducted on a roving basis.
Class III: Stations at which the sediment yield is from
ordinary or secondary tributaries are generally grouped into this
category. Stations that are representative of small watersheds with
a drainage area less than 300 to 500 km2 in arid regions or 100 to
200 km2 in wet regions also belong to this category. In Class III
stations, simplified methods of taking measurements may be used,
such that the sediment load in flood events is estimated with
acceptable accuracy. Measurements are frequently taken on a
roving basis.
This idea of classification of sediment measuring
stations is useful in the planning and implementation of stream
gauging networks.
6.1.4
Total load
The purpose of sediment measurements at hydrometric stations or
specific locations in a river is to monitor the total sediment load
flowing through the section. Ideally, total load is the summation of
the suspended load and the bed load, in view of the type of movement of the sediment. However, in practice, measurements cannot
be performed very well in zones very close to the river bed, where
the sediment concentration is the greatest. Sometimes, there may
be an overlap in the portion of depth covered by the bed load
sampling apparatus, and part of the suspended sediment may be
included in the sample collected by the bed load sampler.
Furthermore, sediment covering a large range of areas in different
types of movement can have different types of behavour from a
CHAPTER 6 — OPERATIONAL METHODS OF SEDIMENT MEASUREMENT
hydraulic point of view. The idea of total load should be kept in
mind as a basis for taking sediment measurements.
6.1.5
Sedimentation surveys
Erosion in upland watersheds produces sediment in river systems.
In the entire process of transportation from upland areas to the sea,
the sediment may be deposited in some reaches, or scoured from
the river bed at some other reaches. In managing the sediment
problems in a river, the fluvial process, including the status of
sedimentation, must be well known. As far as the amount of deposition or erosion in a river reach is concerned, it is usually far less
than the amount of sediment load transported through the river
system. In some cases, the amount of deposition or erosion in a
river reach has an order of magnitude equivalent to the tolerance
limit of errors involved in the sediment measurement at hydrometric stations. Therefore, sedimentation surveys have to be
conducted in the studied reach to provide more reliable and accurate data on the amount of sedimentation, rather than the
estimation made with data obtained through hydrometric stations
and some other reconnaissance investigations. Measuring techniques are explained in Chapter 5.
The geomorphologic data of a river may be obtained by
conducting a topographic survey, including a land survey and
underwater surveying, or by repetitive surveying on pre-determined ranges. Besides the surveying data, bed material must be
sampled and its size distribution analysed. The dry density or unit
weight should be determined with undisturbed samples that may
be collected occasionally. The data obtained through sedimentation surveys in river reaches are irreplaceable for understanding
the fluvial process and the status of erosion or sedimentation of
the river reach under study. They also provide basic data to study
the response of the river in its fluvial process to the modification
of incoming flow by human activities. A geomorphologic river
study provides a practical basis for the assessment, protection and
enhancement of the physical environment of the river system. A
practical guide on the application of the geomorphologic approach
to river management was provided by the Environment Agency of
the United Kingdom (Universities of Nottingham, Newcastle and
Southampton, 1998).
6.1.6
Parameters to be collected for a complete sediment
data set
For the collection of non-cohesive sediment data, guidelines were
issued by the American Society for Testing and Materials (ASTM
D5387-1997) describing the parameters that should be measured
or collected to obtain a complete sediment and hydraulic data set.
A complete data set should include the following parameters:
(a) Sediment parameters: sediment discharge or sediment
concentration of suspended load; bed load; size distributions
of suspended load, bed load, bed material and their specific
gravity;
(b) Hydraulic parameters: water discharge, velocity, width,
depth and slope, gauge height;
(c) Other parameters: temperature;
(d) Description of field conditions such as bed forms present at
time of data collection; methodology and instrumentation;
site description.
If bed load is not measured, or the sediment load in the
unsampled zone is to be evaluated, the data set can be used to
compute sediment transport using any prominently known and
119
verified transport formulae. With the data set, the total load may
be evaluated or estimated from the measured sediment load, for
instance by applying the Modified Einstein Procedure (Colby and
Hembree, 1955; Stevens, 1985).
6.2
6.2.1
6.2.1.1
MEASUREMENT OF SUSPENDED SEDIMENT
Method of measurement
MEASUREMENT OF SUSPENDED SEDIMENT DISCHARGE IN
A VERTICAL
Suspended sediment discharge over an entire cross-section is
usually measured by dividing the cross-section into a number of
sections. Sediment discharge passing through each section is
obtained by taking measurements along the vertical within the
portion of the section it represents. It has been shown by field data
that the vertical distribution of sediment concentration for various
size groups is quite different. Even for sizes finer than 0.062 mm a
gradient exists (Nordin, 1981). An example of the vertical distribution of sediment concentration is shown in Figure 6.1.
The conventional methods used to measure sediment
concentration in a vertical are sampling by point or depth integration and/or in situ measurements. The measuring method is closely
related to the instrument used for taking the samples. Both timeintegration samplers and instantaneous samplers are used for taking
samples. For a time-integration sampler, the nozzle of the sampler
used for point or depth integration should be isokinetic, or, in other
words, the velocity at the entrance of the nozzle should be equal or
very close to the ambient velocity. The same requirements are also
valid for some in situ measuring apparatuses. Some apparatuses,
such as nuclear gauges, ultra-sonic or vibration apparatuses, etc.,
have been used for the in situ measurement of sediment concentration. The measurement of sediment discharge at a point involves
collecting the accumulation of sediment in a specific period by
means of apparatuses such as the Neyrpic sampler or the Delft
bottle; these are integration samplers. By integration over a time
period, fluctuations of sediment concentration existing in natural
rivers may be minimized and temporal mean data can be obtained.
(1) Sampling by point integration in a vertical
The selection of measuring points in a vertical has been
proposed by standards or manuals issued by various countries.
The number of points can vary according to the depth of the river
and the size of sediment in suspension. In multi-point methods,
it is common to sample at five points, i.e. at relative depths 0,
0.2, 0.6, 0.8 and 1.0 (ratio of the depth of the sampler to the
Concentration g.1–1
Sediment concentration of whole sample
d <0.01
d <0.025
d <0.05
d <0.1
d = Sediment size (mm)
Figure 6.1 — Vertical distribution of sediment concentration for
several size fractions.
120
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
stream depth). In a frozen river, the bottom of the ice cover is
used instead of the surface of the flow. Accuracy also depends on
the grain size of the suspended sediment and the shape of the
distribution curve.
In practice, in the interest of lessening the work involved
in taking and processing samples during a flood event, samples
may be taken at fewer points, such as three points at relative
depths 0.2, 0.6 and 0.8, two points at relative depths 0.2 and 0.8,
or one point at relative depth 0.5 or 0.6. A composite sample may
be obtained by direct mixing according to a proportion of the
samples determined through experiments in the field. In other
words, the concentration at each point should be weighted against
the proportion of discharge it represents. Such methods should be
adopted only after their results are checked against measurements
obtained with multi-point or other more accurate methods.
The sediment discharge per unit width in each vertical is
determined either by graphical integration of the product of velocity and sediment concentration throughout the depth, or by
Equation 6.1.
qs =
d
m
n
∑k C V
(6.1)
i i i
i =1
where qs is the sediment discharge per unit width in kg s–1 m–1, m is
the number of measuring points, Ci is the sediment concentration at
the measuring point as determined in a field laboratory or directly
by in situ instruments in g 1–1 or kg m–3, Vi is the velocity at the
measuring point in m s–1, d is the depth in m, ki is the fraction of
depth each measurement represents, and n is the sum of the
weighting factors at a vertical distance. In Equation 6.1, fractions
of depth ki are considered as a weighting factor to be applied to the
Table 6.1
Value of factor ki
Measuring at relative depth
Number of
measuring points
in the vertical
5
3
2
1
n
0
0.2
10
3
2
1
1
3
1
1
0.5
0.6
3
1
1
0.8
1.0
2
1
1
1
or 1
products of velocity and sediment concentration. Values of factor ki,
as recommended in the Chinese Standards, are given in Table 6.1.
The average sediment concentration in a vertical can be
computed by dividing qs by q, the water discharge per unit width at
the vertical, which is obtained directly from discharge measurements.
It should be pointed out that the constant value of factor
ki is assigned to each measuring point purely by numerical integration. In practice, no sample can be taken exactly at relative
depth 1.0. Sampling at the bottom of the river bed is usually taken
within a varied relative depth ranging from 0.94 to 0.98 depending
on the structural design of the sampler, i.e. the lowest position of
the sampler relative to the river bed. The gradient of concentration
for coarse sediment is very large in the vicinity of the river bed.
Hence, there is an error induced from the computation of sediment
discharge by using the proposed factor ki. The error is a systematic
error in nature and should be minimal.
(2) Sampling by depth integration in a vertical
Depth integration is usually performed with depth integrating samplers. Water and sediment mixture can then be
sampled continuously while the sampler is moving at a constant
Table 6.2
Maximum transit rate ratios and depths for sampler bottle/nozzle configurations*
US sampler
DH-81
D-74
D-77
P-61
P-72
Nozzle size (mm)
Nozzle color
Container size (l)
Maximum depth (m)
Max.ratio Rt/Vm
3.17 (1/8 in)
4.76 (3/16 in)
6.35 (1/4 in)
7.93 (5/16 in)
3.17
4.76
6.35
7.93
3.17
4.76
6.35
3.17
4.76
6.35
6.35
7.93
4.76
4.76
4.76
4.76
White
White
White
White
White
White
White
White
Green
Green
Green
Green
Green
Green
White
White
Blue
Blue
Blue
Blue
0.4732 (1 pint)
0.4732
0.4732
0.4732
0.9464 (1 quart)
0.9464
0.9464
0.9464
0.4732
0.4732
0.4732
0.9464
0.9464
0.9464
3 Liter
3 Liter
0.4732
0.9464
0.4732
0.9464
4.57 (15 ft)
4.57
2.74 (9 ft)
1.83 (6 ft)
4.57
4.57
4.57
3.05 (10 ft)
4.57
4.57
2.74
4.57
4.57
4.57
4.57
4.57
54.86 (180 ft)
36.58 (120 ft)
21.95 (72 ft)
15.54 (51 ft)
0.2
0.4
0.4
0.4
0.1
0.2
0.4
0.4
0.2
0.4
0.4
0.1
0.2
0.4
0.1
0.2
0.4
0.2
0.4
0.2
* Quoted from ASTM D6326-98. Here, only some versions are listed, for illustration. The standard United States samplers are designated by D for depth integration, P for point
integration, DH for hand-held depth integration, and also, by the year in which the version was developed.
CHAPTER 6 — OPERATIONAL METHODS OF SEDIMENT MEASUREMENT
121
Table 6.3
Allowable errors in sampling method along verticals in a cross-section
Relative standard error (%)
Systematic error due to improper
sampling methods in a vertical (%)
Systematic error due to insufficient
numbers of verticals (%)
Sampling points
in a vertical
All sediment
All sediment
Bed material
±1.0
±1.5
±3.0
±2.0
Station class
I
II
III
6.0
8.0
10.0
Number of
verticals
2.0
3.0
5.0
±1.0
±1.5
±3.0
Depth (feet)
transit rate along the vertical. If the ratio of intake velocity to
ambient velocity is equal to 1, the volume of samples at each point
will be proportional to the local velocity. The sediment concentration of the sample taken by the depth integration method is the
discharge-weighted average concentration in the vertical.
Sampling may be carried out by round trips of lowering
and lifting or by just a single trip either from the surface to the
bottom or from the bottom to the surface. Electromagnetic devices
may be installed to open and close the intake. The transit rate of
lowering and lifting the sampler should not exceed four-tenths of
the mean velocity in the vertical and is also limited by the rate of
air compression in the sampling bottle. In order to obtain a representative sample, the container should not be filled entirely during
sampling.
In fact, the maximum transit rates are controlled by the
compression rate and the approach angle, and are functions of the
size of both the nozzle and the sampler container. It varies from
0.1 to 0.4 V m . For the United States, a series of isokinetic
suspended samplers, maximum transit rate ratios and depths for
sampler nozzle–container size configurations has been established.
It appears in Table 6.2 [ASTM D6326-1998].
As illustrated by Edwards and Glysson (1999), a series
of graphs used for determining the appropriate transit rate can be
constructed for various nozzle/container size combinations. As an
example, the graph as shown in Figure 6.2 is quoted. It was developed for a nozzle size of 3/16 in (4.76 mm) and a container size of
1 pint (0.473 l). For round trip depth integration, the transit rate
used in raising the sampler need not be the same as the one used in
lowering, but both rates must be kept constant.
Transit rate divided by mean velocity
Figure 6.2 — Example of transit rate determination using graph
developed for nozzle size 4.76 mm (3/16 in) and 1 pint sample
container (after Edwards and Glysson, 1999).
Bed material
±5.0
6.2.1.2
MEASUREMENT OF SEDIMENT DISCHARGE IN A
CROSS-SECTION
(1)
Selection of verticals based on the transverse distribution
of concentration. The number of verticals required for sediment
discharge measurements depends on the size distribution and
concentration distribution of the sediment, as well as on the
desired accuracy of data acquisition. Verticals should be spaced
closely in zones with large transverse variations in sediment
concentration and in the main currents. In measuring sediment
discharge, it is usual to measure the velocity simultaneously with
the sediment concentration. For new hydrometric stations, the
number of sampling verticals is usually approximately half those
along which velocities are measured. It is suggested in the Guide
to Hydrological Practices (WMO, 1994) that for taking a
discharge measurement, in general, the interval between any two
verticals should not be greater than one twentieth of the total
width, and that the discharge between any two sediment sampling
verticals should not be more than 5 per cent of the total discharge.
(2) Multi-point and multi-vertical method. These
methods are used to determine as accurately as possible the sediment concentration, size distribution and sediment discharge along
a vertical and across the entire section of a stream. They also
provide the basis for simplified measuring methods. Sampling by
these methods thus establishes the standard by which the adequacy
of measurements made by other less detailed schemes or methods
is judged.
According to the requirement suggested in the Chinese
Standard (Chinese Standard, 1992) quoted in Table 6.3, the accuracy of conventional methods of sediment measurement should be
assessed by conducting experiments at the station. If the error of
the method currently in use exceeds the tolerable limits specified
in the Standard, the method currently in use should be improved to
reduce the error. The error limits specified in the table are somewhat more restrictive than those used elsewhere.
For the middle or lower alluvial reaches of a river, flows
over flood plains often take place during floods. For a sedimentladen river, the distribution of water and sediment discharge in the
main channel and the flood plain should be investigated. The sediment distribution should be taken into account in the arrangement
of the verticals. To shorten the duration of sampling, simplified
methods and a lesser number of verticals are usually used for
measurements over the flood plain.
(3) Selection of verticals based on equal discharge
increment. In this method, verticals are arranged according to the
distribution of water discharge across the section. Each sampling
vertical represents approximately an equal portion of discharge.
The transit rate for each vertical may not be equal, but the sample
volume for each vertical should be kept approximately equal. For
round trip depth integration in a vertical, the transit rate during
122
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
descending and ascending should be the same. The method is
illustrated in Figure 6.3 and is known as the equal discharge increment, or EDI, method. It is suggested in the Guide to Hydrological
Practices (WMO, 1994) that three to ten equal sections of
discharge be selected. If the volumes of sediment-water mixtures
sampled at verticals are the same, a composite sample may be
obtained by mixing all the samples to yield a cross-sectional
average sample from which the average concentration, as well as
size gradation, can be determined by laboratory analysis. This
method is simple as regards sampling work and computation. The
discharge distribution across the section must be estimated prior to
the sampling work. If the main current shifts its position
frequently, or drastic scour or deposition takes place in the crosssection, sampling points representing equal portions of the
discharge should be promptly adjusted according to the variations.
This may be difficult during floods.
(4) Selection of equally spaced verticals. The channel
width at the water surface is divided into sections of equal width
corresponding to the number of verticals required. The Guide to
Hydrological Practices (WMO, 1994) suggests that the whole
width be divided into six to ten equal segments for taking depthintegrated samples. When the depth integration method is
employed, the transit rate of the sampler for all the verticals
should be kept the same, that is, established at the deepest and
fastest vertical in the cross-section. In round trip depth integration,
the descending and ascending transit rates should also be kept the
same. The same nozzle is used at all verticals. The sample bottle
should not be allowed to fill completely. Ideally, the sample
volume will be directly proportional to the water discharge represented by the vertical. The average concentration in the
cross-section will be the concentration of the composite sample
made up by combining all samples at the cross-section. This
method, known as the equal width increment (EWI) method, is
illustrated in Figure 6.3. It has an advantage over the EDI method
in that the distribution of flow in a measuring section is not needed
before sediment samples are taken.
Equal transit rate for
all verticals
Cross-sectional average
concentration obtained by
composite sample
Equal width increment (EWI)
Samples taken at
vertical through
centroid of areas of
equal discharge
Sample transit rate
adjusted so that equal
sample volumes are
taken at each vertical
Equal discharge increment (EDI)
Figure 6.3 — Sketch of methods for measuring sediment discharge
using the depth integration method.
(5) Simplified Method. During a flood, adequate
sampling using conventional methods may not be carried out due
to rapid changes in both discharge and sediment concentration.
Hence, there is a need to develop a sampling method of greater
ease and simplicity of operation to take samples to define the
temporal variation of concentration during the entire flood. Such a
simplified method is called an index-sampling method. In the
United States, it is sometimes referred to as the Box Sample.
Index samples should be taken at the same time that the
conventional method is being used. The concentration of the index
samples is correlated with the cross-sectional average concentration obtained by a conventional method. If the relationship is
stable, the ratio of cross-section concentration to index sample
concentration is plotted against discharge or stage and used to
convert the index sample concentration to the cross-sectional
average value. Various methods for collecting index samples have
been employed in rivers with different characteristics. Obviously,
some of these are rather complicated and may even be considered
as conventional methods. It has been shown from actual data that
the sediment distribution varies with flow conditions.
In small streams, structures already in existence or built
especially can be utilized to take sediment measurements by
installing sampling apparatuses or in situ instruments. Pumping
samplers of various designs, radioisotope gauges, turbidity meters
and depth-integration samplers have been used by many countries
such as Indonesia, Italy, the United Kingdom and the United
States, etc. (International Association of Hydrological Sciences
(IAHS), 1981). These devices can be used to monitor the variations in sediment concentration during flash floods. Samples taken
by such devices are equivalent to index samples.
In some cases, the sediment carried by a current is
mostly wash load and the distribution across the whole crosssection is fairly uniform. Samples taken at any point in a crosssection should be representative of the average value. However,
for large alluvial rivers the situation is more complicated. In
reaches where erosion and deposition may take place on a large
scale, an index sample taken at a fixed point in the river cannot be
expected to be representative, and the relationship of the concentration of the index sample to that of a cross-sectional average
sample will not be stable.
In summary, there are no definite and reliable rules for
the selection of measuring points for taking an index sample. It
serves as a supplement to the conventional method for measuring
suspended sediment discharge. By analysing data obtained by
precise and/or conventional methods, the sampling position may
be chosen, bearing in mind the desired accuracy. The following
are recommended:
(a) When the variation in sediment along the transverse direction is relatively large, three to five verticals arranged in an
equal discharge increment basis should be used for taking
index samples whenever possible. Depth integration is
preferred, but sampling at one to three points may be used;
(b) If the river bed in a wide river undergoes drastic changes
during floods, it would be impractical to determine the
centroid of equal portions of discharge accurately. Three to
five verticals covering roughly the deepest parts of the
section may be selected for taking samples. Depth integration is preferred, but sampling at one or two points in each
vertical may be used. Samples should be combined for laboratory analysis;
CHAPTER 6 — OPERATIONAL METHODS OF SEDIMENT MEASUREMENT
(c)
(d)
Three verticals at 1/6, 1/2 and 5/6 of the stream width for
mountain streams, or arranged by other appropriate divisions
such as one in the main current, and two on both sides, etc.,
are proposed by the Indian Standard (1966);
In a flood event, when more verticals are precluded from
use, one vertical may be used in order to shorten the duration
of the measurement. One vertical located near the main
current is sometimes used to represent the cross-sectional
average conditions. It can be seen from the example of field
data showing the transverse distribution curve of sediment
concentration in Figure 6.4 that there are verticals located on
either side of the main current at which the ratio of local
sediment concentration to the cross-sectional average
concentration equals unity. The exact position varies with
velocity and sediment concentration. However, if it varies
within a narrow range and is relatively stable, the vertical
may be used for taking the index sample with fair accuracy.
When the transverse distribution of sediment is fairly
uniform, the sampling position may be fixed at a point determined by analysing actual field data. During a flood event,
one vertical located near one bank is allowed only when the
pre-assigned position for sampling is inaccessible. Also,
sampling on the water surface can only be allowed if other
methods cannot be used under practical conditions. Results
should be corrected by analysing more detailed actual observational data.
6.2.1.3 SAMPLING FOR SIZE ANALYSIS
The purpose of sampling for size analysis is to provide information on temporal variations in grain size and to compute the
sediment discharge of each size group. The distribution of sediment size along a vertical and across a transverse section can also
be used to assess the accuracy and reliability of the measurement
of suspended sediment discharge.
A precise method for determining the size distribution
over a cross-section may also be simplified so that more samples
may be taken during a flood period. In general, samples used for
determining concentration are used for determining size gradation.
In selecting simplified methods, including methods for taking
€
Figure 6.4 — Transverse distribution of sediment concentration.
123
index samples, the representatives of size distribution of the
sample should be considered.
Along with velocity and channel shape, etc., sediment
size is a major factor influencing the non-uniform distribution of
sediment concentration across a section. If coarse particles, such
as those greater than 0.062 mm, constitute only a small fraction of
the total suspended sediment, the concentration obtained by a
simplified method may be representative of the total suspended
sediment, but not for coarse particles. Vertical and transverse
distribution of suspended sediment is affected by hydraulic
elements such as water depth, slope, etc., as well as sediment characteristics such as grain size. The exponent z in the expression of
sediment distribution in a vertical based on diffusion theory may
be used as an index (ISO, 1977b; Vanoni, et al., 1975):
z=
ω
κU∗
(6.2)
In Equation 6.2, ω is the average settling velocity for the
size group under study, κ is the Karman constant, and U* is the
friction velocity.
For sizes finer than 0.1 mm, settling velocity varies with
the square of particle diameter. Under the same hydraulic conditions, the value ω or z may differ one-hundred-fold for particles of
0.1 and 0.01 mm in diameter. Different patterns of sediment distribution are found for these two size groups: for 0.01 mm sediment,
the vertical and transverse distributions are rather uniform, while
for 0.1 mm sediment, large gradients exist in a vertical and across
the stream. Errors, which may be involved in adopting simplified
methods, should not be overlooked when coarse sediment particles
are present in appreciable amounts.
When selecting a measuring method, a compromise has
to be made between simplification and desired accuracy. In
general, the selection of measuring verticals or points for index
sampling has to take the characteristics of size distribution into
account if a better understanding of the sediment transport of
various size groups is desired. Errors which may be induced by
the simplified methods will be discussed in later sections.
Judging by the experiences gained from field observations on sediment-laden rivers, the variation in particle size with
time may be less than the variation in sediment concentration.
As far as sampling frequency is concerned, more sampling
should be carried out during floods in order to define clearly a
sediment hydrograph. If samples are expected to be representative of both concentration and grain size, a composite sample
taken on the basis of equal portions of discharge by combining
multi-point samples or depth integrated samples is recommended. Errors involved in simplified methods, such as an
index sample taken at one point in a vertical, would be too
large, particularly if the sediment load contains an appreciable
amount of coarse particles.
6.2.1.4 FREQUENCY AND TIMING OF SAMPLING
The desirable timing and frequency of sampling depends on the
runoff characteristics of the basin. For many streams, an average
of 70 to 90 per cent of the annual sediment load is carried down
the river during the flood season. Suspended sediment should be
sampled more frequently during the flood period than during low
flow periods. During floods, hourly or even more frequent
sampling may be required to define sediment concentration accurately. During the rest of the year sampling frequency can be
124
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
reduced to daily or even weekly sampling. For watersheds with a
wide variety of soil and geological conditions and an uneven
distribution of precipitation, sediment concentration in the stream
depends not only on the flood event in the year, but also on the
source of the runoff in the basin. Under such conditions, no definite sediment measurement schedule can be assigned. Besides, the
sampling of sediment concentration should be properly timed to
check the temporal variation in sediment. In general, the accuracy
needed from the sediment data determines how often a stream
should be sampled. The greater the required accuracy and the
more complicated the flow system, the more frequently it will be
necessary to take measurements.
6.2.2
Computation of sediment discharge
When point samples of suspended sediment are taken for each
vertical, the sediment discharge per unit width is obtained by
Equation 6.1. Sediment discharge of the entire cross-section can
then be computed by integration of the sediment discharge per
unit width along the entire width of the stream. In practice, this is
carried out by summing the products of the sediment discharge per
unit width and the section width each vertical represents.
If the sampling is conducted using the depth integration
method (either the EDI or EWI method), all samples are combined
into a single representative discharge-weighted sample. The sediment discharge in the entire cross-section is then computed as:
Qs = Cm Q
(6.3)
where Qs is the sediment discharge of the entire cross-section in
kg s–1, Q is the water discharge expressed in m3 s–1, and Cm represents the cross-sectional average concentration expressed in
kg m–3. If other units are used in expressing the parameters, a
coefficient must be applied.
Some types of instruments, such as the Delft bottle
sampler or the Neyrpic sampler, can only sample the accumulated
sediment passing into the sampler nozzle over a certain period of
time. Sediment discharge per unit area can then be computed by
dividing the weight of sediment accumulated by the sampling time
and also by the area of the intake nozzle and the efficiency of the
sampler (Jansen, et al., 1979).
The computation of average size gradation along a vertical or over the entire cross-section can be calculated by weighting
Table 6.4
Classification of suspended-sediment samplers
Classification
Operation
Basic
feature
Sample
volume (L)
Instantaneous
Point Sampling
Horizontal
0.5, 1.0 or 2.0
Pressure
adjusted by
chamber
0.47
0.5
Description
May be opened or closed by spring
dropping hammer or electro-magnetic
switch; operated by rod or suspended
by cable
US-D or US-DH series with nozzles
in three different sizes;
Bottle sampler with intake nozzle
pointing to the flow and air exhaust;
Depth
limitation
None
4.5 m round trip
4.5 m round trip
Depth integration
Pressure
adjusted by
collapsible
bag
Plastic nozzle exchangeable; while
used in deep water the volume of
sample may be increased by using
large plastic bags
Depends on
bag size
0.47 or
0.94
US-P series with nozzles in three
different sizes;
25–40 m,
max 55 m
1.0–2.5
JLC or AYX series with 4 mm nozzle
Pressure
adjusted by
collapsible
bag
1.0–3.0
Plastic nozzle exchangeable;
plastic bag
Nozzle exchangeable; rubber bag
specially made
Intake
velocity may
be adjusted
Practical
no limit
Pressure
adjustable by
opening or
closing value
Integration
Point integration
1.0–8.0
1.0–2.0
0.47
Accumulation
of sediment
Direct
accumulation
of sediment
Water flows
out while
sediment
retained
Vacuum chamber used for
adjusting pressure; may be
used near bed surface
Intake velocity adjusted by
varying pump speed; may be
used near bed surface
Single-stage sampler; used in
flashy streams
Delft bottle or Neyrpic-type
for measuring suspended bed
material; discharge correction
factor should be applied
Depends on
bag size
Depends on
bag size
None
None
CHAPTER 6 — OPERATIONAL METHODS OF SEDIMENT MEASUREMENT
the amount of sediment in each size class in each sample according to the flow rate represented by the sample. The sediment sizes
should be divided into groups to meet data analysis requirements.
In some countries, they are divided into three size groups, such as
sand (2.0 to 0.062 mm), silt (0.062 to 0.004 mm) and clay (finer
than 0.004 mm). If necessary, the number of size groups may be
increased. In India, suspended sediment coarser than 0.075 mm
(Indian Standard 6339, 1971) is classified as coarse. In the ISO
standards (ISO, 1982a), the division line between coarse and fine
sediment is set at 0.06 mm.
6.2.3
Measuring devices and instrumentation
6.2.3.1 SAMPLER FOR TAKING REPRESENTATIVE SAMPLES
Since 1947, a series of suspended sediment samplers, designed on
the basis of time integration and isokinetic nozzles, have been
developed through the Federal Interagency Sedimentation Project
(FIASP) in the United States. This series of standard samplers
includes samplers with different types of suspension, i.e. using
rods or cable reels; different container sizes, i.e. 1 pint (0.473 l) or
1 quart (0.946 l); and different construction materials, i.e.
aluminium, bronze or plastic. A set of exchangeable nozzles with
different sizes varying from 0.3 to 0.8 cm is available for most
samplers. In addition, epoxy-coated versions of all samplers are
available for collecting trace metal samples (Edwards and
Glysson, 1999).
Later on, various similar samplers were also developed
in other countries, however, they are not listed in the Table. Basic
types of samplers are classified in Table 6.4. Although they may
differ in structural design, type of suspension, sample volume, and
nozzle size, etc., they may be classified in one of the categories
listed in the Table 6.4.
Samplers are selected to meet data collection requirements in consideration of suitable measuring methods. More than
one type of sampling device is sometimes found at key hydrometric stations, to meet various flow conditions. A comparison of the
results obtained with different samplers should be made if consistent data are to be obtained. Sometimes, it may be necessary to
make small modifications to the sampler to cope with local river
conditions, without sacrificing their basic properties.
Samplers designed on the basis of time integration have
been widely adopted all over the world. Random errors due to
fluctuations may be eliminated to a certain degree, improving the
reliability of the results. During flash floods or the frozen season,
when abundant debris or ice floes exist in the flow, which may
block the intake nozzle of an integration-type sampler, instantaneous samplers may be used instead. Instantaneous samplers are
also used when sediment concentration is very high, because they
are simple and easy to operate; however, errors due to fluctuations
in velocity and sediment concentration are inevitable and should
be compensated by repetitive sampling.
6.2.3.2 BASIC REQUIREMENTS FOR AN IDEAL SAMPLER
The basic requirements for an ideal sampler may be summarized
as follows:
(1) The intake velocity of the nozzle for a time-integration
sampler should be equal or close to the ambient velocity. To
ensure sampling accuracy, it is better to calibrate the intake
velocity of the nozzle. It has been proven by experiment that
the error in the measurement of sediment concentration is
less than 5 per cent if the ratio of the intake velocity to
125
ambient velocity is kept within 0.8 to 1.2 (USGS, 1976). It is
specified in China that the ratio should be 0.9 to 1.1 at a
confidence level of 75 per cent in flows with a velocity less
than 5 m s –1 and a sediment concentration less than
30 kg m–3. For flows with very high sediment concentrations, the ratio would fall below the above range, however,
no appreciable differences in the observed sediment concentrations have been found;
(2) The sampler should be able to collect samples close to the
bed so that the unsampled zone can be kept as small as
possible. The distance from the centerline of the nozzle to
the bottom of the sampler should preferably be less than
15 cm. This figure is 10 to 12 cm for American series
samplers;
(3) Enough weight should be available for the sampler to maintain its stability under water. Ease of operation and
maintenance is essential;
(4) The sampling volume should be sufficient to fulfil minimum
requirements for determining concentration as well as size
gradation. Repetitive sampling may be necessary to fulfil the
minimum requirements for sample quantity.
In the design of a time-integration sampler, the intake
velocity is adjusted by pressure equalization in the sampler
container. Limitations as to the depth within which the adjustment
is effective should be strictly observed. For instance, present
American point-integration samplers can be operated to a depth of
16 to 37 m, with a maximum of 55 m, while the United States
depth-integration sampler series can be used to a flow depth of
less than 4.5 m for round trip operation (Edwards and Glysson,
1999).
6.2.3.3 SOME DEVELOPMENTS OF MECHANICAL DEVICES
Collapsible-bag samplers have been developed in the United
States and China. One version of the American bag sampler, with
a prefabricated plastic cap incorporated with an intake and air vent
nozzle, is attached to a plastic bottle in which a lightweight plastic
bag is inserted as a collapsible bag. With a newly developed solenoid valve, it can be used either as a depth-integration or
point-integration sampler (Stevens, et al., 1980; Szalona, 1982). A
new series of bag samplers that are more streamlined and have a
lower unsampled zone is now being developed in the United
States. The Chinese version uses a specially made rubber bag as
the collapsible bag. By intercomparison, it was found that they
perform similarly and the concentration of samples taken by both
samplers corresponded closely in flows with a velocity from 0.7 to
3.0 m s–1 and concentrations from 4 to 90 kg m–3 [Long and
Nordin, 1989]. This type of sampler apparently has potential for
use in the field.
Automatic pumping devices have been used in small
rivers, canals and experimental basin outlets, etc. One of the characteristics of this type of sampler is its ability to collect samples at
regular time intervals or in response to a rise or fall in stream flow
at a definite point in the river. The entire variation in sediment
concentration during a flash flood may be followed. Sufficient
samples can be obtained automatically to define the variations in
sediment concentration during a flood. It is particularly useful for
stations located in remote areas. However, all automatic pumping
systems are vulnerable to pipe blockages and may also require
efficient flushing systems. Different versions of the automatic
pumping sampler developed from 1969 to 1982 have been tested
126
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
and evaluated by FIASP. It was found that almost all of the types
were not isokinetic samplers, and improvements were needed to
overcome shortcomings (Edwards and Glysson, 1999).
Portable pumping samplers may be used for taking
point-integrated or depth-integrated samples at any point or vertical in a cross-section. A sampling nozzle may be mounted on the
streamlined sounding weight, together with velocity- or depthmeasuring devices such as a propeller meter or an echo sounder
transducer, etc. A device for measuring sediment and velocity
distributions in rivers and estuaries has been described by
Crikmore (1981). A pumping sampler with an attached filtering
device has also been developed and used in Pakistan.
6.2.3.4
SOME DEVELOPMENTS IN THE IN SITU MEASUREMENT OF
SEDIMENT CONCENTRATION
In situ monitoring of sediment concentration has been developed
and applied in some countries with promising results. The
measurement of sediment concentration by in situ nuclear gauges
has been carried out in some rivers in Italy, Hungary, Poland and
China. In general, the following features are common for various
types of radioisotope gauges (Berke and Rakoczi, 1981; Lu Zhi,
et al., 1981):
(1) Range of measurement: different ranges are specified for
gauges of various designs. The lowest detectable concentration within the tolerance of allowable error for hydrometric
measurements is in general 0.5 g 1–1. The maximum concentration may well exceed 1 000 g 1–1;
(2) Accuracy and reliability are ensured by calibration at certain
intervals of time or by comparison with traditional sampling
methods. The result of a field experiment indicates that the
lower the concentration, the greater the relative error;
(3) The measured zone for portable nuclear gauges may extend
to only 5 cm from the bed. In general, the unmeasured zone
extends 15 cm or more from the bed;
(4) Am241 or Cs137 is used as the source;
(5) A continuous record of the temporal variation in concentration may be achieved by installing the sensor at a definite
point in the cross-section. This is one of the advantages with
which none of the existing apparatuses can compare;
(6) Sampling still has to be performed for size analysis.
The development of the photoelectric turbidity meter is
based on the principle of attenuation of light transmitted through
sediment-laden water. From light scattering theory, the photodensity (the ratio of intensity of the transmitted light and incoming
light, I/Io) depends not only on the concentration but also on the
particle size existing in the medium. It would be possible to establish a relationship between the sediment concentration and a
photo-density reading only if the grain size were relatively
constant. In operation, the instrument must be calibrated carefully
to establish such a relationship. Determination of sediment
concentration on the basis of the photoelectric effect can only be
adopted in rivers where variation in grain size is very small and
the concentration is fairly low. The upper limit of application is 1
to 5 g l–1 (Brabben, 1981; Grobler, 1981).
There are two types of turbidity sensors based on light
scattering and absorptiometry (light attenuation). The former is
mainly of value for the lower end of the turbidity range below
0.5 g l–1, but can be relatively sensitive to variations in sediment
properties. The absorptiometric systems tend to extend further
up the turbidity range but are less sensitive at the lower concen-
tration end. Some works have used both systems in parallel
(Leeks, 1999).
A vibration device was developed at the Institute of
Hydraulic Research, YRCC. The apparatus has been installed at
Sanmenxia Hydropower Station for monitoring the sediment
concentration passing through turbine runners (Ma and Zhao,
1994). The Institute of Civil Engineering of the University of
Florence, Italy, has developed an optical ultrasonic device to
measure sediment concentration and mean particle size in the
field. By taking relative readings on two meters reflecting the
ultrasonic effect and the photoelectric effect, respectively, sediment concentration and particle size can be interpolated by graphs
obtained by calibration in the laboratory (Billi, et al., 1981).
A method has been developed based on the scattering of
ultrasound (4.4 MHz) from suspended sediment particles. By
measuring the frequency as well as the intensity of the Doppler
signal within a sediment suspension, both the velocity and the
sediment concentration can be measured simultaneously. It is
reported that the instrument has been successfully applied for
offshore measurements (Jansen, 1978).
For low sediment concentrations such as those found
under tidal conditions, a method is required which permits the
sampling and handling of a large volume of water (for example,
50 1) in order to obtain a reliable average value of the concentration. Delft Hydraulics Laboratory has developed a pumping
sampler that is interfaced with a device for the separation of water
and sediment using a filter method. Sample volume is determined
by means of a calibrated vessel. Comparisons with the acoustic
Doppler method in the field gave satisfactory results (van Rijn,
1980). As discussed in the previous section, an efficient flushing
system is required to prevent pipe blockages.
The new developments in the measurement of sediment
concentration cited in the above examples show promising results.
Needless to say, these instruments are still in the process of being
developed. More research work has to be done before they can be
adopted for use in routine work.
6.2.3.5 INTERCOMPARISON OF MEASURING DEVICES
To ensure accurate and comparable results, observations with
conventionally used sampling devices and/or in situ measuring
instruments should be compared for the standardization of sediment samplers. The need for a better understanding of the
probable error involved in sediment measurement further emphasizes the importance of the intercomparison of sediment
measuring devices. For time integration samplers, the hydraulic
efficiency of the nozzle should be checked prior to its adoption for
routine work, both in the laboratory and in the field. Sampling
efficiency may also be checked by comparison with a reference
nozzle that has a sampling efficiency of 100 per cent.
It is recommended that comparisons be made by means
of parallel sampling with traditional samplers and new samplers
before the latter are adopted. Attention must be paid to operational
techniques to avoid any systematic errors. When data are collected
for intercomparison, several samples should be collected and
analysed to minimize errors due to fluctuations. In situ measuring
devices have to be checked for deviations from the calibration
curve determined previously in the laboratory. It is suggested that
the results of parallel sampling (including measurements taken by
in situ apparatuses) should not deviate by ±5 per cent at the 75 per
cent confidence level.
CHAPTER 6 — OPERATIONAL METHODS OF SEDIMENT MEASUREMENT
An intercomparison of four point-integration samplers
was made jointly by the Delft Hydraulics Laboratory
(Netherlands) and the Cerni Institute (Yugoslavia) on the Danube
River near Belgrade in 1979. Velocity of the stream at the
sampling point was approximately 1.0 m s –1 . The sediment
concentration was 0.1 to 0.2 kg m–3 with a mean diameter of
0.2 mm (Dijkman and Milistic, 1982).
Through a WMO project, an intercomparison of
suspended sediment samplers has also been carried out at Chutuo
hydrometric station on the Changjiang River in China, in which
several point integration samplers developed by different Chinese
agencies were inter-compared and a USP61 type sampler was used
as a basis for comparison (Gao and Li, 1988). The sediment
concentration at the site is in general several kilograms per cubic
meter. Later, similar work was carried out at Tongguan Station in
the Yellow River, where the sediment concentration is much higher.
The results of the intercomparison are informative, not only with
regard to the results from sediment transport values, but also for the
characteristics and performance of the various samplers. It was
found that at low sediment concentrations, the performances of the
properly designed point integration samplers are similar and the
measured sediment concentrations are comparable. For concentrations of more than 30 kg m–3, the ratio of the intake velocity to
ambient velocity is less than 1. It appears that further studies are
needed on the performance of an integration-type sampler in
heavily sediment-laden flow (Gao and Li, 1988).
6.3
MEASUREMENT OF BED LOAD
Bed load movement is an important type of sediment transport in
rivers. The bed load, composed mainly of coarser particles, has
important effects on the fluvial process, even though its quantity
may be not as large as that of the suspended load. Bed load movement is quite uneven in both the transverse and longitudinal
direction and fluctuates considerably. In practice, it is more difficult to measure the bed load discharge accurately than it is to
measure suspended load. Research into the improvement of
sampling techniques is necessary.
6.3.1
Direct measurement of bed load discharge
The direct method measures the bed load discharge by taking
samples directly from the stream with a properly designed
sampler. Apparatuses or samplers used in the direct method may
be classified into the basket-type, pressure-difference-type, pantype and pit-type categories. The weight of the sample taken by
these samplers in a specific time interval represents the bed load
discharge over the width of the sampler.
The advantage of the direct method is that the
samplers are portable and are relatively easy to operate if proper
hoisting facilities are available. Temporal and spatial variations
may be observed, and the sampling work may be laborious and
time-consuming. Sampling efficiency should be obtained by
calibration in laboratory flumes and also in the field when the
bed load discharge can be determined by other reliable methods.
The efficiency of a sampler is defined as the ratio of the quantity of sediment trapped in a bed load sampler to that being
actually transported as bed load in the space occupied by the
sampler. Efficiency varies greatly from 10 to about 150 per cent
for different types of samplers (ISO, 1981a; Hubbell, 1964;
Xiang, 1980).
6.3.1.1 CHARACTERISTICS OF BED LOAD MOVEMENT
The factors affecting bed load transport are the hydraulic
conditions in a river reach (velocity, depth and width, slope,
size, shape and unit weight of bed composition, and morphology of bed forms, etc.) and the availability of sediment from
the source area. Measured data appear to be rather random in
nature, with large fluctuations under relatively stable hydraulic
and supply conditions. Figure 6.5 presents an example of the
variations in bed load discharge as measured in the field
(CWRC, 1980).
Generally speaking, the bed load discharge increases
very rapidly with increasing velocity. Consequently, the temporal
distribution of the bed load is characterized by its intensive transport during the flood season, particularly during several heavy
floods. For example, at Wutongqiao Station on the Changjiang
River in China, 60 per cent of the total bed load in 1965 was
carried down the river in just one day.
The spatial distribution of the bed load transport rate
over a cross-section is also not uniform. Heavy transport may
take place over only fractions of the bed width, while the transport rate outside these strips may be very small or seem to
approach zero. Although bed load transport is strongly influenced by local currents and the availability of bed materials, it is
quite common that the maximum velocity occurs within a strip
other than where the bed load transport is the highest. An
example of the transverse distribution of bed load transport for
sand and gravel measured at Yichang Station on the Changjiang
River is shown in Figure 6.6.
The variation in bed load transport rates along the river
course is also pronounced. Measured data in the East Fork River in
the United States reveal that there is an orderly progression in bed
Bed load discharge
Transport rate for gravel (kg s–1)
(g s–1 m–1)
Time
Figure 6.5 — Fluctuations in bed load discharge measured in the field
(CWRC, 1980).
127
Width (m)
Figure 6.6 — Transverse distribution of bed load transport rate
measured at Yichang Station, Changjiang River, China).
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
and medium flow conditions. On the Changjiang River, in China,
the number of measurements taken in a year to monitor the entire
process of bed load transport usually exceeds 100.
Discharge (m3 s–1)
128
Number of sampling verticals
Number of sampling verticals
Mean bed elevation (m)
6.3.1.2 FREQUENCY OF MEASUREMENTS
The frequency of measurements depends on the data requirements
for the computation of the total amount of bed load discharge for a
specific flood period. The measurement of bed load discharge over
an entire cross-section is laborious and time-consuming. In the
measurement of suspended sediment, simplified methods are
usually adopted for routine work. However, fluctuations observed
in bed load transport are far larger than those in suspended sediment. Simplified methods may induce appreciable error and
should not generally be used.
In general, the measurement of bed load discharge
should be planned to cover a large variation in water discharge.
The frequency of measurements should be much higher during
floods than in the low flow season. If bed load measurement
cannot be carried out satisfactorily during the rising limb of a
large flood, the bed load discharge may be extrapolated from the
discharge-to-bed load transport relationship established under low
6.3.1.3 SELECTION OF SAMPLING VERTICALS
Sampling verticals are chosen to check the transverse variation of
bed load movement. According to the experience gained in the
Changjiang River, sampling verticals should be in conformity with
the transverse distribution of the bed load transport, i.e. more
verticals, less than 15 m apart, are placed within the zone where
intensive bed load transport takes place, or any two adjacent verticals should cover no more than 15 per cent of the total bed load
transport, and three to five repetitive samples are taken in each
vertical. Only a few verticals are placed in the weak bed load
zone. The portion of the bed where intensive bed load movement
occurs should be identified by trial, prior to selecting the sampling
verticals (Huang, et al., 1983).
Experience gained in the East Fork River, in the United
States, has shown that the collection of about 40 individual bed
load transport rate measurements in a cross-section is, in most
cases, practical and economically feasible. Three different
methods have been used. In the first method, called the single
equal-width increment (SEWI) method, samples are collected at
each traverse in a round trip at 20 equally spaced intervals in the
cross-section. In the second method, called the multiple equalwidth increment (MEWI) method, samples are collected to and fro
at four or more evenly spaced verticals, taking one sample at each
vertical in one traverse until a total of 40 samples are collected. In
the third method, called the unequal width increment (UWI)
method, samples are taken at unequal space width increments until
a total of 40 samples are collected. It is clear that the SEWI
method is appropriate to define the transverse distribution in bed
load transport rate, whereas the MEWI and UWI methods are
more effective to define the temporal variations at each vertical
(Edwards and Glysson, 1999).
The duration of sampling, namely the time the sampler
is left on the river bed to take a sample, is limited by the transport
rate and the volume or capacity of the sampler. In general, the
quantity of a sample should not exceed two thirds of the effective
volume of the sampler. The experiment conducted in the
Changjiang River shows that the duration is preferably 3 to 5 min
for gravel, and 0.5 to 3 min for sand bed load.
Owing to the extreme variability of the bed load movement at different sites, at present it would be difficult to set up
Days from 1 May 1980
Figure 6.7 — Variations in bed elevations (East Fork River, United
States).
load transport rate from pool to riffle, reflecting the phenomenon of
the temporary storage of bed material (see Figure 6.7) (Emmett,
et al., 1981, 1983).
Figure 6.8 — Variation of sampling errors in bed load samples.
CHAPTER 6 — OPERATIONAL METHODS OF SEDIMENT MEASUREMENT
definite criteria in selecting the number of sampling verticals and
the number of required repetitions to be used at a particular site.
Some compromise must be made to achieve a balance between
the representation of both the spatial and temporal variations.
Experiments are encouraged at hydrometric stations to determine
the most appropriate sampling method to use for routine measurements. Probable relative error, which may be induced by an
insufficient number of verticals and repetitions, has been reported
by the Council for Mutual Economic Assistance (COMECON),
as quoted by Operational Hydrology Report No. 16 (WMO,
198l). Figure 6.8 is taken from that report.
6.3.2
Indirect method
6.3.2.1 SEDIMENTATION PROCESS
If bed load constitutes the major part of deposits in a reservoir, the
measurement of the deposit volume by repetitive surveys should
give an average bed load rate. In the evaluation of bed load, fine
material transported into the reservoir mainly as suspended sediment should be deducted from the total volume of deposits. The
unit weight of the deposits may be determined fairly accurately by
field measurements. Preferably, systematic suspended sediment
load data should be obtained at both inlet and outlet hydrometric
stations. The amount of bed load is then the amount of deposited
sediment, which is the difference between the amount of incoming
and outgoing sediment load. This indirect method of bed load
measurement gives only an average rate of bed load discharge in a
period between two successive surveys, rather than the instantaneous rate. If the bed load discharge is not very large, a long
period of time is necessary between repetitive surveys to obtain a
fair degree of accuracy (Shandong Provincial Office of Hydrology,
1980).
6.3.2.2 DUNE TRACKING
The dune tracking method of measuring bed load discharge
involves measuring the rate of bed material movement in duneshaped forms in the direction of flow. It is generally difficult to
measure the bed load in an alluvial river that consists mainly of
fine sands by means of existing measuring methods. The dune
tracking method has the advantage that only hydrographic surveying techniques are employed. With this method, a sounding system
should be established which permits the recording of bottom
profiles along pre-fixed courses in a river reach. Bed load rate can
be estimated from the propagation of dunes, calculated by successive surveys. The accuracy of the dune tracking methods relies on
the accurate determination of the bed elevation and positioning of
the measuring points (Havinga, 1981).
Two methods are used in monitoring the movement of
dunes:
(1) Moving boat technique: Longitudinal profiles are measured
repetitively by an echo sounder mounted on a boat. The
length of the traversed reach should be long enough to
include 20 to 25 well-defined dune forms. Usually, a straight
reach is selected for this purpose. Accurate records of time
and the boat position should be maintained. In the active bed
zone of the reach, five or more longitudinal profiles are
usually measured during each survey;
(2) Echo sounding: Continuous soundings taken at a fixed
point or several points in the flow cross-section monitor
the variation in depth and thus the movement of the
dunes. The time for taking such measurements should be
129
sufficient for at least 20 to 25 dunes to pass the point of
measurement.
6.3.2.3 TRACER METHOD
The tracer method, as well as the dilution method, is based on the
detection of the sediment movement by tracers. This method is
feasible for measuring bed material discharge and sediment
dispersion. However, there are large variations in the techniques
used. Selecting the appropriate technique depends on the study
purpose and the river conditions in the measuring reach. The
procedures and techniques involved are the selection and labelling
of the sediment tracer particles, the method of introducing the
tracer into the flow system, and the method of detection. Field
data collection includes tracing the labelled particles, sampling the
bed material and measuring hydraulic elements in the river reach
under investigation. The latter two are usually measured using
conventional methods.
Four labelling methods are available for use with the
tracer method. The fluorescent tracer, radioactive tracer and stable
isotope tracer can all be used in rivers where the bed material is
composed of relatively coarse particles such as gravel and sand.
However, only the radioactive tracer seems to be suitable for use
in places where the bed material is composed mainly of fine sand,
silt and clay. Fluorescent and stable isotope tracers have to be
detected in laboratories from samples taken in the field, but the
radioactive tracer can be detected in situ with a portable instrument. With the fluorescent tracer method, the movement of
radioactive tracers of different sized sediments can be measured
by dyeing them various colours to represent particles in different
size ranges. However, it is rather difficult to trace the movement of
radioactive tracer particles of different sizes. In contrast to a
radioactive tracer, stable isotope tracers have no environmental
impact since they do not involve radioactivity until the samples
taken from the field are neutron-activated in the laboratory.
Magnetic methodologies have also been used. The magnetic properties of sediment can be enhanced (by heating, inserting iron or
using electric coils). The particles are then traced using metal
detectors or specially designed detectors (Leeks, 1999). In all
cases, the labeled particles should have the same hydraulic behaviour after labelling as before and should resist leaching, abrasion
and decay of their traceability.
6.3.2.4
INVESTIGATION OF THE LITHOLOGIC PROPERTIES OF
SEDIMENT
Bed load sediment is originally composed of rock fragments
formed through weathering and wear during the transport
course over long distances. Lithologic properties vary with the
geological conditions of individual watersheds. If, for instance,
the bed load content of the tributary is known and the lithologic
composition differs distinctively from that of the main stem, the
lithologic composition of the bed load may be utilized as naturally labelled tracers in the estimation of bed load in the main
stem of the river.
In practice, the proposed method has been used to evaluate gravel bed load at Yichang in the Changjiang River. However,
this method is rather laborious and time-consuming, since a
tremendous amount of field and laboratory work has to be carried
out if a fair degree of accuracy is expected. The method can still
provide a feasible means of estimating bed load discharge when
other methods are impossible or too expensive.
130
6.3.3
6.3.3.1
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
Measuring devices
TECHNICAL REQUIREMENTS FOR AN IDEAL BED LOAD
SAMPLER
The technical requirements for an ideal sampler may be listed as
follows:
(1) The sampler should exert minimum disturbance on the flow,
especially in the vicinity of the sampler mouth;
(2) The sampler should have a moderately high sampling efficiency, for example, one exceeding 30 per cent, for different
sizes of bed load. The sampling efficiency should be calibrated;
(3) The sampler should have a simple design and be robust. A
portable version should be sufficiently heavy and easy to
operate;
(4) The size of the entrance should be adequate to cope with the
measurement of suspended sediment and also be at least 1.5
times the maximum size of the bed load;
(5) For pressure-difference samplers, the ratio of the intake
velocity to ambient velocity should be equal to or slightly
higher than 1.
Techniques involved in the measurement of bed load are
rather complicated. A commonly used bed load sampler may not
fulfil all the above requirements, but good results may still be
obtained if great care is taken in handling the sampler in the
appropriate manner.
6.3.3.2 VARIOUS KINDS OF BED LOAD SAMPLERS
The bed load measuring devices or samplers currently in use may
be classified into four types: basket-type, pressure-difference-type,
pan- or tray-type and slot- or pit-type. A variety of bed load
samplers have been developed. Here, only some samplers are
briefly described.
(1) Basket-type sampler. A basket-type sampler is
generally adopted for sampling coarse bed load material such as
gravel and pebbles. Metal or nylon mesh is put on the side and top
of a metal frame. Loosely woven iron rings or other elastic materials may be put at the bottom to deal with variations in bed surface.
The average sampling efficiency of a basket-type sampler calibrated in the laboratory is reportedly about 45 per cent, although
this may vary from 20 to 70 per cent (Hubbell, 1964). Experience
in China indicates that the sampling efficiency of this type of
sampler may still be much lower than this average value.
As an example, in order to take a direct measurement of
gravel bed load, several versions of basket type samplers were
developed and used in the Upper Changjiang River and some of its
major tributaries. The MB-2 sampler, weighing 700 kg, with an
opening size of 50 (height) × 70 cm (width), was used in the
Mingjiang River and Qingyijiang River. It may be used in mountain streams with a velocity under 6 m s–1 and a depth of less than
5.5 m, with bed load sizes of 5 to 500 mm. The Y80-2 sampler,
weighing 200 kg, with an opening size of 30 × 30 cm, and its
former versions have been used on the main stem of the Upper
Changjiang River under flow conditions with a depth under 30 m
and a velocity of less than 4 m s–1. The maximum size of the bed
load to be sampled is 250 mm.
(2) Pressure-difference-type sampler. The main feature of
a pressure-difference bed load sampler is that the ratio of the intake
velocity to ambient velocity and the hydraulic efficiency does not
differ much from 1. The pressure difference is obtained by enlarging
the flow section beyond the intake. Bed load sediment is collected
by the meshed bag at the rear or at the bottom of the flow section.
The following is a general description of several samplers.
(a) The Changjiang Y-78 bed load sampler. Several versions of
Y-78 samplers, i.e. type 78-1, weighing 50 kg (framework
not included), and type 78-2, weighing 14 kg, are available.
They may be used in streams with a velocity of less than
2.5 m s–1 and a depth of less than 10 m. Type 78-1 has an
opening of 10 × 10 cm and an effective capacity of 16 kg for
taking samples, while type 78-2 has an opening of 7 × 8 cm.
The main feature of this sampler is the position of its centre
of gravity, which is maintained in the front part of the
sampler by heavy lead strips and by a buoy in the rear part. A
protective plate in front of the sampler prevents unnecessary
settling and excessive scour around the entrance. The
hydraulic efficiency is close to 100 per cent. The sampling
efficiency is about 60 per cent. The sampler is suitable for
streams with bed material that is predominantly sand sized
(Zhou Dejia, et al., 1981);
(b) The BfG bed load sampler. This sampler was developed and
is used by the Federal Institute of Hydrology in Germany.
The intake nozzle is 8 (height) × 16 (width) cm. The collecting basket has a large capacity, allowing it to sample 6 kg of
bed load without affecting its efficiency. The inlet and the
collecting basket are connected by a flexible sleeve of reinforced plastic of 15 cm length (Federal Institute of
Hydrology, 1992);
(c) Helley-Smith (HS) bed load sampler. The intake section is
7.62 × 7.62 cm. The area ratio of nozzle exit to entrance area
of the original version is 3.22, and the bed load is caught in a
nylon bag with a mesh opening of 0.25 mm. The hydraulic
efficiency is 1.54. The overall sampling efficiency as calibrated in the field is close to 100 per cent (Emmett, 1979). A
laboratory study with varying bed materials and a range of
transport rates carried out by Hubbell in 1985 indicates that
the sampling efficiency varies with particle size, and that the
transport rate displays an approximate sampling efficiency of
150 per cent for sand and small gravel, and close to 100 per
cent for coarse gravel (Edwards and Glysson, 1999).
In order to sample larger sizes of bed load, a modified
version with the intake opening enlarged to 15.2 × 15.2 cm
has also been developed. The hydraulic efficiency is also
well over 100 per cent. However, its sampling efficiency is
about 100 per cent when used on medium to coarse sand and
gravel beds with a bed material size of 0.5 to 16 mm. It can
be used in streams with a velocity of less than 3 m s–1. The
Toutle River Sampler (TR2) is another modified version of
the Helley-Smith sampler, but it can be used to take samples
that are 2 to 150 mm in size. The intake section is 15.2 ×
30.4 cm. During intercomparison work carried out in China,
the sampler weight was increased to 230 kg by adding lead
pieces to the sampler. The apparatus was used in flow with a
velocity less of than 5 m s–1 (Xu, 1988);
After some modifications to the frame of the HelleySmith sampler, using a nozzle of the same size and an
expansion ratio of 1.4, a new version designated US-BL-84
was developed and adopted as a standard bed load sampler
by the United States Geological Survey (USGS);
(d) Slot- or pit-type sampler or sampling method. Emmett of
the USGS set up a bed load trapping system to collect bed
load sediment. Concrete troughs or trenches 0.4 × 0.6 m are
CHAPTER 6 — OPERATIONAL METHODS OF SEDIMENT MEASUREMENT
constructed across the river to a width of approximately
20 m. The slot is divided into eight sections fitted with
gates. Along the bottom of the concrete trough a rubber belt
0.3 m wide is threaded around drive and guidance pulleys,
and then returns overhead. Sediment falling into the open
slot is carried laterally to a sump in the riverbank. After
continuous sieving and weighing, the sediment is returned
to the river downstream of the trap by a conveyor belt. The
measurement of bed load using this type of installation is
reliable and accurate. However, it is adaptable mainly for
relatively small rivers and particularly for experimental
studies or the calibration of samplers (Emmett, 1979;
Leopold and Emmett, 1997);
Bed load traps made of metal plates or other suitable
material can be inserted into the riverbed to collect sediment
moving as bed load. The length of the trap along the flow
direction may be 100 to 200 times the grain size. Instead of
sampling at regular intervals, this type of sampling method is
used primarily to obtain the total amount of bed load in a
flood period, since it is not easy to remove or replace the
traps during floods. Bed load traps can also be used to study
the bed load transport in small experimental basins. A
caisson mechanical trap was developed at the Bureau of
Hydrology of Jiangxi Province, China. The top of the inner
container may be adjusted to make it even with the riverbed.
The height of deposition in the trap can be recorded and the
sampled material can be extracted from the trap using a
submerged slurry pump. Traps such as vortex tubes have
been used successfully in Nepal, China and other countries
to discard sediment moving in the vicinity of the bed of
canals or streams. These can also be used as a bed load
measurement device.
6.3.3.3 NEW DEVELOPMENTS
(1)
Intercomparison of bed load samplers. Bed load with
material of different sizes such as large gravel or fine sand has to be
sampled with different apparatuses such as a basket-type or
pressure-difference-type sampler. To study the behaviour of
different types of samplers, intercomparisons of bed load samplers
were carried out in the United States and China under a cooperative
study programme from 1986 to 1988. The samplers used for the
comparison included the basket sampler (types MB2, Y80) and the
pressure-difference sampler (types Y78-1, HS, TR2). The fieldwork
for the intercomparison was carried out in several rivers at sites with
gravel bed or sandy bed load. Although the range of the size with
which a specific sampler is applicable may be different, under the
same size range, the sampling result obtained by different samplers
is still comparable. Relative sampling efficiency can be obtained,
131
which reflects the behaviour of the samplers under comparison (Gao
and Xu, 1989).
(2)
Development of a new bed load sampler for gravel and
coarse particles. Several important ideas were deduced through
intercomparison work. For sampling on gravel-bed rivers, the
flexible bottom of a basket-type sampler may cope better with the
river bed, but the sampling efficiency may sometimes be too low
due to its low hydraulic efficiency. The high hydraulic efficiency
of pressure-difference samplers such as the Helley-Smith sampler
led to a high sampling efficiency, however, too much fine sediment
may sometimes be sampled due to the suction effect. Also, a
scouring effect may take place at the entrance due to the nonflexible bottom of the sampler. The different behaviour of these
two types of samplers was brought to light through the
aforementioned intercomparison work. A new version of bed load
sampler to be used mainly for bed load of gravel and pebble size
was therefore developed through a cooperative study by several
institutions in China. The new bed load sampler (Type AYT)
provides a flexible bottom at the entrance and an expansion section
to create a pressure difference. The sampler was designed,
manufactured and calibrated through extensive studies by
experiments in flumes with scale models in various sizes. The
hydraulic efficiency is 1.02. The sampling efficiency is a function
of bed load discharge (η = 48.5 Qs0.058, Qs in g s–1 m–1). Several
versions of this type of sampler are available, as shown in Table
6.5 (Gao, et al., 1995).
6.3.4
Calibration of samplers
A bed load sampler has to be calibrated for its sampling efficiency,
with which the measured transport rate can be converted.
6.3.4.1 DIRECT FIELD CALIBRATION
Efficiency is determined by directly comparing the result of
measurement obtained by the sampler under study with the bed
load measured directly by a more accurate and reliable method.
The transverse slot with a conveyor belt installed on the East Fork
River is a typical example of an accurate method for calibrating
the sampler and measuring the bed load (Leopold, 1997).
In practice, facilities similar to that installed at East Fork
River are not available for calibrating various types of bed load
samplers. It would appear that a carefully conducted intercomparison of bed load samplers in the field would be a feasible way of
obtaining their relative efficiency. If the efficiency of the sampler
serving as an index is known, then the efficiency of the sampler to
be compared may be determined.
If there is a highly turbulent section within the measuring reach, sediment that normally moves as bed load will be
suspended, and can be sampled by a suspended-sediment sampler.
Table 6.5
Basic versions of AYT sampler series used for coarse bed load
Dimension (mm)
Effective
capacity
Serial No.
Entrance
1
2
3
Range of application
Weight
Overall
Particle size
Depth
Velocity
Width
Height
Length
Max. height
kg
kg
mm
m
m s–1
120
300
450
96
240
360
760
1 900
2 850
176
438
657
10
60
180
40
320
600
2–100
2–250
2–400
40
40
30
4.0
4.5
5.0
132
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
A good estimate of bed material discharge within the unmeasured
zone, including the bed load, may be obtained by measuring the
difference of the sediment discharge at both the turbulent and the
normal sections using standard suspended-sediment sampling
techniques. This principle has been used by the USGS to evaluate
the total sediment discharge in a turbulence flume built in a
natural stream (Vanoni, et al., 1975).
6.3.4.2 LABORATORY CALIBRATION
Bed load samplers can be constructed to scale and tested in laboratory flumes. However, the sampling efficiency obtained by
flume experiments in a laboratory with a model sampler is usually
larger than the true efficiency. Large differences were observed in
the experiments carried out by CWRC in using model samplers
with a different scale ratio. It was found that the efficiency of a
sampler is not constant but varies with the flow parameters, transport rate, particle size and local bed conditions. For instance, the
efficiency of a basket sampler with a flexible bottom for sampling
gravel is very small at the moment when the gravel just starts to
move. The efficiency becomes greater as the transport rate
increases. For a pressure-difference sampler the efficiency changes
with the flow velocity.
The calibration of samplers on a reduced scale will lead
to scale effects; it is therefore advisable to test the full-scale
instrument. Hubbell of the USGS has reported recent refinements
in calibrating bed load samplers. Calibration curves, rather than
efficiency percentages, were derived by two independent methods
using data collected with prototype versions of the Helley-Smith
bed load sampler. The tests were conducted in a large calibration
flume capable of continuously measuring transport rates across its
width. The flume was 2.7 m wide, 1.8 m deep and 83 m long, with
a discharge as large as 8.5 m 3 s –1 . An adjustable width slot
extended across the full width of the channel, dividing it into
seven lateral sections. The facility was designed to re-circulate bed
load particles ranging in size from 2.75 mm at rates up to 12 to
20 kg s–1. Apparently, with this type of facility the results obtained
by laboratory calibration should be much more reliable than the
earlier laboratory calibrations using scale models (Hubbell, 1981;
Druffel, et al., 1976).
As discussed in previous sections, different types of bed
load samplers are designed for different bed conditions. Sampling
efficiency is different for different types of samplers. At present, a
sampling efficiency of over 50 to 60 per cent should be considered
to be satisfactory for sand and gravel. In any case, the sampling
efficiency of the bed load sampler should be determined by calibration in the field and also studied in the laboratory to correctly
interpret the measured data.
6.3.5
Computation of bed load discharge
Bed load discharge per unit width measured at each vertical may
be computed from the following equation:
qsb = 100 k Wb / (η b t)
(6.4)
where qsb denotes the bed load discharge per unit width after
modification according to the sampling efficiency η expressed in
%, Wb is the weight of the sample collected in a period of time t
(sampling duration), b represents the width of the sampler inlet,
and k is a coefficient inserted for the conversion of units expressed
for various parameters.
The total bed load discharge over the entire cross-section
can then be computed by numerical integration along the stream
width. This is done either by a graphical or analytical method. In
the graphical method, the bed load discharge is plotted as the ordinate, and the horizontal distance along the entire width is plotted
as the abscissa. In the graph, the distribution of mean velocities is
also plotted to make a visual inspection of the reasonableness of
the measured results. The analytical method involves the computation of bed load discharge by a trapezoidal formula, assuming a
linear variation between two adjacent verticals.
The analytical method of computing the bed load
discharge over a cross-section may be illustrated by Figure 6.6. It
is called the mid-section method and is expressed as (Edwards and
Glysson, 1999):
QB = qb1b1/2 + Σ qbi [(bi + bi+1)/2] + qbi+1 + bi+1/2
(6.5)
The results obtained for each individual bed load
measurement can also be related to some hydraulic parameter
such as the discharge, or the stream power, during the period of
measurement. This relationship, together with the rating curve at
the same site, can provide a necessary tool in the further computation of the total bed load. In flood events, it is difficult to take a
representative bed load sample. In this case, the bed load
discharge may be extrapolated through relationships between the
bed load discharge and relevant hydraulic parameters which are
established with measured data obtained in other periods.
6.4
MEASUREMENT OF TOTAL SEDIMENT
DISCHARGE
6.4.1
Direct methods
There are three types of direct methods for evaluating the total
sediment discharge, i.e. measurement of suspended and bed load
discharge at a specific cross-section; measurement of sediment
accumulation in reservoirs; and measurement by turbulence flume.
6.4.1.1
MEASUREMENT OF SUSPENDED SEDIMENT AND BED LOAD
DISCHARGE
The most direct and intuitive method is to take separate measurements of the suspended and bed load discharge simultaneously at
the cross-section. However, the total sediment discharge is not the
simple summation of the measured suspended sediment discharge
and bed load discharge. The reason for this is that there is an
unmeasured zone when the suspended sediment is measured using
the depth integration method. The lowest sampling point can only
be set at 0.94 to 0.98 relative depth, and some suspended bed
material load in the vicinity of the bed may not be collected by the
sampler. In sampling using the point method, some errors may be
induced by using the weighting factors in numerical integration.
Sometimes, there may be an overlap in the portion of depth
covered by the bed load sampling apparatus, i.e. a part of the
suspended load may be included in the sample taken by the bed
load sampler.
Under present technical conditions, the measurement of
bed load is both time-consuming and labour-intensive. In the
lower reaches of an alluvial river, bed load usually constitutes only
a relatively small portion of the total load. Therefore, only a few
stations attempt to take both suspended load and bed load
measurements as routine work. Nevertheless, in evaluating the
total sediment discharge, the transport in the so-called unmeasured
CHAPTER 6 — OPERATIONAL METHODS OF SEDIMENT MEASUREMENT
zone must be accounted for. Direct measurement of the bed load is
encouraged and should be carried out whenever possible. For river
management, it is necessary to have knowledge of the relative
response of bed load and suspended load since they represent
different management problems. In such cases, the direct measurement of bed load is indispensable.
Installations and devices similar to those used by the
United States Geological Survey on the East Fork River,
supplemented by regular suspended sediment sampling over the
section, can satisfy the measurement requirements. However, they
require careful design and may be too expensive to operate on a
routine basis. Structures or weirs across a small river to
concentrate the sediment-laden flow have been constructed in
some experimental basins in Italy. At the bottom of the weir, a
vortex tube is built to collect bed load materials. Devices such as
automatic-pumping samplers or other types of samplers may be
used for suspended-load sampling (IAHS, 1981). Needless to say,
such measuring installations can only be constructed on relatively
small rivers. They would be impractical for normal sediment
measurement networks in large or medium-sized rivers. However,
for some experimental reaches where the measurement of the total
load is significant, they provide an effective method worth
adopting.
6.4.1.2 MEASUREMENT BY MEANS OF TURBULENCE FLUME
This method can be used at certain narrow constrictions in
sandbed streams with sections so turbulent that nearly all sediment
particles moving through the reach are in suspension. The turbulence flume was so named in literature because artificial
roughness elements were put on the floor of the flume to produce
intense turbulence. In such a flume, measurement of the total sediment transport may be conducted by taking only suspended
samples. The turbulence flume set up at Dunning on the Middle
Loup River contained a series of baffle piers in a criss-cross
arrangement on top of a concrete base. The base was placed at the
same elevation as the original river bed. The additional turbulence
thus created was effective in putting all the sediment into a state of
suspension. Suitable samplers could be used for depth integration
(Vanoni, et al., 1975). Another example is at the outlet of a stilling
basin below a dam, where the flow is so turbulent that all the sediment is in a state of suspension. It provides a place to take samples
representing the total sediment load passing through the dam
outlet structures.
The idea of a turbulence flume is practical. Its advantage
is that a conventional method can be used without modification to
obtain the transport rate of the total load. The total sediment
discharge can be measured reliably and directly. However, this
type of construction may be not feasible for use on the main stem
of large rivers.
6.4.1.3 MEASUREMENT BY SEDIMENT ACCUMULATION
The total deposition or the growth of deltas in a small reservoir
over a certain period of time can be determined through repetitive
surveys. The volume of deposition, converted to weight and
divided by the duration, will give the average rate of accretion of
sediment discharge in the reservoir. The sediment passing through
the reservoir can sometimes be measured accurately by taking
only suspended
€ samples in fully developed turbulence sections at
the outlets. This amount can be added to the deposition to determine the total sediment discharge.
133
Methods for conducting a reservoir survey are discussed
in Chapter 4. A certain degree of accuracy can be achieved in
determining the total sediment discharge if the surveying work is
done strictly, according to the accepted standards.
6.4.2
Computation of total sediment load from measured
suspended sediment discharge data at a hydrometric
station
Owing to the complexities of bed load movement, less labourintensive techniques are still not very well developed for
measuring bed load discharge in rivers. In contrast, the measurement of suspended load, after long-term research and
development, now yields acceptable results in most sedimentladen rivers. However, except in some experimental reaches or
basins, there are still no reliable means of measuring the total sediment load in a river; neither can the suspended sediment discharge
which exists in all verticals be accurately estimated by ordinary
sampling procedures. Schroeder and Hembree (1956) pointed out
that in wide and shallow streams, the total quantity of bed load
and suspended load within the unmeasured zone may well amount
to 20 to 60 per cent in some cases. It may exceed 100 per cent for
coarse particles. Chien and Wan (1998) pointed out that errors
exist for either the depth integration method or sampling by
points, and described the methods of evaluating the correction
coefficient for both sampling methods; these will be discussed
later in the section.
6.4.2.1
THEORETICAL BACKGROUND
The basic idea of computing the total sediment load
from the measured suspended sediment discharge data at a hydrometric station may be illustrated by the following equation:
QT = QM
Q SC
Q SCM
(6.6)
where QT is the total sediment discharge over the entire depth,
including bed load; QM is the actual measured suspended sediment
discharge, QSC is the computed theoretical sediment discharge
over the entire depth, and QSCM is the computed theoretical sediment discharge for the measured zone.
For measurements in a vertical, the ratio of the
computed sediment discharge over the entire depth to that in the
measured zone in a vertical may be evaluated by the Einstein total
load transport theory. If a case depth integration method is used, it
may be expressed as follows (Einstein, 1964):
 E z −11 − A z (1 + PI1 + I 2 ) A
iT q T
=  

i S q SM  A  1 − E  (PI1 + I 2 )
E
(6.7)
where iTqT and iSqSM are the total sediment discharge over the
entire depth and the suspended sediment discharge over the
measured zone, respectively, expressed in size fraction and per
unit width, E is the ratio of the thickness of the unmeasured zone
to the flow depth, A is the ratio of thickness of the bed load layer
to the flow depth (in the original Einstein formula, A stands for
2D/d, where D is the grain diameter, and d is the depth of flow), z
is the exponent in the sediment distribution formula, and equals
ω/κU* (where ω is the settling velocity of the sediment grains, U*
is the shear velocity, and κ is the universal coefficient), and the
parameter P is computed by:
134
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT

d 

P = 2.303 × lg30.2
ks / χ 

ibqsb = 11.6 (A • d • CA • U*')
where ks is the dimension of roughness elements, χ is a function
of the ratio of ks and thickness of laminar sublayer as ks/δ, and I1
and I2 are two definite integrals which are functions of A and z:
I1 = 0.216 ×
I 2 = 0.216 ×
1
A z −1
(1 − A) z
1
A z −1
(1 − A)
z
∫
A
∫
A
z
1 1 − y 

 dy
A y 
z
1 1 − y 

 ln ydy
A y 
(6.9)
(6.10)
For sampling by points, the ratio of the computed sediment discharge over the entire depth to that evaluated by data
obtained at points may be deduced as follows (let θ denote the
ratio):
ι q
θ = n Τ ST
(6.11)
d
kC u
∑
i
yi yi
1
The denominator in the expression is the formula used to
compute the sediment discharge in a vertical. In the expression, ki is
the weighting factor to be applied to each measuring point, and the
theoretical values of the point concentration are Cyi, and point velocity uyi and total sediment discharge iT qST are given by the equations
presented below:
C yi = C 2 D
d − y   A 
i



 yi  1 − A 
(6.15)
(6.8)
where CA is the sediment concentration in the size group in the
bed layer (at a distance 2D from the bed in the original Einstein
formula), D is the mean grain size of the size group, z is the exponent in the sediment distribution formula, and equals ω/κU*',
where ω is the mean settling velocity of the size group and κ is the
universal coefficient, and U*' is the grain shear velocity;
After derivation, an equation for computing θ is
obtained:
θ = 4.648
(1 + PI1 + I 2 )
(1 − A) z
A
z −1
n
∑
i =1
1 − x
i
ki 
 xi
z
 (P + 2.303 lg xi )

(6.16)
where xi is the relative depth. Once the weighting factor ki is determined for the specific sampling method (by points), θ can be
evaluated.
A graph quoted from Chien and Wan (1998) is shown in
Figure 6.9. In the graph, the ratio is expressed by 1/θ for measurements taken at three points in a vertical with a weighing factor of
1:2:1 and assuming P = 13. It can be seen that the ratio of the
amount of unmeasured to measured sediment discharge would be
too large to be meaningful if the suspension index z exceeded 0.6
to 0.8. In other words, for coarse sediment, direct measurement of
the bed load and improvement of sampling methods are needed to
obtain reliable data.
z

y 
u yi = 5.75U∗' log 30.2 i 
ks χ 

iTqST = ibqsb (1 + PI1 + I2)
(6.12)
(6.13)
(6.14)
Figure 6.9 — Corrections applied to the measured data in a vertical by
sampling at three points.
6.4.2.2 THE MODIFIED EINSTEIN PROCEDURE
The modified Einstein procedure (MEP), first proposed by Colby
and Hembree in late the 1950s, has been widely used in some
rivers in the United States to estimate the total load. In practice,
the proposed MEP method was formulated on the basis of the
Einstein total load transport formula with some modifications, and
it is applicable for computing the total sediment discharge over the
whole cross-section for streams where the bed material is
composed mainly of sand and small gravel. It has been verified in
medium and small rivers where total sediment load data are available. Schroeder and Hembree (1956) have applied this method in
large rivers with sandy beds. Pemberton (1972) of the United
States Bureau of Reclamation also proposed a modification to the
Einstein formula for use in the planning and design of hydrological projects. Stevens (1985) worked out a computer program to
facilitate the computation. The procedure uses the data obtained in
the measurement of suspended sediment discharge, such as
discharge, width, average velocity or depth, water surface slope,
average sediment concentration within the sampled zone, water
temperature, size gradation of measured suspended sediment and
bed material to estimate the unmeasured suspended sediment and
bed load discharge.
The MEP method was developed for rivers with bed
material predominantly of sand and small gravel on the basis of
the depth integration sampling method. The study conducted by
Lin (Lin and Liang, 1997) indicated that the method could also be
applied to the Middle and Lower Yellow River with fine sand and
coarse silt bed materials and to stations in which the field data
were based on points-method sampling. They made some modifications to the MEP computer program as proposed by Stevens to
make it suitable to the Yellow River. Li and Long applied this
modified program to compute the total load using data collected
CHAPTER 6 — OPERATIONAL METHODS OF SEDIMENT MEASUREMENT
during the measurement of the suspended sediment discharge at
several hydrometric stations in the Lower Yellow River, and
applied the coefficient to correct the daily suspended sediment
load which was derived from measured data after proper data
processing (Li and Long, 1994).
Deposition or erosion in river reaches or in the reservoir
of the Middle and Lower Yellow River may be evaluated by the
sediment balance equation (i.e. difference of sediment load
measured at two terminal stations taking account of the intermediate in or out flows). If the total load was computed by the
aforementioned method instead of the hydrometric stations’
measured suspended sediment load in the balance equation, the
result would correspond much better to the amount of sedimentation obtained directly through repetitive range surveys (Li and
Long, 1994).
6.4.2.3 CORRECTION COEFFICIENT
To correct the measured sediment concentration in a vertical, the
ratio of iT·qT /iS·qSM expressed in section 6.4.2.1 may be considered as a correction coefficient. For the depth integration method,
the ratio may be computed by Equation 6.6, and similarly, for the
sampling by point method, it may be computed by Equation 6.15.
The measured suspended-sediment discharge of a given size fraction multiplied by θ will give the total sediment discharge per unit
width of the size fraction at the vertical. Obviously, the summation
of the total sediment discharge of all the size fractions will give
the total sediment discharge in the vertical. These can be summed
for all verticals at a section to give the total sediment discharge at
the measuring site of the stream.
In considering the correction of the measured sediment
€ accounting for the unmeasured zone in the depth
concentration
integration method, errors of discharge measurement should also
be considered. For the point integration method, for instance, with
five points in a vertical, the lowest point for taking either velocity
measurements or sampling for sediment concentration is usually
around a relative depth of 0.95 instead of the theoretical position
of relative depth of 1.0. On the one hand, the measured sediment
concentration at this point will be lower than that at the river
bottom. On the other hand, the measured velocity will be greater.
Hence, the deviation of the computed sediment discharge per unit
area from the true value depends on the relative magnitude of the
velocity and the concentration, or distribution, of the product of
Uy and Sy.
6.4.2.4
RATIO OF BED LOAD DISCHARGE TO SUSPENDEDSEDIMENT DISCHARGE
Most published sediment data are limited to the suspendedsediment discharge. For a rough estimate, the ratio of the bed
135
load to suspended load may be used empirically for the estimation of the total sediment discharge. Following the same line of
approach expressed in the preceding paragraphs, the ratio may
be written as:
r=
ib q sb
1
=
iT q T 1 + PI1 + I 2
(6.17)
where r is the ratio of bed load discharge to suspended-sediment
discharge, and iTqT is the total sediment discharge per unit width
for a certain size fraction.
The ratio varies with the diameter of transported sediment and the boundary conditions of the flow, and may be
estimated by Equation 6.16. According to an estimation based on
field data from some hydrometric stations on the Yellow River, the
average value of r may vary from 0.14 to 0.88 per cent. The
maximum value of r of various stations varies from 0.8 to 4.2 per
cent. However, for rivers with relatively stable boundary and
inflow conditions, the range of variation may not be so wide
(Zhang and Long, 1998).
Maddock made a summary of the ratio of the bed load to
suspended load based on annual loads, as shown in Table 6.6. The
ratio varies with different bed compositions and suspended sediment concentrations (Vanoni, et al., 1975).
In an alluvial river, the suspended bed material load and
the bed load being transported may have the same correlation, as
they are related to the hydraulic conditions of the flow. The transport rate of wash load depends more on the available supply of the
fine material contributed from the watershed. For this reason, the
ratio of the bed load to suspended load is valid only for bed material load.
6.4.3
Comments
The following pertinent points are worth mentioning regarding the
evaluation of the total sediment discharge. In the first place, the
total load may be classified as bed material load and wash load.
Wash load transport depends on the availability of the sediment
from the source area, and moves essentially as suspended load. An
accurate estimation of wash load relies mainly on reliable
measurement in the field, either by sampling or in situ measurement. An indirect method for estimating the total sediment
discharge, as discussed in the previous section, would only give an
evaluation of the bed material discharge and not of the wash load.
Bed material discharge depends fundamentally on the transport
capacity of the flow, which may be evaluated by transport formulae for given hydraulic and morphological conditions. Transport
formulae should be verified or modified if necessary using
observed data. The amount of wash load can be estimated by
direct measurement.
Table 6.6
Estimation of ratio of bed load to suspended load
Suspended sediment
concentration (ppm)
Bed composition
Suspended load
composition
Ratio (r)
< 1 000
Sand
Gravel, consolidated clay
Similar to bed
Small amount of sand
0.25–1.50
0.05–0.12
1 000–7 500
Sand
Gravel, consolidated clay
Similar to bed
25% sand or less
0.10–0.35
0.05–0.12
> 7 500
Sand
Gravel, consolidated clay
Similar to bed
25% sand or less
0.05–0.15
0.02–0.08
136
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
In small rivers, the overall ratio between bed load and
suspended sediment may be fairly constant over the years because
it is controlled by the source properties. However, at instances
between floods it can be highly variable.
Evaluation of the total load based on measured
suspended sediment data is a promising approach. However, the
reliability of the computation of the total load relies upon the
accuracy of the proposed sediment transport formula used in evaluation of the correction factors. The idea of applying a correction
factor to the measured sediment discharge has been expressed. In
this report, Einstein’s bed load function, as well as the total load
transport formula, is used to illustrate the theoretical background,
and is used after some modifications in the modified Einstein
procedure (MEP). However, verification of the Einstein formula,
since it was originally proposed with field data, indicates that the
computed results do not match the field data very well, particularly in relatively low flows. On the basis of Einstein’s theory,
Wang, et al., proposed a new transport formula (Wang, et al.,
1995). The same line of approach was followed in the development of this new formula and some parameters originally used in
the Einstein formula were replaced by the results of newly
conducted experimental studies. The proposed formula has been
verified both by measured bed load and suspended load at hydrometric stations in the Middle and Lower Yellow River, with
satisfactory results (Zhang and Long, 1998). The new transport
formula will provide a sound theoretical basis for evaluating the
ratio QSC/QSCM expressed in Equation 6.6.
It should be noted that the sediment moving in the vicinity of the bed is composed of coarse bed material particles. The
correction coefficient is much greater for coarse sediment than
fine sediment. In an alluvial river, particularly the downstream
reaches, the bed load and the suspended bed material load being
transported in the vicinity of the bed may constitute only a small
fraction of the total load. However, it is important in sedimentation
studies in determining the total transport load. It also plays an
important role in the fluvial process and reveals a proper relation
between the transport rate and the hydraulics of the flow, which is
a basic characteristic in the study of fluvial processes.
6.5
LABORATORY PROCEDURES
6.5.1
Determination of sediment concentration
Suspended sediment samples obtained in the field must be treated
in the laboratory for the determination of sediment concentration
and particle size. Evaporation, filtration and displacement methods
are generally used in laboratories to determine the sediment
concentration. The method is chosen on the basis of the quantity
and the composition of sediment in the sample and the desired
accuracy. In the Chinese Standards, a minimum weight of the
sediment in the sample is required, in accordance with the
sensitivity of the weighing apparatus, in order to make it possible
to use evaporation and filtration methods and specific gravity flasks
of different capacities in the displacement method. In the United
States, it was found that the filtration method might best be used on
samples containing sand concentrations of less than 10 000 mg/l
and clay concentrations of less than 200 mg/l. The evaporation
method is applicable to samples ranging from 0.2 to 20 l in volume,
from 5 to 500 000 mg/l in sediment concentration, and having less
than 35 000 mg/l in dissolved-solid concentration. In addition, the
wet sieving method is used if two concentration values are required:
one for sand size particles and one for a combination of silt and clay
sized particles. The sample is separated by a sieve with 0.062 mm
square apertures. The coarse fraction is treated by the evaporation
method and the fine part, after splitting, may be weighed either
through filtration or evaporation (ASTM Standard D3977-97).
In general, sediment concentration is determined by the
weight of the dried sediment contained in the sample, divided by
the volume of the sediment-water mixture sample. An indirect
method, for example, is to take a reading from a physical apparatus, such as a turbidity meter, to obtain the sediment concentration
from a calibration curve that expresses the relationship between
the reading and the sediment concentration.
Sediment concentration is expressed in three different
ways: CS represents the weight of dried sediment contained in a
unit volume of sediment-water mixture commonly expressed in
mg/1, g/l or kg m–3; CSG represents the weight of dried sediment
divided by the weight of the sediment-water mixture and may be
expressed in percentage of weight (%) or in parts per million
(ppm); and C SV represents the volume of sediment particles
contained in a unit volume of the sample, expressed in per cent
(%) or as a ratio.
6.5.1.1 EVAPORATION METHOD
In the evaporation method, the wet sediment sample, after the
supernatant liquid is decanted from the vessel, is transferred to an
evaporation dish and dried in an oven at a temperature slightly
below the boiling point until the visible moisture is evaporated.
The oven temperature is then raised to 105ºC for two hours. If the
dissolved solids exceed 2 per cent of the sample weight, their
concentration should be determined separately in the original
water. The content of dissolved solids should be subtracted from
the dried sediment weight in computing the sediment concentration. The dry weight of the evaporation dish is usually precisely
determined beforehand. In routine operations, it should be
checked to avoid any possible error.
6.5.1.2 FILTRATION METHOD
Filtration is used to determine concentration that is low. The
quality of the filter material influences the accuracy of this
method to a great extent. Experiments should be carried out to
test the filter material before it is finally selected. The first
experiment is to determine the amount of sediment that may be
leaking through the filter material. If the leak exceeds 2 per cent
of the total sampled sediment, better quality filter material
should be used. The second experiment is to determine the
content of soluble matter in the filter material. By comparing the
dry weight of the filter material before and after immersion in
water, the weight loss can be determined and used to correct the
dry weight of the sediment obtained by filtering. In the United
States, it is considered that the filter pore size (filter ratings) and
filter diameter are critical in the filtration method. Filters with
retention ratings of 1.5 micron and a filter diameter exceeding
24 mm are commonly used in the sediment laboratory (Edwards
and Glysson, 1999).
In the United States, a crucible is used in conjunction
with various types of filter material in the filtration method. The
crucible is a small porcelain cup of about 25 ml in capacity with a
perforated bottom. Glass fibre filter disks have proved satisfactory
for the filtration of most types of sediments. Force filtering may be
used, in which air pressure is applied to the water surface to speed
up the filtering process. If much fine-grained material is contained
CHAPTER 6 — OPERATIONAL METHODS OF SEDIMENT MEASUREMENT
137
Table 6.7
Allowable error in determination of sediment concentration*
Method
Evaporation
Filtration
Displacement
Random error (%)
Systematic error (%)
Volume
measurement
Sediment
weight
Loss during
decantation
0.5
0.5
0.5
1.0
1.0
2.0
1.0–2.0
Dissolved
solids
Leakage through
filter
Absorption of
filter paper
1.0–2.0
1.0–2.0
1.0–2.0
* ranges of error are set for different class stations.
in the sample, a glass fibre filter disk may be used in conjunction
with an asbestos mat. The crucible is adapted to an aspirator
system and vacuum filtration to speed up the filtration process
(ASTM Standard D3977-97).
6.5.1.3 DISPLACEMENT METHOD
The displacement method involves determining the difference in
weight between a sample of sediment-laden water and an equal
volume of clear water. This method can only be applied to samples
with a relatively high sediment concentration. The dry weight of
sediment is computed by the following equation:
WS = k (WWS + WW)
(6.18)
k = ρS/(ρS – ρ)
(6.19)
where:
where WS is the sediment weight to be determined in g, WWS is the
weight of the specific gravity flask plus the weight of sediment
water mixture in g, WW is the weight of the specific gravity flask
plus the clear water weight with volume and temperature equal to
that of the sediment-water mixture (during weighing the water
temperature should be constant), ρS is the density of sediment
particles, and ρ is the density of water.
In routine work, the values of k and WW have been tabulated in forms for a given water temperature. The density of
sediment particles, ρS, should be checked occasionally. At normal
temperatures, k varies from 1.59 to 1.61 for a range of ρS from
2.65 to 2.70. WW also varies with the temperature. In laboratories,
for commonly used specific gravity flasks (usually 50, 100, 200 or
250 ml in volume), it is calibrated once a year and its value can
easily be determined once the temperature is known. To ensure
accuracy, water temperature in the flask should be measured to
0.1°C and WW should be weighed to 0.001 g.
In calibrating W W , the original water may be used
instead of distilled water. If raw water is used in weighing WW,
and the influence of dissolved solids on the concentration is negligible, no correction is needed. If the dissolved solids vary
substantially in a year, it would be better to calibrate the weight of
the flask by using distilled water and to make necessary corrections for the dissolved solid content.
6.5.1.4 ACCURACY REQUIREMENT
The allowable error in the measurement of sediment concentration
is put forward in the Chinese Standards, as shown in Table 6.7.
Accuracy in determining sediment concentration relies
mainly on accuracy in weighing. For balances with different
sensitivity used for weighing, a minimum amount of sediment is
required to ensure an acceptable accuracy such as specified in the
Chinese Standards. It is clear that either the balance should be
chosen according to the sediment weight available to be sampled,
or the quantity of samples should correspond to the available
Table 6.8
Size analysis methods commonly used in China and the United States
Range of application
(diameter in mm)
Concentration (g/l)
Fine sediment--------Settling in clear water (two-layer system)
Siltmeter
0.062–0.5; may be more if
longer tube is used
Visual accumulation
tube
0.062–2.0
Fine sediment------- Settling in dispersed medium system
Pipette
0.002–0.062
0.002–0.062
Photo-sedimentation
0.005–0.062; may be
used for 0.005–0.1
Hydrometer
0.005–0.062; may be
used for 0.002–0.05
Coarse sediment
Sieve
Direct measurement
0.062–20.0 or more
0.062–32.0
Required sample weight (g)
0.3–5.0
0.05–15.0
3.0–20.0
2.0–5.0
<1.0
3.0–20.0 in 1 000 ml
1.0–5.0
<1.0
15.0–30.0
15.0–30.0 in 1 000 ml
100–200 if done independently
More than 20 for coarse particles;
min. 0.05
Sufficient quantity
138
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
laboratory apparatus. Nevertheless, errors may be easily introduced
in the measurement of sediment concentrations if sediment samples
are not properly treated. The procedures put forward in relevant
standards should be strictly observed.
6.5.2
Size analysis
6.5.2.1 METHODS FOR SIZE ANALYSIS
There are many methods available for size analysis. The size
distribution of a sediment sample may spread over a wide range.
Two or even more methods may be necessary to analyse the whole
sample. For instance, the sieve method may be used for small fractions of coarse particles while the visual accumulation (VA)
method or its equivalent size-analyser method is used for particles
greater than 0.062 mm, and the pipette or photo-sedimentation
method is used for particles smaller than 0.062 mm. The methods
commonly used for size analysis in routine work in China and the
United States are listed in Table 6.8.
Sediment size as obtained by different methods has
different meanings. When the sediment particle is directly
measured by a rule, its size is measured in three mutually perpendicular directions denoted by a, b and c, in which c and a are the
shortest and longest axial lengths respectively. The mean diameter
is the summation of a, b and c divided by three. The shape factor
SF is given by the expression c (ab)–1/2. The nominal size is
expressed by the diameter of a sphere with the same volume as the
particle that is obtained by immersing the particle in water and
measuring the volume of displaced water. Sediment size determined by sieve analysis is called the sieve diameter. Sediment size
determined by methods based on the settling principle is defined
as the diameter of the sphere that has the same settling velocity
and the same density as the given particle. It is called the settling
diameter, or sedimentation diameter.
There are overlaps in the range of application of the
sieve method and methods based on the settling principle. The
relationship between sizes with different definitions has been
studied and noted in relevant literatures or standards. Owing to
the different meanings in the definitions of particle size evaluated by the various methods, the size distribution curves will not
coincide with each other at the junction portion when two
methods based on different principles are employed in size
analysis. Empirical revisions or corrections are necessary at the
junction point. For fine sediment, methods based on the settling
principle are recommended, as no discontinuity in the gradation
curve is induced by the definition of size implied in the methods
used for size analysis.
The results of size analysis are usually expressed by a
size gradation curve with an accumulated percentage finer as the
ordinate and a sediment diameter in the logarithm scale as
abscissa. A log frequency curve may also be used. Nevertheless,
characteristic figures can always be interpolated from the curve,
such as D50, D35, D65, D90, P0.05 and P0.025, etc., where P represents the percentage of the total sample that is finer than the
indicated size. The mean diameter and mean settling velocity of
each size fraction is usually expressed by its geometric mean
value, i.e. √D1D2 or √ω1ω2.
In the ISO draft Standard and the Indian Standards, size
analysis for suspended load is performed by subdividing the total
suspended load into three size groups: larger than 0.2 mm, 0.2 to
0.075 mm and smaller than 0.075 mm, representing coarse,
medium and fine particles. For bed load and bed material, the
sediment sample is subdivided into two portions: smaller and
larger than 0.6 mm. Conventional methods are used for the
detailed analysis of each portion.
The treatment of suspended samples in three parts is a
kind of simplification of the method employed in the determination of the percentages of each portion. Just as in the
simplification methods used in sampling suspended sediment, the
simplification for size analysis is worth studying. For instance, in
an alluvial river with a bed composed mainly of coarse sand, silt
and clay particles are the wash load. If the amount of sand (greater
than 0.05 mm) could be roughly determined during floods by
considering a simplified analysis of the index samples, a better
understanding of the role played by coarse particles in the fluvial
process would be revealed.
The classification of sediment sizes in the size gradation
of a sediment sample involves dividing the sizes into size groups.
The demarcation is set at sizes so that the latter is twice as large as
the former, in ascending order. The nomenclature and division are
shown in Table 6.9. In the Geological Department, the sediment
size is usually represented by φ, which is defined as φ = – log2D.
(1) Sieve analysis. Sieve analysis is a traditional method
used for the mechanical analysis of sand and gravel. From a practical point of view, the direct measurement of particles greater
than 20 to 32 mm contained in a sample may be more convenient
than sieving. National standards for sieves, as well as operational
specifications used for analysis, have been established in most
countries. When sieve analysis is adopted for the size analysis of
fluvial sediment, two methods may be used. The wet-sieving
method carries out the analysis by immersing the whole sample in
water while the sieving operation is performed, or a small water
Table 6.9
Nomenclature and division of sizes
Name
Gravel
Sand
Size range (mm)
φ phi-system
Very coarse
Coarse
Medium
Fine
Very fine
64–32
32–16
16–8
8–4
4–2
–6
–5
–4
–3
–2
Very coarse
Coarse
Medium
Fine
Very fine
2–1
1-0.5–
0.5–0.25
0.25–0.125
0.125–0.062
–1
0
+1
+2
+3
Size range (mm)
φ phi-system
Coarse
Medium
Fine
Very fine
Coarse
0.062–0.031
0.031–0.016
0.016–0.008
0.008–0.004
0.004–0.002
+4
+5
+6
+7
+8
Medium
Fine
Very fine
0.002–0.001
0.001–0.0005
0.0005–0.00025
+9
+10
+11
Name
Silt
Clay
Only the upper limit is expressed in the φ system.
CHAPTER 6 — OPERATIONAL METHODS OF SEDIMENT MEASUREMENT
size D1, and C1 is the sediment concentration at position h,
corresponding to sediment size D1. During the period 0 to t1,
the total sediment passing position h should be:
jet is used to rinse all the particles to speed up the process. In the
dry-sieving method, sieving is performed in the usual way. The
sieves are shaken to speed up the process. To ensure accuracy in
sieve analysis, a comparison of the results obtained with sieves
used in routine work should be compared against those obtained
with standard sieves on a regular basis. Corrections should be
made if necessary.
At present, the lower limit of sizes within which sieve
analysis may be applied is 0.062 mm. It is recommended that
settling diameter, rather than sieve diameter, be used in analysing
suspended sediment. Also, it is preferable to use methods based on
the settling principle, such as the VA-tube method or the size
analyser method for the analysis of sizes ranging from 0.062 to
1.0 mm, which are commonly found in suspended sediment. For
bed materials, however, the major part of the sample will be in the
sand range, and sieves are more convenient for this analysis. The
characteristics of bed material are usually expressed directly by
size, while for suspended sediment it is more common to give
characteristics in terms of settling velocity or settling diameter
rather than sieve size.
(2) Methods based on the settling principle. According
to the settling medium, methods based on the settling principle
may be classified into two groups: settling in clear water, or the
two-layer system, and settling in sediment-laden water, or the
dispersed system (Allen, 1977).
(a) Two-layer System. This is also called the stratified system.
The settling tube is filled with clear water (distilled water)
prior to analysis, and sediment is inserted into the tube from
the top. Different-sized particles will separate automatically
in the tube according to their own settling velocities. The
right-hand side graph in Figure 6.10 is a sketch illustrating
how this settling system works.
At time t, all particles of size D, which have fall velocity
h/t1, have settled to position h. Sediment discharge per unit
area passing through the cross-section of the cylinder should
be C1ω1, where ω1 is the settling velocity of the particle of
∫01Cωdt
t
P> D1
t = t1
t=0
t = t1
Figure 6.10 — Sketch illustrating the settling process in two systems
for determining size distribution of fluvial sediment.
0
T
(6.21)
0
(b)
t=0
∫ C ω dt
=
∫ C ω dt
t1
Depth (cm)
€
(6.20)
This is the part of the sediment with size equal to or
greater than Di in the total sediment sample. The greater
percentage in the total sediment sample could be expressed
as:
(b) Stratified system
(a) Dispersed system
139
where T is the time required for the settling of all the particles in the sediment sample.
This is the basic principle of the visual accumulation
(VA) tube and the size-analyser method. In practice, the
settled sediment weight is obtained directly in both methods.
In the VA tube method, the height of accumulation at the
bottom is recorded and converted into weights by the relationships obtained from previous experiments. In the
size-analyser method, settled sediment from the tube is withdrawn at prescribed intervals and the weight can be
determined directly. An apparatus developed by Delft
University of Technology in the Netherlands is called DUST,
and has similar functions.
Settling in a clear water system is suitable for size
analysis of 0.062 to 1.0 mm sediment, i.e. medium and fine
sand. In practice, the settling velocity of a small group of
sediment particles is measured instead of the settling velocity of a single particle. This does not fulfil the requirement
set forth in the assumption on which the formula for settling
velocity is based. It has been shown by experiments that tube
size and the quantity of the sample have an influence on the
analysis results. A method for correcting the results obtained
by the VA tube method was suggested in the report of the
United States FIASP. Size analysis was performed by the VA
tube method, and corrections could be made by comparison
of the results with the known size-gradation curve, if there
were any differences. A standard sample was prepared by
subdividing a sediment sample into groups by sieve analysis.
One hundred particle grains were then picked out from each
group. Settling velocities were determined in the tube for
each individual particle, and the size distribution within each
group could be calculated. The composition of the size
distribution curve for each group according to the weight of
each group would give the size distribution curve of the
composite sample, which was the known gradation of the
standard sample to be used for comparison (FIASP Report,
1963). Based on a similar principle, a correction method for
size analysis by the size-analyser method has also worked in
China (Xiang and Li, 1994).
Dispersed System. Size analysis methods that adopt the
dispersed system and are commonly used in various countries include the pipette, hydrometer, bottom withdrawal tube
and photo-sedimentation, etc. These methods are suitable for
sizes of less than 0.062 mm, in the silt and clay range, or, in
practice, from 0.062 to 0.002 mm. With the pipette method,
140
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
water and sediment are mixed in a cylinder as shown in the
left-hand sketch of Figure 6.9. At time t, a small volume of
mixture v1 (in ml) is withdrawn at distance h below the water
surface. After treatment, the dry sediment weight w1 in the
small sample can be obtained. The percentage P by weight
of sediment finer than D1 can be computed by:
PD < D1 (%) =
w1 V
v1 W
(6.22)
where V is the volume of the water-sediment mixture in the
test cylinder, and W is the total sample weight placed in the
test cylinder.
In the pipette method, a 1 000 ml graduated cylinder is
usually used for the analysis. After sufficient dispersal of the
mixture in suspension, five or six samples (25 ml) are withdrawn from the centre of the cylinder intermittently at
positions 5, 10 or 20 cm from the water surface at predetermined times. A period of 10 seconds is allowed for each
withdrawal by pipette. From experience, the optimum
concentration recommended for the suspension is 0.5 to 2.0
per cent by weight. Evaporation or other appropriate
methods may be used to determine the dry sediment weight
contained in the pipette samples.
Photo-sedimentation is a method used extensively by
various industries for determining size gradation. It is a
simple, rapid method particularly suitable for size analysis of
fine sediment of silt size. It is based on the principle of the
scattering of light transmitted through a sediment-laden
water medium. From the light scattering theory, it can be
deduced that the intensity of transmitted light is related to
the intensity of light before transmission, as follows:
I = Io exp [–k L C/D]
(6.23)
(
)
where I and Io are the intensity of transmitted light and that
before transmission respectively, k is the extinction coefficient, L is the distance between the light source and the
detector, C is the sediment concentration in g 1–1, and D is
particle size.
Photo-density (the ratio of I/Io) depends not only on the
concentration C but also on the particle size D existing in the
medium. Figure 6.11 shows the relation of I/Io versus C
using d as parameters, obtained from experiments with
Yellow River sediment. The extinction coefficient k does
vary with size, but approaches a constant when the particle
Figure 6.11 — Relation of I/Io versus C.
size exceeds 0.02 mm. Experimental data of the extinction
coefficient k versus grain size fits quite well with that
deduced from theory (Lu, et al., 1983). It would therefore be
possible to establish a relationship between the sediment
concentration and a photo-density reading only if the grain
size was relatively constant. In operation, the instrument
must be calibrated carefully to establish such a relationship.
Many comparisons have been made for the results
obtained with the pipette and photo-sedimentation methods
by analysing the same sample. The average deviation for a
number of samples is less than 1.0 to 1.5 per cent, with a
maximum deviation for a single sample of less than 4.5 per
cent. Repetitive analysis by the photo-sedimentation method
shows that the deviation from the average value of percentage finer is less than 5 at the 80 per cent confidence level.
To ensure the reliability and consistency of the size
analysis, some standards have recommended that the adoption of new methods for size analysis should be based on the
results of comparisons with traditional methods. The allowable error is specified (Chinese Standards, 1992).
Comparisons may be made with the percentage finer for a
specific index size or another index size by which a size
gradation curve can be defined. It was found through the
intercomparison of size analysis methods conducted in
China and the United States that the results obtained by the
photo-sedimentation method are comparable to those
obtained by the traditional pipette method (Long, et al.,
1989; Lu, 1995).
A semi-automatic pipette withdrawal apparatus has been
developed, as reported by FIASP. The auto-pipette is an apparatus that makes six scheduled withdrawals (for particle sizes
of 2, 4, 8, 16, 31, 62 mm) automatically in the pipette size
analysis procedure. A fixed-elevation, 12-depth siphon
sampling scheme is used instead of mechanically lowering the
pipette to a predetermined depth for each withdrawal. An
optical water level sensor stops the siphon when the correct
volume of sample is obtained. Flushing of the siphon line
precedes each of the scheduled sub-samples (Beverage, 1982).
6.5.2.2 TREATMENT OF SAMPLES FOR SIZE ANALYSIS
Sediment samples should be treated in preparation for size analysis. The state of sediment particles moving in natural streams is
quite complicated. Flocculation, coagulation and various physical
phenomena have been observed in natural rivers when the sediment particles are transported, eroded or deposited throughout the
river course. It should be noted that these processes continue after
collection. Dissolved salts, organic matter and flow turbulence
influence the physical state of sediment particles. This is why
samples should be treated prior to size analysis.
For suspended sediment analysed by settling methods,
there are two schools of thought on the treatment of samples. The
first one is to treat the sample to achieve a standard state so that
the analysis results obtained at different times can be compared to
each other. The other one is to keep the sample in a state as close
as possible to its natural state. Since the influence of water quality
and the physical state of the particles on the settling property of
fluvial sediment is still not well known, it is difficult to study the
settling property of sediment particles in different environments.
This is true in particular for fine sediment such as fine silt and clay
particles, among which flocculation easily takes place. It has been
CHAPTER 6 — OPERATIONAL METHODS OF SEDIMENT MEASUREMENT
shown by experiments that flocculation occurs easily when there
are appreciable amounts of Ca++ and Mg++ ions present in the
original water, or when the organic matter absorbed or attached to
sediment particles exceeds 1 per cent of the sediment weight. The
influence varies with sediment concentration in the river. If the
sample has not been treated for organic matter and no dispersing
agent has been applied to the suspension medium, no reasonable
explanations can be given for the results of size analysis in the
original water, due to the complicated relationships among the
variables. In some rivers, flocculation varies with the season,
while in other rivers no change is noticeable. For this reason,
analysers generally find it preferable to disperse the sediment
sample and use distilled water as a settling medium. In other
words, size analysis is preferably carried out in a standard state
instead of a natural state. In general, samples are treated for
dissolved salts and organic matters, and dispersing agent is added
to the sample according to specifications or standards.
During field sampling for size analysis, appreciable
amounts of high organic sediments or fragments are sometimes
present in the sample. The density of coal powder is quite different from that of ordinary sediment particles that are composed
mainly of minerals. The separation of coal particles is necessary
to minimize the probable error induced by the difference in
densities.
From the above discussion it may be concluded that,
with the present state of knowledge, it is better to make size analysis by a standard method of sample treatment, keeping the
sediment in a state of dispersion. Research, such as parallel analyses, should be conducted to study the potential flocculation and
the influence of organic matter on size distribution. The chemical
analysis of water should occasionally be carried out with raw
water while collecting samples for size analysis.
6.5.2.3 MEASUREMENT OF PHYSICAL PROPERTIES
Density or specific gravity is an important physical property of
sediment particles, which may be measured with a specific-gravity
flask. In general, the specific gravity of sediment particles varies
from 2.60 to 2.70. For quartz sand particles a value of 2.65 is
usually assumed. Sediment particles are composed of various
kinds of rock fragments, mineral fragments and clay minerals.
Specific gravity determined by standard methods represents an
average value of the composite sample. If an appreciable amount
of coal powder is present, it should be separated from the sample.
The unit weight or dry density of the bed material is also
an important parameter in the study of sediment transport. The
method of sampling undisturbed samples for the determination of
unit weight is discussed in Chapter 4.
6.6
DATA PROCESSING
Sediment data acquired by various means has to be processed in a
unified manner. Daily, monthly and annual sediment load and
variations in size gradations are computed by appropriate
methods. The results are tabulated and published together with the
observed stream gauging data.
6.6.1
6.6.1.1
Data processing for suspended load
COMPUTATION OF SEDIMENT DISCHARGE AND CROSSSECTIONAL AVERAGE SEDIMENT CONCENTRATION
Ideally, sufficient sediment discharge measurements should be
taken routinely in cross-sections to define time and space varia-
141
tions. However, in practice, simplified methods such as index
sampling have to be used, particularly during floods. If the
concentration of the index sample is closely related to that of the
cross-sectional average, and deviations from the regression line
are less than ±10 –15 per cent at a frequency of 75 per cent, the
relationship may be considered relatively stable. The correlation
can be used to convert the index sample concentration to the crosssectional average value. Sometimes the correlation may be
established according to a variation of stages or to different
seasons of a year.
Another conversion method is to compute the proportional coefficient that is the ratio of measured cross-sectional
average concentration to the corresponding index sample
concentration. The coefficient is plotted on a hydrograph and
the line representing the variation of coefficient with time may
be used for interpolation. The cross-sectional average concentration can then be obtained by multiplying the index sample
concentration by the proportional coefficient interpolated from
the graph.
Here, further comments on index sampling are called
for. In section 6.1, the idea of taking index samples is considered
merely as a simplified method to supplement a conventional
method. If three to five verticals, arranged on an equal discharge
increment basis, are adopted as an index sampling method, the
result may be acceptable. However, if only a single vertical is
used, the position for taking index samples should be carefully
selected so that the index sample can be better correlated to the
average sediment concentration of the whole section. As can be
seen from the transverse distribution, there should be a position, or
one or two verticals in a cross-section, where the ratio of average
concentration in the vertical to the cross-sectional average is equal
to 1. If this position is relatively stable, it can be used for taking
the index samples. In some streams where the concentration is low
but varies greatly in the transverse direction, or in some untrained
rivers, significant and time-dependent differences may exist even
in the higher concentration ranges. The advisability of adopting
this kind of index sampling should be carefully examined by
studying actual data and should be determined in the light of experience. At sites where optical or nuclear concentration gauges are
used along with an automatic pumping device mounted at a fixed
point of the cross-section, the data collected are equivalent to an
index sample concentration. The relationship with the crosssectional average value should be examined to estimate the
applicability of these kinds of physical apparatuses.
6.6.1.2
COMPUTATION OF AVERAGE DAILY SEDIMENT DISCHARGE
AND CONCENTRATION
During the low-flow season or when the water discharge shows
little variation, only one sample is taken daily or even over several
days. Average daily sediment concentration is usually obtained by
interpolation of an appropriate value from the sediment hydrograph for the day. Sometimes, samples taken on successive days
are combined for treatment. The resulting concentration may be
used as the average concentration for the period. If there is no
appreciable change in discharge but the sediment concentration
shows variations, several samples may be taken in a day. The
arithmetic mean of the concentration may then be used as the
average value.
If there are appreciable variations in discharge and sediment concentration during a day, the errors resulting from the
142
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
computation of daily average sediment concentration by the
methods discussed above will be intolerable. In such cases, the
concentrations should be weighted with the water discharge in the
computation of daily average sediment discharge or concentration.
The most common methods may be summarized as follows:
Assuming the temporal variation of sediment discharge
is linear, the mean sediment discharge in time period ti to ti+1 is:
1
(6.24)
(Qi ρi + Qi +1ρi +1)
2
where Q and ρ are the discharge and sediment concentrations,
respectively. The daily sediment discharge is then:
Qsi =
Qs =
1
48
n
∑[(Q ρ + Q
i +1ρi +1
i i
)∆ti ]
(6.25)
1
In other words, the sediment discharge should be
weighted by the time interval it represents to give the mean daily
sediment discharge.
If it is assumed that the discharge Q and sediment
concentration vary linearly, then the daily mean sediment
discharge may be computed by:
1
QS =
96
n
∑(Q
i −1 + Qi )
(ρi + ρi +1)∆ti
(6.26)
1
The errors involved in this method may be smaller than
those in the method of Equation 6.27 under conditions in which
both discharge and sediment concentrations change drastically
during a day and the number of measurements is insufficient to
delineate the changes. Nevertheless, the above methods are
approximate methods. To be exact, the average sediment discharge
in a period should be computed by integration.
QS =
1
t i +1 − t i
∫
t i +1 −t i
0
Q ⋅ ρ ⋅ dt
(6.27)
After integration and simplification:
QS =
1
1
(Qi ρ Si + Qi +1ρi +1) + (Qi ρi +1 + Qi +1ρi )
3
6
(6.28)
The daily sediment discharge is computed as follows,
dividing a day into n time periods:
Qs =
1
72
n
∑
1
[(Qi ρi + Qi +1ρi +1)∆ti ] +
1
144
n
∑[(Q ρ
i i +1 + Qi +1ρi
)∆ti ] (6.29)
1
In one computation of daily sediment discharge for
several stations in a tributary of the Yellow River, it was found that
the error induced by the approximate method using Equation 6.29
ranges from -0.6 to 2.8 per cent, while for the method using
Equation 6.28, it ranges from +1.0 to 6.5 per cent.
6.6.1.3 SEDIMENT TRANSPORT CURVE
The relationship between water discharge and suspended sediment
discharge, or sediment concentration, is sometimes called the sediment transport curve, or the sediment rating curve. If a sufficient
number of sediment discharge measurements is taken, the sediment transport curve can be plotted and used for interpolation or
extrapolation purposes. The transport curve can be drawn from
measurements at different stages of a flood. If the peak of the sediment discharge lags behind the peak of discharge, a clockwise
loop is usually obtained, and vice versa. In cases where insufficient observed data are available to define the loop, an average
line is drawn through the data as an approximation. However, the
result is less accurate or reliable than if the loop in the rating curve
can be drawn. Glysson described in detail the process of developing sediment transport curves, including the choice of dependent
and independent variables, procedures for developing a transport
curve, and the effects of seasonal variations, major sediment transport events and timing of peaks on the shape of transport curves.
Curve fitting methods are also discussed for using the transport
curve to estimate the sediment load for periods when measured
sediment data are not available (Glysson, 1987).
The relationship of water discharge to sediment concentration may be drawn for different time intervals such as
instantaneous, daily, monthly, annual or flood period. The instantaneous curve may reflect the effect of different factors on basic
transport characteristics. However, it is not theoretically applicable
to the direct computation of daily sediment discharge from daily
water discharge, except for days on which the rate of water
discharge is approximately constant throughout the day. Daily or
instantaneous water-sediment discharge curves, adjusted for
factors that account for some of the scatter from an average curve,
may be used to compute approximately the daily, monthly and
annual sediment discharge (Mimikou, 1982).
6.6.1.4 DATA PROCESSING FOR SUSPENDED SEDIMENT SIZE
The percentage finer for a certain size is commonly used to
express sediment size in computation. Usually, only a limited
number of precise measurements for the distribution of sediment
concentration and sediment size over an entire cross-section is
available in a year. Therefore, the index sample used in the
measurement of sediment concentrations has also been used for
size analysis to define the variations in sediment size with time.
Again, the relationship between the percentage finer for the unit
samples and for the cross-sectional average samples can be used
for determining the average size distribution. It is recommended
that deviations from the average correlation line should not exceed
±3 to 5 per cent for 75 per cent of the points for coarse particles
and not exceed ±5 to 10 per cent for 75 per cent of the points for
fine particles. The percentage finer for a certain size of the index
sample can be converted to a cross-sectional average value by
means of this relationship.
Since the vertical and transverse distributions of sediment concentration have different characteristics for different
sediment sizes and vary with the hydraulic elements of the flow, it
is impossible to obtain a simple correlation between the sediment
concentration for various size groups of the index sample and that
of the cross-sectional average sample. The method discussed
above is merely an approximation for practical purposes. For
rivers with a large amount of fine material, the errors induced may
be negligible. However, if coarse particles are predominant, the
errors should not be overlooked. The discharge of coarse particles
is usually underestimated.
Average daily, monthly and annual values of the percentages finer for a certain size of suspended sediment can be
computed by weighting individual values with the sediment
discharge. If the sediment discharge is relatively stable, an arithmetic mean may be used without introducing appreciable error.
CHAPTER 6 — OPERATIONAL METHODS OF SEDIMENT MEASUREMENT
Another approach in determining the mean size distribution of
suspended sediment is to divide the sediment into size groups. For
each size group, the procedures discussed in the previous sections
should be followed in the computation of average daily, monthly
and annual sediment concentrations, and that of sediment
discharge.
6.6.2
Data processing for bed load
Bed load is part of the bed material load. It varies with flow
velocity and other hydraulic properties. The measured bed load
discharge should be plotted on the hydrograph of water stage,
discharge and suspended sediment discharge to detect any inconsistencies. Abrupt changes or deviations from an average
tendency should be checked for reliability in the measurement of
bed load.
If sufficient bed load measurements are made over a
section, a hydrograph may be plotted to show bed load movement
for the duration of the hydrometric investigation. Daily bed load
transport rates may be read directly from the hydrograph. The
accuracy of each measurement relies, of course, upon the proper
selection of sampling verticals and the sampling techniques.
Although bed sediment moves at random under
average conditions, a definite relationship exists between the
bed load transport and hydraulic elements such as discharge or
stream power, which can be used for computing the daily bed
load. Empirical relationships between bed load discharge and
hydraulic parameters established on the basis of measurements
in low and medium flows may be extrapolated to high flow
conditions. After verification with the measured bed load, wellknown formulae may be used in the calculation. With size
analysis data, relationships can also be established for different
size groups and can be used for computing the bed load
discharge in size groups.
6.6.3
Examination of processed data and data processing
using computers
The processed data, including the average daily, monthly and
annual sediment concentration and sediment discharge, should be
carefully examined for their reasonableness, and all calculations
should be checked. For the data obtained at a single station, the
relationship between the sediment concentration of an index
sample and a cross-sectional average concentration, and relationships between sediment discharge or sediment concentration
versus water discharge as shown by the data obtained over a year,
should be compared with the relationships used in previous years.
If there have been no changes in the operational methods, either in
the measurement of sediment discharge or in the collecting of
index samples, the trends in the relationships should not vary.
Points deviating from the trend should be checked for correctness,
or possible reasons for the deviation should be explored.
Hydrographs of discharge, water stage and sediment concentration
should be drawn to detect any unreasonable bias. Inconsistency
can usually be judged by experience, and should be rectified if
necessary.
Sediment and water balance data should also be used in
the examination of processed data. The monthly and annual
sediment discharges at stations located on the same river should be
tabulated according to a sequence from upstream to downstream.
Inflows from tributaries should be added to the sediment load at
upstream stations and compared with the sediment load at
143
downstream stations. Amounts of sediment withdrawn from the
river, sediment inflow from intermediate regions and the amount of
deposition or erosion should be estimated or measured in order to
detect any bias. This process can be expedited by applying a
sediment-balance equation (WMO, 1994; Ministry of Water
Resources, 1975b).
In the published yearly report, explanations should be
given concerning the major factors and procedures followed in the
data acquisition and processing stages to help users judge the
quality of the data for their specific purposes. An explanation of
data processing should include: (a) Operational methods for
sampling suspended sediment, instrumentation, methodology,
sampling frequency and problems to be solved, etc.; (b) Data
analysis, checking for reasonableness and interpolation method, if
applicable; (c) Assessment of accuracy and reliability of the data;
and (d) Suggestions for future work and unsolved problems, etc.
The data-processing method discussed in the previous
sections is the traditional method performed manually in many
hydrological offices. However, fundamental rules still have to be
observed if computers are adopted for data-processing purposes
such as the recording and transmission of observed data, processing of data according to a definite program and the storage,
retrieval and publishing of the processed data.
Depending on the policy adopted by different countries,
the analysed hydrological data may or may not be transmitted to a
centralized office for further processing. This is particularly true
for sediment measurements such as samples taken in the field that
have to be sent to regional or district laboratories for size analysis,
even though sediment concentration is usually determined in field
laboratories. Except in a few cases, it is unnecessary to transmit
sediment data on a real-time basis. However, recent developments
in automatic observation systems, as well as the widespread adoption of computerized systems for data processing, have created the
need for the efficient transmission of observed data after preliminary processing.
Different transmission systems may be selected according to the speed at which data are required and the availability of
proper installations. Procedures for the transmission of hydrological data may include manual, semi-automatic and fully automatic
methods. Details of the transmission methods will not be
discussed in this report. Some general guidelines are discussed in
the Guide to Hydrological Practices (WMO, 1994). The vast
quantities of observed and processed data being gathered call for
careful consideration regarding data storage. At present, most of
these data are transferred to magnetic tape or discs for working
storage.
After processing and tabulation in standardized formats
convenient for various uses, sediment data are published. The
annual report is a common form of presentation for all fundamental hydrometric data, including sediment data. Since sediment data
are always used in conjunction with water-flow data, they should
be published as a complete set rather than separately.
In general, the annual report is divided into volumes
according to the river drainage basins. Every year, data obtained at
various stations located in the same drainage basin receive preliminary processing at each station, and are then compared to rectify
any processing errors. If there are any unreasonable results, there
should be a careful examination of the field and office work and, if
necessary, additional field investigations should be conducted. The
next procedure is a final examination of the data obtained at
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
various stations in the same drainage basin, within which water
and sediment balance should be achieved. The publication of the
hydrological data is the final step in the annual data-processing
exercise. Needless to say, the work should be published promptly,
with errors kept to a minimum. In recent years, computer technology has been universally adopted for data processing, including
publication. When processing data using computers, the normal
fundamental procedures have to be followed if reliable data are to
be expected. Clear responsibility for the long-term stewardship
and long-term security of the raw and published data needs to be
established.
The formats used for publication may be greatly influenced by whether or not computers are used. If the data have been
collected on machine-readable media, or if manually-collected
data have already been transferred to machine-readable media,
tabulation can be performed by computer line printers or by photo
composition much faster and more economically than by
manually-typed copy. Standardized data formats are usually used
in publications. Guidelines on data processing issued by relevant
agencies are available for use. WMO/HOMS reference manual
components may also be of use.
6.7
ASSESSMENT OF ACCURACY AND
RELIABILITY IN MEASUREMENT OF
SEDIMENT TRANSPORT
6.7.1
General description
Measurement errors may be classified as systematic or random
errors. Random errors, represented by the precision of measurement, are caused by many independent factors. As the number of
measurements is increased, the distribution of the deviations of
observed data from the mean value tends to follow a normal distribution. Thus, if there are no systematic errors, a mean value can be
determined which approaches the true value as the number of
observations increases. However, if there are systematic errors, the
problem cannot be eliminated by merely increasing the number of
observations. Hence, systematic errors will accumulate with an
increase in the number of observations. Systematic errors may
constitute only a small fraction of the total amount of observed
sediment discharge, yet intolerable errors can result if the measured
sediment load is used in the estimation of the total amount of
erosion and deposition for certain reaches. Both random and
systematic errors should be controlled within allowable limits. The
elimination of systematic error in a measurement is a key problem
with regard to improving the reliability of sediment data.
6.7.2
Major factors influencing the reliability of
measurement of sediment transport
6.7.2.1 APPARATUS
The apparatus used in routine measurement should be chosen
carefully and maintained to minimize probable errors. For timeintegrating and depth-integrating suspended sediment samplers,
the ratio of intake velocity to ambient velocity is an important
factor that must be controlled. For a sediment size less than
0.45 mm, the error would be less than ± 5 per cent if the velocity ratio could be controlled within a range of 0.8 to 1.2
(Edwards and Glysson, 1998). Errors may also be induced by
misuse of depth-integrating samplers. The transit rate of the
sampler should be kept uniform and should be less than 0.4
times the average velocity in the vertical; otherwise, samples
may not be representative.
Beijiazuan Station
Concentration (g 1–3)
144
Huayuankou Station
Time (mins)
Figure 6.12 — Fluctuations of sediment concentration in the Yellow
River and its tributary.
For an instantaneous trap-type sampler, natural fluctuations in sediment movement have a large influence on the
observed sediment concentration. The fluctuations vary with the
characteristics of flow as well as with sediment concentration.
Figure 6.12 gives examples obtained by means of radioisotope
gauges at two hydrometric stations, one on the main stem of the
Yellow River and the other on a tributary. It is clearly shown
that fluctuations in sediment concentration appear less intensive
under high concentration than under low concentration. The
concentrations of the samples taken with horizontal trap-type
instantaneous samplers at more than 10 stations on several large
rivers in China were analysed for errors resulting from fluctuations in concentration. The study showed that the relative
standard error in measured concentration due to fluctuations
could reach ± 10 per cent.
6.7.2.2 CHARACTERISTICS OF MEASURING SECTIONS
The boundary and hydraulic conditions of the measuring section
are closely related to the accuracy of the measurement. If the
measuring section is sited at a narrow constriction of the river and
the bed is composed mainly of gravel and pebbles, sufficient
mixing will take place to suspend sand material due to the flow
turbulence. The distribution of sand-size material should be fairly
uniform, both vertically and transversely. Under such conditions,
samples taken by conventional or even simplified methods can be
considered representative and accurate, in comparison with
samples taken at reaches in wide alluvial channels with a sand
bed. At an ordinary cross-section in a river reach, however, the
distribution of the concentration of coarse sediment is not uniform
either in transverse or along a vertical. The gradient of sediment
concentration for coarse particles in the vicinity of the riverbed
increases very rapidly. Errors involved in disregarding the sediment load transported in the so-called unmeasured zone are
inevitable. In the measurement of suspended sediment, the error
involved in sampling coarse particles is far greater than that for
fine particles.
6.7.2.3 SAMPLING FREQUENCY
For rivers where the sediment source is from upland erosion
caused by storm rainfall, the major part of the annual water and
sediment flow are concentrated in the flood season, and particularly in large floods. Continuous records of sediment
concentration have been kept on the River Creedy in England, as
reported by Walling, et al. (1981). These records have shown that
80 per cent of the total yearly sediment load is transported in 3 per
cent of the time. In the Yellow River, on average, 68 per cent of
the total sediment yield is transported in only two months of the
CHAPTER 6 — OPERATIONAL METHODS OF SEDIMENT MEASUREMENT
145
Table 6.10
Minimum weight of sample required for size analysis based on settling principle
VA tube
Suitable range
Sand size
Minimum weight
required
0.05–15.0
Size analyser
Pipette
Hydrometer
BW tube
Photo-sedimentation
0.5–1.8
<0.5
Silt and clay size
0.3–5.0
1.0–5.0
3.0–20.0
15.0–30.0
Source: Ministry of Water Conservancy, 1975a; Edwards and Glysson, 1998.
Table 6.11
Precision and bias for sediment concentration test methods
Concentration
added
Concentration
recovered
Evaporation
mg/l
10
100 000
Standard deviation of
single operator (So)
Filtration
Evaporation
Filtration
Evaporation
Filtration
8
2.5
2.6
2.3
2
Bias (%)
Evaporation
Filtration
mg/l
9.4
100
1 000
Standard deviation of
test method (St)
91
976
5.3
961
36.8
100 294
–6
–20
5.1
20.4
15.9
532
–9
14.1
–2.4
360
–3.9
0.3
Source: ASTM Standard D3977-97.
Table 6.12
Precision of photo-sedimentation method of size distribution
Particle size (mm)
0.005
Settling system
Deviation of cumulative percent
finer at 80% confidence level
0.01
0.025
0.05
Disperse system
1.8
2.3
year. In 1977, the sediment transported in three floods lasting only
10 days amounted to at least 70 per cent of the annual load. The
temporal variation in sediment discharge should be considered
when sampling frequency is selected.
0.05
0.10
0.25
0.5
3.0
2.5
Clear water system
3.2
4.8
0.6
3.0
comparison with the sediment concentrations obtained by traditional methods.
6.7.2.5
MEASUREMENT OF CONCENTRATION AND SIZE ANALYSIS
IN THE LABORATORY
6.7.2.4 IN SITU MEASUREMENTS
Radioisotope gauges and turbidity meters have been used in some
countries to measure sediment concentration in situ. The accuracy
of in situ measurements obtained using nuclear gauges depends on
the characteristics of each apparatus. Other conditions being
equal, the precision of measurement is closely related to the
counting rate of the instrument in receiving radioactive signals
from the source. The higher the counting rate, the higher the precision and the lower the smallest detectable concentration. As
regards the nuclear gauges currently in use, the lowest detectable
concentration is approximately 0.5 g/1, with an allowable relative
error of 10 per cent. In the low concentration range, measurement
error increases with the decrease in sediment concentration. The
lowest concentration for which the use of a nuclear gauge is
permitted can then be determined by setting an allowable error for
the measurement of sediment concentration. To ensure the desired
accuracy, attention must be paid to field calibration or to field
checks on the calibration curve by means of other reliable
sampling methods. Changes in water quality and mineral composition of the sediment may induce variations in the calibration
curve. It is important to calibrate the instrument in the field by
Errors involved in the treatment of sediment samples are one of
the error sources when sediment concentration and size gradation
are determined. The volume of the sample required to ensure a
certain degree of accuracy in the determination of sediment
concentration should be considered with reference to the sensitivity of the balance available in the laboratory. The sample weight
should fulfil the minimum requirements of the size analysis
method. The minimum requirements are listed in Table 6.10.
The errors involved in the laboratory treatment of
samples for sediment concentration and size analysis have been
analysed in some countries. Systematic errors can easily result if
some important procedures are not followed, for instance,
correction for dissolved solids and calibration of the specific
gravity flask, etc. The precision and bias for the concentration test
method put forward in the ASTM Standard is as follows. Samples
for collaborative testing were prepared by dispersing a specially
prepared dry powder in approximately 350 ml of water. Mixtures
were shipped in sealed glass containers to the nine participating
laboratories, where three Youden pairs at each of the three
concentrations were tested. The results of the test for the three
methods specified in the Standard are shown in Table 6.11.
146
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
The precision of different size analysis methods has been
evaluated by many parallel experiments conducted by YRCC.
Each sample is divided into more than 30 parts, and repetitive
measurements using the same analysis method are made to
determine deviations from the mean value. Table 6.12 gives the
results of the experiments determining the precision of sediment
size distribution using the photo-sedimentation method.
6.7.2.6 COMPUTATION METHOD AND DATA PROCESSING
The main purpose of processing sediment transport data is to
calculate the total amount of sediment transported in a month, a
year or in a flood period for a given river. Methods for data
processing may be grouped into extrapolation and interpolation
methods. By interpolation, the sediment concentration is
determined from the actual measured values, and the total
sediment load is computed by integration of the product of
discharge and sediment concentration with time. By
extrapolation, a sediment transport curve is established which
defines the average relationship between instantaneous sediment
concentration and water discharge. Other parameters, such as
different sources and seasonal variations, etc., may be used if
necessary to improve the co-relationships. Sediment discharge
determined by means of the discharge hydrograph of a given
period, together with the sediment transport curve, can only be
used for a very rough estimate.
Walling, et al. (1981) studied the effect of various dataprocessing techniques and frequency of sampling on the
accuracy of calculated sediment yield by using continuous
records of sediment concentration extending over seven years on
the River Creedy, in the United Kingdom. The ratio of sediment
yield estimated by taking concentrations at different sampling
frequencies to the actual measured sediment discharge obtained
by detailed computation is used as an index of precision. In the
study, sediment discharge obtained by the interpolation of sediment concentration and weighted by discharge provided a result
with a relatively high accuracy. If the concentration is not
weighted by discharge, the total sediment load is seriously
underestimated. In assessing the reliability of data-processing
methods, both accuracy and precision should be considered
(Wallin, et al., 1981).
The average daily sediment concentration of the Yellow
River is usually computed by one of the following methods.
Average daily concentration may be obtained by taking the
average concentration value interpolated from the sediment
concentration hydrograph, or it may be obtained by computing
sediment discharge, integrating with time to obtain a daily amount
of sediment and then dividing by the mean daily discharge. A
comparison of the sediment load during a flood event at Lintong,
Weihe River, shows that the difference in the two methods
amounts to nearly 9 per cent of the total sediment load transported
in the flood. It is recommended that discharge-weighted sediment
concentration be used rather than the average concentration
method in the computation of sediment discharged during floods.
6.7.3
Major factors influencing the reliability of bed load
measurement
The operational method for the measurement of bed load
discharge differs considerably from that for suspended sediment
due to the spatial and temporal variations in bed load movement.
Experience on the East Fork River indicates that verticals
Figure 6.13 — Variation of relative error to the suspension index z.
densely distributed across a river and measurements taken on
double traverses are necessary to obtain an accurate and reliable
bed load discharge. In routine measurements, such requirements
are not easily satisfied. The fact that the sampler efficiency is not
stable and that bed load transport varies spatially and temporally
makes it very difficult, if not impossible, to obtain reliable bed
load data.
In alluvial rivers, bed load material, including the
discharge of bed load and part of the suspended load, should be
closely related to the hydraulic and boundary conditions of the
flow. Direct measurements taken under relatively stable conditions
can be used to establish or verify such relationships. An estimate
of the yearly sediment yield can be made by extrapolation, using
mathematical or physical models in which the total sediment
transport rate has been verified for stable flow conditions.
6.7.4
Analysis of systematic errors
The systematic errors involved in sediment measurement are illustrated here by two case studies. In the first case study, long-term
data on the amount of erosion and deposition obtained through a
sedimentation survey were accumulated for Sanmenxia Reservoir
and the Lower Yellow River. The amount was compared with that
computed by the sediment balance equation. The elements
involved in the sediment balance equation included the difference
in sediment load at two terminal hydrometric stations, the amount
of bank erosion, inflow from the intermediate drainage basin, sediment withdrawn together with water for irrigation, and the unit
weight of deposits. It was found in this case study that the systematic errors involved in the sediment measurement at the inflow
hydrometric stations might be slightly greater than 2% and that
coarse sediment constituted a major part of the deviations.
According to the second case study, systematic errors induced by
sediment measurement in a vertical may be estimated by the
methods proposed in section 6.4. Einstein’s total load transport
formula is used to estimate the probable error in sediment concentration for different size groups. According to the actual data
obtained at some stations on the Yellow River, the quantity P
usually varies from 10 to 16 and A varies from 10–5 to 10–3.
Assuming an average value of 13 for P and 10–5 for A, the value
of the relative error may be computed. The factors P and A are
defined in section 6.4. The relative error is the ratio of the
CHAPTER 6 — OPERATIONAL METHODS OF SEDIMENT MEASUREMENT
147
Table 6.13
Allowable error of suspended sediment measurement
Random error uncertainty X
Station
A
(%)
L
V
W
Systematic error (%)
A
L
V
W
Grade I
10
4.2
12
4
±1.0
–2.0
±1.0
±1.0
Grade II
16
4.2
16
6
±1.5
–3.0
±1.5
±1.5
Grade III
20
4.2
20
10
±3.0
-4.0
±3.0
±3.0
difference between total and measured sediment discharge as
percentages of the total sediment discharge. Variation of the
computed relative error versus z, the suspension index in the
sediment distribution formula, is shown in Figure 6.13. It can be
seen that as the value of z becomes greater than 0.47, the relative
error is greater than 10 per cent for all the simplified methods for
the measurement of sediment discharge in a vertical. The greater
the value of z, the larger the relative error will be.
6.7.5
Analysis of random errors
Random errors may be eliminated by repetitive measurements, as
discussed in previous sections. However, this is true only for
certain independent variables and not for quantities such as water
stage, discharge and sediment concentration, which are unsteady
in nature. Nevertheless, with statistics from long-term records it is
still possible to obtain an average value of the variable under
certain conditions. The accuracy of sediment measurement is in
general not very high. Random errors within a tolerance limit are
easier to deal with than systematic errors.
There are several sources of errors involved in the
measurement of cross-sectional average sediment concentration or
sediment discharge. The first category relates to the measurement
of width, depth and velocity and to the sampler’s performance and
efficiency. The second concerns the fluctuation properties of the
velocity and sediment concentration. The third category belongs to
the errors involved in the laboratory analysis of samples. The
fourth category concerns errors related to the method of taking
measurements, such as the number of points in a vertical or representatives of the index sample, etc.
When assessing the probable errors involved in the
measurement, experiments carried out on site are required to
compare the results obtained by a conventional method or
instrument
€ with those obtained by a more precise method or a
standard instrument. The random uncertainty for a measurement
of cross-sectional average sediment concentration is composed
of two sources of errors. The errors involved in average sediment
concentration include: (i) errors inherent to the sampler, which
are the deviation of the results obtained with the apparatus used
in comparison with those obtained by a standard apparatus (A);
(ii) errors involved in the laboratory analysis of the sample (L);
(iii) fewer number of points in a vertical or method of
computation for the average concentration in a vertical (V); (iv)
errors involved in the evaluation of the cross-sectional average
sediment concentration caused by an insufficient number of
verticals or method of computing the average sediment
concentration in a cross-section (W); (v) errors caused by an
insufficient sampling duration, due to a temporal fluctuation of
sediment concentration (T).
For instance, in a vertical, if the average sediment
concentration Csm obtained by taking measurements at more than
five points in a vertical, for example, seven points, is used as a true
value of the average concentration in the vertical, and the average
sediment concentration Cs is obtained by using fewer points, then
the relative standard error is:
 1 
σ 2 =

 n −1 
∑( E − E )
i
2
s
(6.30)
where Ei = (Csi/Csm) — 1, and Es = 1/n ∑ Ei.
Similarly, the relative standard error in the measurement
of the cross-sectional average sediment concentration may be
expressed by similar equations. In the Chinese Standards, the
uncertainty of a measurement is expressed by a percentage. For
normal distribution, the random uncertainty X should be 2σ in its
value at a confidence level of 95 per cent. It is specified that the
random uncertainty and systematic error involved in the sediment
measurement should be limited, as shown in Table 6.14 (Chinese
Standard GB 50159-92).
In Table 6.13, A denotes the uncertainty induced by the
instrument used in the measurements; it is obtained by intercomparison with the standard calibrated instrument. L is the
uncertainty induced in the treatment of sediment samples. V is the
uncertainty of the mean sediment concentration in a vertical,
which is induced by limited sampling points in the vertical
(including that induced by the method of calculation of the mean
concentration in the vertical). W is the uncertainty of the average
sediment concentration in the cross-section, which is induced by
the number of verticals and also by the method of computing the
cross-sectional average concentration.
The total random uncertainty and systematic error of a
measurement of cross-sectional average sediment concentration
should be determined by a mixture of all the errors. i.e.:
XCT = [Xw2 + (1/(m+1)) (XA2 + XL2 + XT2 + XV2)]1/2
(6.31)
where X represents the uncertainty value, the subscripts represent
the errors specified in the previous section, and m is the number of
verticals. The total random uncertainty of a measurement of sediment discharge is computed by:
XQS = [XCT2 + XQ2]1/2
(6.32)
where XQ represents the total random uncertainty of discharge
measurement expressed as a percentage.
It is well known that systematic and random errors are
inherent in sediment measurement. Systematic errors should be
148
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
minimized by improving measurement methods, or eliminated by
applying corrections to the measured data. Random errors should
be minimized by enhancing the precision of the measurement,
including the operational methods used in the field, the treatment
of sediment samples in the laboratory and data-processing
methods.
Similarly, the error in sediment deposition measurement
by survey should be studied. A 134-km long reach of the Lower
Yellow River was surveyed in the early 1960s by the range
method, with an average distance between ranges of about 1 to
3 km. Xiong, et al., (1983) found that the relative standard error in
computing the amount of erosion or deposition with only half of
the ranges amounted to 12 per cent.
Up until now, only limited research has been carried out
to assess the accuracy and precision of sediment transport
measurement. The measurement of sediment discharge, and of the
bed load in particular, is relatively crude compared to the wellestablished methods for stream gauging. The systematic error
involved in the measurement of total sediment discharge has
created indefinite factors in the evaluation of sediment deposition
by the difference in the sediment load method. Deviations between
the range method and the difference in the sediment load method
are commonly found in river reaches or in reservoirs.
Improvements in measurement methods are necessary to enhance
the accuracy and precision of the measurements.
6.8
SUMMARIES AND RECOMMENDATIONS
A better understanding of sediment yield, sediment transport and
erosion or deposition is of vital concern to all engineers engaged
in the planning and development of water resources. The proper
selection of operational methods for sediment measurement relies
not only on the basic knowledge of sediment movement in rivers
or in reservoirs, but also to a large extent on the accuracy required
for data acquisition. To summarize, the following recommendations are listed for reference.
6.8.1
Fundamental concepts
The data-acquisition programme for the study of sedimentation
problems in river basins is given in Table 6.14.
6.8.2
Implementation of measuring programmes
On sediment-laden rivers where sediment management in the river
basin is a problem, a programme of sediment measurement should
be worked out to evaluate the amount and variation of sediment
transport with existing and supplementary hydrometric networks.
Sedimentation surveys should be carried out periodically in
important river reaches and reservoirs for a better knowledge of
the spatial variation of erosion or deposition.
For important river reaches or reservoirs, the inflow
hydrometric stations should be able to measure the input from
more than 80 per cent of the drainage basin. Measurement of the
total sediment discharge should be carried out at such stations. For
ordinary reservoirs, a minimum of 60 per cent of the drainage
basin should be represented by inflow gauging stations at which
sediment measurement is taken. Inflows from tributaries contributing more than 10 per cent of the total sediment inflow should also
be measured.
6.8.3
Measuring site
Channel conditions, including the bed material composition and
flow conditions in the main channel and over the flood plain, etc.,
should be thoroughly investigated by reconnaissance. If it is
necessary to measure the total sediment discharge, a section is
preferred at which all sediment is well mixed in the flow by fully
developed turbulence. Such stations can be located at the outlet of
a dam, or at localities where artificial roughness can be set up. At
such stations, conventional suspended sediment measuring techniques may be employed to obtain the total load data. For small
rivers, measuring structures may be constructed in which vortex
tubes or trenches can be installed to collect the bed load. In rivers
where fine suspended sediment constitutes the major part of the
total sediment load, an estimate of the total sediment discharge by
taking only suspended sediment measurements should provide
data with a fair degree of accuracy. However, the probable bed
load discharge can only be estimated by analytical methods. The
operational method for suspended sediment measurements should
be chosen carefully.
Table 6.14
Programme of data acquisition according to the International Hydrological Programme (IHP)
Purpose of study
Items of measurement
Surveying
Annual sediment discharge
Sediment transport
Relevant items
Total sediment discharge or
concentration at hydrometric stations
Water discharge, etc.
Erosion and deposition in
river reach or reservoir;
depletion of reservoir
capacity
Sedimentation survey by
ranges in a river reach or
reservoir
Total sediment discharge at inflow
and outflow gauging stations
Size distribution and/or unit
weight of deposits
Fluvial processes in river
reaches or in backwater
reaches of a reservoir
Repetitive survey over
entire reach or in localities
of interest: aerial photographs
if possible
Bed material discharge at inflow
stations
Relevant hydraulic and
sediment parameters such as
water surface slope, bed
material composition, velocity,
depth and width, water
temperature, size distribution
of sediment
Source: UNESCO, 1982.
CHAPTER 6 — OPERATIONAL METHODS OF SEDIMENT MEASUREMENT
6.8.4
Measurement of suspended sediment discharge
It is very important to select an appropriate operational method for
measuring suspended sediment discharge. For each measuring
station, field data obtained by multi-point methods should be
analysed to establish simplified methods that can be employed
during floods. The relationship between the sediment concentration obtained from an index sample and the cross-sectional
average sediment concentration obtained by a multi-point method
should be established for conversion purposes. An index sample is
one taken at a pretermined vertical, or set of verticals, by depth or
point integration methods.
6.8.5
Corrections for transport in the unmeasured zone
The sediment discharge value as measured using conventional
methods of suspended sediment measurement is usually inadequate for coarse sediment in the sand-size category. The depth
integration method leaves an unmeasured zone in the vicinity of
the bed due to the fact that the sampler nozzle is above the bed
when the sampler rests on the bed. If the measurement involves
sampling by points in a vertical, some errors will be induced since
it is impossible to take samples right at the bed surface where the
concentration is the greatest. Corrections are necessary if the total
sediment discharge is to be obtained. Methods similar to the modified Einstein procedure may be employed for correction purposes.
6.8.6
Frequency of measurement
A common feature of rivers in which floods are produced mainly
by rainstorms is the non-uniformity of both water and sediment
flow. A sufficient number of measurements during floods is
needed to monitor the entire process. Experience plays an important role in finding a compromise between the proper timing of
measurements and the selection of adequate measurement
methods.
6.8.7
Sampling apparatus — suspended sediment
Time-integration samplers have been used extensively in recent
decades. Besides the well-adapted depth-integrating or pointintegrating series, collapsible-bag samplers or portable pumping
samplers can also be used advantageously. The in situ measurement
of sediment concentration using newly developed instruments
designed on the basis of physics such as nuclear gauges or ultrasonic or vibration type apparatuses incorporated with computer
data processing units should be encouraged. However, the necessity
of calibrating samplers or measuring devices in the laboratory and
in the field prior to their adoption should be emphasized. New or
untested sampling methods should be evaluated by comparing their
data with that obtained by conventional methods in flows with a
wide range of concentrations. One should be well aware of the fact
that most suspended sediment samplers collect samples containing
both bed material and wash load. If morphological predictions have
to be made in which a transport formula is required, the wash load
should be determined by size analysis of the sample and excluded,
since it is the bed material that is of major importance in river
behaviour. However, the wash load may have an influence on the
transport of bed material. As mentioned in the previous sections,
some samplers such as the Delft bottle directly measure the sediment discharge of bed material while others, such as the
pump-filter sampler, measure the concentration of bed material in
suspension. These samplers may be used advantageously to study
transport characteristics in rivers that carry a small amount of
149
sediment. The selection of an appropriate apparatus must be based
on the objectives and technical considerations of the measuring
programme.
6.8.8
Sampling apparatus — bed sediment
All bed load samplers should be properly calibrated to define their
sampling efficiency. An efficiency of more than 50 to 60 per cent
is considered to be satisfactory for use in the field, provided that
great attention is paid to the operation of the bed load sampler so
as to overcome the uncertainties caused by the temporal and
spatial distribution of bed load movement.
When studying the armouring effect on the transport
characteristics of an alluvial river, sampling and analysis of the
bed material are important. Samples of bed material from the sand
and small gravel size categoires can be taken with conventional
samplers currently in use. However, there are still some difficulties
involved in the sampling of coarse gravel.
6.8.9
Computation of total load
Methods for evaluating the total sediment discharge by a combination of field measurement and analytical measures appear to be
promising, and should be studied further. The formulae used in the
analytical methods should be verified with actual measured data,
when available. As regards coarse-grained sediment in the bed
material, the total annual sediment discharge may not be large, but
it is significant in the study of stream behaviour.
6.8.10 Size analysis
Fluvial sediment samples should be analysed in the field or laboratory for size distribution. A rough estimate of suspended
sediment transport may be made for sediment in two to three size
groups. Samples are separated using sieves and any one of the
methods based on the settling principle. If the data are to be used
to study sediment transport characteristics, suspended load, bed
load and bed material should be analysed for size distribution. A
size gradation curve should be prepared instead of only giving the
relative amounts in just the two or three size groups.
6.8.11 Method of size analysis
In the size analysis of fluvial sediment, different methods have
their own applications. In general, for sediment particles greater
than 0.5 to 1 mm, sieving is preferable. For medium and fine sand,
silt and clay, methods based on the settling principle are preferred
because settling velocity is an important factor in the study of
suspended sediment. A system of settling in clear water, such as
the visual accumulation tube method, is suitable for sediment sizes
from 1.0 to 0.062 mm. For sediment finer than 0.062 mm, a
system of settling in a dispersed medium, such as the pipette
method or the photo-sedimentation method, is preferable.
In current practice, it is necessary to consider the
dissolved salts and organic matter in the samples to be analysed.
All the sediment particles should be kept in a standard dispersed
state in still, distilled water for settling. The native water is used in
size analysis only for comparison. The influence of water quality
on the settling characteristics still has to be determined.
6.8.12 Data processing
In the processing of sediment data, the stage-discharge relationship and the relationship of index sample sediment concentration
to the cross-sectional average concentration are of fundamental
150
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
importance. In a sediment-laden river where the bed is subject to
drastic changes during floods, the development of stage-discharge
relationships is difficult. The relationship of index sediment
concentration to average cross-sectional concentration may also be
different for different sediment sizes. The reliability and accuracy
of sediment data rely not only on the measuring method, but also
on the data-processing method. Therefore, in data-processing
work, the careful establishment of these two relationships according to the flow and sediment characteristics, using adequate actual
measured data for the field, is essential. In data-processing work,
checking the original data and some of the computed results for
their reasonableness is an essential and important task that should
be taken seriously. Computer technology is already widely used in
the processing, publication and storage of data, which provides
very useful means for the study of sediment movement in rivers
and reservoirs.
6.8.13 Assessment of accuracy and reliability
Unlike discharge measurement, there is still no established method
for assessing the precision and accuracy of sediment measurement. As analysed in previous sections, systematic and random
errors are inherent in sediment measurement. Systematic errors
should be minimized by improving the measurement methods, or
eliminated by applying corrections to the measured data. Random
errors should be minimized by enhancing the precision of
measurement, including operational methods used in the field, the
treatment of sediment samples in the laboratory and data-processing methods. According to the purposes of the data acquisition,
various degrees of accuracy should be maintained at different
stations engaged in the data acquisition programme. For instance,
if it is necessary to estimate annual sediment yield in some small
tributary rivers, a simplified method of observation may be
allowed. However, for an alluvial reach in the main tributary of a
sediment-laden river, if the measurement of sediment transport is
required for studying the fluvial processes of the reach, a relatively high standard of accuracy is required, particularly for bed
material discharge. Data on the total sediment discharge, size
distribution and relevant hydraulic parameters should be measured
and filed for further analysis.
6.8.14 Monitoring for sediment quality
There is an increasing need for improved data collection for the
study of sediment quality, as the latter is closely related to the
environmental impact of a river. Sampling procedures similar to
those used in measuring sediment discharge may be adopted, but
the standardization of analysis and careful operation are essential
if reliable results are to be expected. Sediment is a pollution
carrier and may be harmful to engineering works as a result of
settling in reservoirs and silting of canals, etc. However, sediment
can also be turned into a resource if it is well managed or
controlled. The scope of sediment measurement programmes
should be broad enough to cover the quantity as well as the quality
of the sediment in order to obtain a better understanding of sediment transport.
REFERENCES
American Society for Testing of Materials, 1997: ASTM
Standards. D5387-97, D3977-97.
American Society for Testing of Materials, 1998: ASTM
Standards. D6326-1998.
Allen, T., 1977: Particle Size Measurement. second edition,
Chapman and Hall, London.
Bajiazui Reservoir Experimental Station, 1980: Cylindrical
Revolving Sampler for Taking Undisturbed Samples of Soft
Mud. Bajiazui Reservoir Experimental Station, China.
Berke, B. and L. Rakoczi, 1981: Latest achievements in the development of nuclear suspended sediment gauges. Proceedings
of the Florence Symposium, IAHS.
Beverage, J.P., 1982: Theory and operation manual for the semiautomatic pipet withdrawal apparatus. Report Z, Federal
Interagency Sediment Project, St. Anthony Falls Hydraulic
Laboratory, Minneapolis, Minn.
Billi, P., et al., 1981: Continuous measurement of suspended sediment transport by an optical-ultrasonic device. Proceedings
of the Florence Symposium, IAHS.
Brabben, T.E., 1981: Use of turbidity monitors to assess sediment
yield in East Java, Indonesia. Proceedings of the Florence
Symposium, IAHS.
Changjiang Water Resources Commission, 1980a: Compilation of
Techniques for Bed-load Measurements.
Chien Ning, 1952: The efficiency of depth integrating suspended
sediment sampling. Transactions, AGU, Volume 33.
Chien Ning and Wan Zhaohui, 1998: Mechanics of Sediment
Transport. ASCE Press.
Chinese Standards, 1992: Standard for Measurement of Suspended
Sediment in Rivers. GB 50159-92, Ministry of Construction.
Chinese Technical Standards, 1992: Technical Standard SL 42-92
for Determination of Sediment Particle Size in Open
Channels. Ministry of Water Resources.
Colby, B.R. and C.H. Hembree, 1955: Computations of total sediment discharge, Niobrara River. Water Supply Paper 1357,
USGS.
Crikmore, M.J., 1981: Recent developments in pump samplers for
the measurement of sand transport. Proceedings of the
Florence Symposium, IAHS.
Dijkman, J.P.M. and V. Milistic, 1982: Investigations on Suspended
Sediment Samplers Based on Measurements in the Danube
River, May 1979. Joint publication S410, DHL and JCI.
Druffel, L., et al., 1976: Laboratory hydraulic calibration of the
Helley-Smith bedload sampler. Open File Report 76-752,
USGS.
Durette, Y.J., 1977: Hydac 100-An automated system for hydrographic data acquisition and analysis. Technical Bulletin 105,
Water Survey of Canada.
Durette, Y.J., 1981: Preliminary Sediment Survey Equipment
Handbook. Water Survey of Canada.
Edwards T.K. and G.D. Glysson, 1999: Field Methods for
Measurement of Fluvial Sediment. Book 3, Techniques of
Water Resources Investigations, USGS.
Einstein, H.A., 1964: Handbook of Applied Hydrology. (ed.) V.T.
Chow, Section 17, Sedimentation, Part 11, River
Sedimentation, McGraw Hill, New York.
Emmett, W.W., 1979: Field calibration of the sediment trapping
characteristics of the Helley-Smith bed-load sampler. Open
File Report 79-411, USGS.
Emmett, W.W., et al., 1981: Measurement of bed-load in rivers.
Proceedings of the Florence Symposium, IAHS.
Emmett, W.W., et al., 1983: Some characteristics of fluvial
processes in rivers. Proceedings of the Second International
Symposium on River Sedimentation, Nanjing.
CHAPTER 6 — OPERATIONAL METHODS OF SEDIMENT MEASUREMENT
Engel, R., 1981: The efficiency of basket type bed-load samplers.
Proceedings of the Florence Symposium, IAHS.
Federal Interagency Committee on Water Resources: Reports 1-14.
Subcommittee on Sedimentation 1940-1963, St. Anthony Falls
Hydraulic Laboratory, Minneapolis, Minn.
Federal Institute of Hydrology, 1992: Instruction for Bed Load
and Suspended Load Sampling. Koblenz.
Gao Huanjin, et al., 1995: Development of new type bed load
sampler and study of its sampling efficiency. Technical
Report of the Joint Project, CWRC.
Gao Huanjin and Li Zhaonan, 1988: Intercomparison of
suspended samplers. Series Report 8802, Working Group on
Sediment Measurement, Department of Hydrology, Ministry
of Water Resources, China.
Gao Huanjin and Xu Shenguo, 1989: Intercomparison of bed load
samplers. Series Report 8803, Working Group on Sediment
Measurement, Department of Hydrology, Ministry of Water
Resources, China.
Glysson, G.D., 1987: Sediment transport curves. Open File Report
87-218, USGS.
Grobler, D.C., 1981: Continuous measurement of suspended sediment in rivers by means of a double beam turbidity meter.
Proceedings of the Florence Symposium, IAHS.
Hakanson, L., 1978: Optimization of underwater topographic
survey in lakes. Water Resources Research, Volume 14,
Washington, D.C.
Han Qiwei, et al., 1981: Initial unit weight of reservoir deposits.
Journal of Sediment Research, Volume 1, Beijing.
Havinga, H., 1981: Bed-load determination by dune tracking,
ISO/TC113/SC6NI55 draft.
Huang Guanhua, et al., 1983: Study of bed-load measurement
by direct sampling in the Yangtze River. Proceedings of the
Second International Symposium on River Sedimentation,
Nanjing.
Hubbell, D.W., 1964: Apparatus and techniques for measuring
bed-load. Water Supply Paper 1748, USGS.
Hubbell. D.W., 1981: Recent refinements in calibrating bed-load
samplers. Water Forum, Volume 1, ASCE, New York.
Indian Standard 4890, 1966: Method for Measurement for
Suspended Sediment in Open Channels.
Indian Standard 6339, 1971: Particle Size Distribution and
Specific Gravity of Sediment in Streams and Canals.
Inland Waters Directorate, 1977b: SEDEX System Operations
Manual. Water Survey of Canada.
Inland Waters Directorate, 1979b: Automatic Suspended Sediment
Pump Sampler. Water Survey of Canada.
IAHS, 1981: Proceedings of the Florence Symposium on Erosion
and Sediment Transport Measurement, IAHS.
International Organization for Standardization, 1977a: ISO
Standard 3716: Requirements for Suspended Sediment
Samplers.
International Organization for Standardization, 1977b: ISO
Standard 4363: Method for Measurement of Suspended
Sediment.
International Organization for Standardization, 1977c: ISO
Standard 4364: Liquid Flow Measurement in Open
Channels — Bed Material Sampling.
International Organization for Standardization, 1981a: Technical
Report on Methods for Measurement of Bed-load Discharge.
ISO-CI13-SC6NI59.
151
International Organization for Standardization, 1981b: Technical
Report on Methods of Sampling and Analysis of Gravel Bed
Material. ISO FFXC I 13/SC6N 152.
International Organization for Standardization, 1982a: Draft
Standard ISO/DIS 4365: Determination of Concentration,
Particle Size Distribution and Relative Density.
International Organization for Standardization, 1982b: Draft
Standard ISO/DIS 6421: Methods for Measurement of
Sediment Accumulation in Reservoirs.
Jansen, P., et al., 1979: Principles of River Engineering. Pitman,
London.
Jansen, R.H.J., 1978: The in situ measurement of sediment transport by means of ultra-scattering. Publication 203, Delft
Hydraulics Laboratory.
Leeks, Wassand, 1999: Personal correspondence.
Leopold, L.B. and W.W. Emmett, 1997: Bed load and river
hydraulics — Inferences from East Fork River, Wyoming.
Professional Paper 1583, USGS.
Liang Guoting 1999: User’s Manual for RGTOOLS. Institute of
Hydraulic Research, Yellow River Conservancy Commission
(YRCC).
Li Songhen and Long Yuqian, 1994: Methods of modification of
the observed sediment load and its application. Journal of
Sediment Research, Volume 3, Beijing.
Li Zhaonan, 1980: Reduced Ranges for Reservoir Sedimentation
Survey. YRCC.
Lin Binwen, 1981: Application of Einstein's bed-load function in
the computation of total sediment discharge in the Yellow
River. Interim Report, YRCC.
Lin Binwen, 1982: Major Causes of the Deviations in Evaluation
of Quantity of Sedimentation by Range Method and
Difference of Sediment Load Method.
Lin Binwen and Liang Guoting, 1997: Corrections of the
measured suspended sediment discharge in sand-bed
streams. Proceedings of the Scientific Exchange, Beijing.
Long Yuqian, et al., 1978: Photo-sedimentation size analysis for
fluvial sediment. Journal of Peoples’ Yellow River, Yellow
River Press.
Long Yuqian and Lin Bingwen, et al., 1982: A preliminary analysis of the error in the measurement of suspended sediment.
Journal of Sediment Research, Volume 4, Beijing.
Long Yuqian, W.W. Emmett and R.J. Janda, 1989: Comparison of
some methods for particle size analysis of suspended sediment samples. Proceedings of the Fourth International
Symposium on River Sedimentation, Beijing.
Long Yuqian and C.F. Nordin, 1989: Intercomparison of collapsible bag suspended sediment samplers. Proceedings of the
Fourth International Symposium on River Sedimentation,
Beijing.
Lu Yongsheng, et al., 1980: Application of photo-sedimentation
technique in size analysis for fluvial sediment. Proceedings
of the First International Symposium on River
Sedimentation, Guanghua Press, Beijing.
Lu Yongsheng, et al., 1983: Automatic particle size analysis of
fluvial sediment by photo-electric techniques. Proceedings
of the Second International Symposium on River
Sedimentation, Water Resources Press, China.
Lu Yongsheng, 1995: Brief introduction on the intercomparison
of size analysis instruments. Journal of Hydrology,
Beijing.
152
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
Lu Zhi, et al., 1981: Development of nuclear gauge for use on
Yellow River. Proceedings of the Florence Symposium,
IAHS, Washington, D.C.
Ma Jingsong and Zhao Yunzhi, 1994: Development and application of computer-aided measurement of sediment
concentration. Proceedings of the First International
Symposium on Hydraulic Measurement, China Research
Institute of Water Conservancy and Hydropower, Beijing.
Mimikou, M., 1982: An investigation of suspended sediment
rating curves in western and northern Greece. Journal of
Hydrological Sciences 27, IAHS.
Ministry of Water Resources, 1975a: Manuals on Hydrometric
Measurement. Beijing.
Ministry of Water Resources, 1975b: Tentative Standards for
Hydrometric Measurement. Beijing.
Ministry of Water Resources, 1978: Tentative Standards for
Reservoir Sedimentation Survey. Beijing.
Niu Zhan, 1990: Method of computation of daily sediment
discharge. Personal correspondence.
Niu Zhan, 1994: Assessment of the uncertainty of sediment measurement. Research Report, Bureau of Hydrology, YRCC.
Nordin, C.F. Jr., 1981: The sediment discharge of rivers — a
review. Proceedings of the Florence Symposium, IAHS.
Pemberton, E.L., 1972: Einstein’s bed-load function applied to
channel design and degradation. Sedimentation Symposium
to Honor H.A. Einstein, (ed.) H.W. Shen, Fort Collins,
Colorado.
Rakoczi, L., 1984: Guide to selection of appropriate technique and
equipment for measurement of river sediment transport.
WMO/HOMS Sequence for Acquisition and Processing of
River Sediment, Budapest.
Rijn, L.C. van, 1980: Measurement of sediment concentration in
tidal conditions. Publication 244, Delft Hydraulics
Laboratory.
Sanmenxia Reservoir Experiment Station, 1980: Optimistic
Density of Ranges for Sedimentation Survey.
Schroeder, K.B., and C.H. Hembree, 1956: Application of the modified Einstein procedure for computation of total sediment load.
Transactions, AGU, Volume 37, Washington, D.C.
Shandong Provincial Office of Hydrology, 1980: Bed-load Transport
Estimated from Reservoir Sedimentation, Shandong.
Singhal, H.S.S., 1981: Sediment sampling in rivers and canals.
Proceedings of the Florence Symposium, IAHS.
Skinner, J.V., et al., 1981: Test and Design of Automatic Fluvial
Suspended Sediment Samplers.
Soil Conservation Service, 1973: National Engineering
Handbook. Section 3, Sediment, Department of Agriculture,
Washington, D.C.
Stelczer, K., et al., 1984: Technique for indirect estimation of bed
sediment load. WMO/HOMS Sequence for Acquisition and
Processing of River Sediment, Budapest.
Stevens, H.H., et al., 1980: Collapsible-bag suspended sediment
sampler. Journal of the Hydraulics Division, ASCE (Hy4),
April, New York.
Stevens, H.H. Jr., 1985: Computer program for the computation of
total sediment discharge by the modified Einstein procedure.
Water Resources Investigations Report 85-4047, USGS.
Stichling, W., 1969: Instruments and techniques in sediment surveying. Reprint Series 22, Inland Waters Directorate, Canada.
Szalona, J.J., 1982: Development of a bag type suspended sediment sampler. Report Y, FIASP.
United Nationas Educational, Scientific and Cultural
Organization, 1982: Sedimentation problem in river basins.
Technical Report, IHP project.
Universities of Nottingham, Newcastle and Southampton, 1998:
River Geomorphology: A Practical Guide. National Centre
for Risk Analysis and Options Appraisal, Environment
Agency, United Kingdom.
US Geological Survey, 1976: Manuals on Techniques of Water
Resources Investigations. Book 3, Chapter C2, Washington, D.C.
US Geological Survey, 1977: Manuals on Techniques of Water
Resources Investigations. Book 5, Chapter C1, United States.
Vanoni, V.A., et al., 1975: Sedimentation Engineering. ASCE,
New York.
Walling, D.E., et al., 1981: The reliability of suspended sediment
load data. Proceedings of the Florence Symposium, IAHS.
Wang Shijiang, et al., 1995: New equation for sediment transport.
International Journal of Sediment Research, Volume 10,
Number 3, Beijing.
World Meteorological Organization, 1981: Measurement of River
Sediments (L. Rakoczi). Operational Hydrology Report
No. 16 (WMO-No. 561), Geneva.
World Meteorological Organization, 1994: Guide to Hydrological
Practices. Fifth Edition, WMO-No. 168, Geneva.
Xiang Zhian, 1980: Measurement of bed-load and data processing.
Changjiang Water Resources Commission (CWRC), China.
Xiang Zhian and Li Keqin, 1994: Experimental Study of the
Method of Correction of the Size Gradation Analysed by Size
Analyser Method. CWRC.
Xiong Guishu, et al., 1983: Analysis of errors in sediment
measurement in the Lower Yellow River. Proceedings of the
Second International Symposium on River Sedimentation,
Water Resources Press.
Xu Shengguo, et al., 1990: Intercomparison of Gravel Bed Load
Samplers at Tizhiyan Station on the Qingyijiang River.
Sichuan Provincial Bureau of Hydrology.
Yorke, T.H., 1976: Ten years of experience with automatic
pumping sediment sampler. Proceedings of the Third Federal
Interagency Sedimentation Conference, Sedimentation
Committee of the Water Resources Council, United States.
Zhang Kuotail, 1981: Application of Computer Technology in
Sediment Data Processing. YRCC.
Zhang Yuanfeng and Long Yuqian, 1996: Improvement of
Einstein’s bed load function. Journal of Sediment Research,
Volume 4.
Zhang Yuanfeng, Long Yuqian, 1998: Adaptability of sediment
transport formula to the Yellow River. Proceedings of the
International Symposium on River Sedimentation, Hong
Kong.
Zhou Dejia, et al., 1981: The development of a sand bed-load
sampler for the Yangtze River. Proceedings of the Florence
Symposium, IAHS.
Zhou Dejia, 1989: Methodology of bed load measurement. Series
of Technology in Hydrometric Measurement, CWRC.
CHAPTER 7
WATER QUALITY RELATED TO TRANSPORT OF SEDIMENT AND TOXIC MATERIAL
7.1
Table 7.1
Characteristic pH value of heavy metal at maximum
adsorption (mg/g)
Element
Zn
Co
Cu
Ni
Pb
pH
7.6
9.0
8.4
9.0
5.5
Max. adsorption
6.65
3.30
8.20
2.15
135.78
(b)
(c)
source material, sorting during transport, and physical
conditions at the point of deposition. Transportation occurs in
a similar fashion in both rivers and lakes, and is a direct
function of water movement. In rivers, water movement is
linear, whereas in lakes water movement is mainly orbital or
oscillatory due to the passage of wind-generated waves. In
lakes, wind stress also induces major water circulation
patterns involving low velocity currents, which influence the
transport directions of wave-perturbed sediment.
Particle-size fractions. The size range (diameter µ) of transported particles ranges upwards from the clay-sized material,
conventionally defined as (<4 µ ). This fraction consists
mostly of clay minerals such as montmorillonite and kaolinite etc., but may also include some other fine minerals and
organic debris. The silt fraction is medium-sized (4 µ–64 µ;
and the sand (2 mm–64 µ) and gravel (>2 mm) make up the
coarser size fraction. These limits are only conventional and
may change slightly from one scale to another. There is a
marked relationship between the particle size and its origin
(rock minerals, rock fragments and pollutants, etc.).
Grain-size influence. The specific surface area is a key particle property which controls adsorption capacity. It is
inversely proportional to particle size and decreases over
three orders of magnitude from clay-sized particles
(10 m2 g–1) to sand grains (0.01 m2 g–1). Therefore, the
finest particles are generally the richest in trace elements.
This effect is particularly evident when separate chemical
analyses are made on different size fractions, as shown for
Cu and particulate matter in the Fly River Basin, Papua New
Guinea (Figure 7.1). When the total particulate matter is
considered, the trace element content is usually directly
proportional to the amount of the finest fraction, as shown in
the Rhine River for the < 16 µm fraction.
1200
1000
Copper
(µg/g)
Copper
(ug/g)
EFFECTS OF SEDIMENT AND HEAVY METALS
ON WATER QUALITY
7.1.1
Absorption of heavy metals in sediment particles
The absorption of heavy metals in sediment particles depends not
only on sediment composition and properties, chemical properties
and forms of the heavy metals, but also on the variety of environmental factors in the body of water. The factors affecting sediment
adsorption include pH, temperature, ionic capacity, adsorbent
concentration, oxido-reduction potential and particle size, etc.
(1) Effect of temperature. Temperature is one of the
important factors relating to how sediment affects adsorption on
metal. For both adsorbent and adsorbate, the adsorptive temperature and the type of adsorbate determine adsorptive capacity.
The study of adsorptive isobars shows that a quantitative relationship exists between temperature and adsorptive capacity.
Since physical adsorption and chemical adsorption are exothermic reactions, adsorptive capacity generally drops when the
temperature rises. Because physical adsorption is a fast process,
a balance is quickly found, and adsorptive capacity drops as the
temperature rises in the experiment. The speed of chemical
adsorption is low and rise in temperature speeds up the adsorption process. It thus appears that adsorptive capacity increases
when the temperature rises.
(2) Effect of pH. The pH value is one of the most
important factors in the adsorption process of metal. The effect
relates to the solubility of metal, the surface adsorptive characteristics of sediment, and the sorption reaction of metal on the
surface of sediment.
In general, the adsorptive capacities of metal on sediment particles, soil and suspended solids increase with the
increase of pH. Heavy metals appear to have the most adsorptive
capacity on sediment at its characteristic pH value. Table 7.1
shows a study of the adsorption of heavy metal on particles in the
Jinsha River, in the upper reach of the Yangtze River, China.
(3) Effect of particle size. Heavy metals in water can
be adsorbed by sediment; its adsorptive capacity for the heavy
metals is firstly determined by the particle size. According to
Fendler’s rule, the smaller the particle size, the greater the
adsorptive capacity. Particle size greatly affects the distribution
of heavy metal. Heavy metals exist on sediment particles finer
than 0.025 mm.
(a) Transport and deposition. As noted previously, sediment can
be defined in terms of particle size and mineralogical
composition, both of which are inter-related. The chemical
composition of the sediment at its point of deposition is a
product of the composition of the source material, size of the
Ok Tedi
Upper Fly
Middle Fly
Lower Fly
Strickland
800
600
400
200
0
<2
2–20
20–63
>63
Grain-size fraction (µm)
Figure 7.1 — Copper in various grain-size fractions in the Fly River
Basin, Papua New Guinea.
154
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
Effects of sediment particles absorbing heavy metals
on water quality
Sediment affects water environment and water quality considerably. The effect has dual characteristics. From one aspect,
sediment is a chief pollutant which comes from a non-point
source, causing physical, chemical and biological pollution on the
water body. It seriously affects the water quality and the aquatic
ecological environment.
From another aspect, on condition that the river has a
specific hydrochemistry, a high sand content and specific sediment
physical and chemical characteristics, many kinds of pollutants,
including heavy metals and toxic organic materials that enter the
water body from sewage, can be adsorbed by sediment. The
amount and intensity of its adsorption is closely related to the
sediment content and size. Sediment adsorption is the process in
which pollutants are divided between water and sand. The result
of sediment adsorbing pollutants is an improvement in water
quality once they are filtered out of the sediment. Therefore, such
adsorption reduces the concentration of pollutants in the water and
improves toxicity and the process of removing and transforming
adsorbed pollutants in the water phase. This process is controlled
by content, time-space distribution and the partition of sediment
particle size in the water.
(1) Case study 1: Sediment and water quality of the
Klagan River in the tropical rainforest of Sabah, Borneo Island. The
study of sediment and water quality was carried out on the Klagan
River, a tributary of Labuk River in Sabah, north-east Borneo. The
river courses through an uneven terrain largely composed of sandstone, limestone, and basalt. The study was designed to gather
information on the water-quality and sediment characteristics of the
above-described riverine ecosystem. Water quality is affected by
sediment and the nature of the rocks of the area.
A number of hydrological parameters of the Klagan
River show wide variation. The data reveal a considerable degree
of erosion of the river banks. River bank erosion is mainly responsible for the increase of the concentration of suspended solids to a
level as high as 328 mg/l.
Phosphate content appears to be linked to the release of
this chemical from sediment under certain conditions of temperature,
anaerobic activity and pH. Desorption of phosphate from ferric
hydroxide at high pH is a distinct possibility. Water content of sulfate
varies with the salinity. Obviously, it is high in the lower reaches.
Regarding the heavy metals in water, Cd, Mn, and Zn
are detectable, whereas Cr, Cu, and Pb are not. Co and Ni
occurred in the first sampling at station KL7. Cd may represent
the sediment-water exchangeable fraction.
Concentrations of heavy metals in the sediment follow
the order: Mn > Ni > Cr > Zn > Cu > Co > Pb > Cd. Except Cd,
Cu, and Zn, which are relatively constant, the remaining metals
decrease in concentration from the upper to the lower reaches of
the Klagan River. Cd occurs dominantly as a water-exchangeable
fraction, and also appears to originate from carbonate compounds.
Co, Cr, Mn, Ni, Pb, and Zn are mainly in the lithogenous fraction,
and have low solubility. The non-lithogenous fraction accounts for
less than 20 per cent of the detectable Cu, which is mainly linked
to the organic fraction. Pb is dominant in the lithogenous fraction.
High Co content is attributed to ultrabasic rocks in the region. Mn
is chiefly found in exchangeable and iron-manganese
oxide/hydroxide fractions. Relatively larger quantities of Ni in the
sediment are derived from basaltic rocks.
(2) Case study 2: Background concentration of heavy
metals in some rivers. A study of background concentration of
heavy metal was carried out for the Yellow River. The comparisons of background of heavy metal in the middle reach of the
Yellow River and in other basins in China and other countries are
shown in Tables 7.2 and 7.3.
(3) Discussion on absorption and flocculation.
Flocculation, because it alters the hydrodynamic properties of
particles in transport, significantly influences the fate and effect of
sediment and associated contaminants. It was found that the
complex structure and composition of a floc would have a significant effect on its physical, chemical and biological behaviour. An
important observation was the apparent structural dominance of
the fibril extra cellular polymeric material within freshwater flocs.
These fibrils are believed to be the dominant material for the
development and stabilization of flocculated material. Each
general component of a floc (organic and inorganic particles, plus
water and pores) is diverse and can possess a specific function
within a floc. The interactions between these constituents and their
functional processes can result in the modification of a floc’s
behaviour; how it is physically transported and settled, how it
adsorbs and transforms contaminants and nutrients chemically,
and biologically, how it develops a diverse microhabitat capable of
Table 7.2
Concentration of heavy metals in filtered water in some
rivers (µg kg–1)
Table 7.3
Concentration of heavy metals in bottom sediment in some
rivers (mg kg–1)
(4) Effect of sediment concentration on the adsorption
of heavy metal. The adsorption capacity of sediment is obviously
affected by the sediment concentration. There is a negative relationship between them. The lower the concentration of sediment,
the more marked the enrichment action of sediment, and the
greater the sediment adsorptive capacity. The total adsorptive
capacity of sediment to heavy metal increases quickly as the sediment concentration increases.
7.1.2
Item
Cu
Pb
Zn
Ni
Cr
14.3
Yellow River
6.89
12.8
40.9
22.7
18.6
1.6
Lakes of the world
43
28
110
66
59
10.0
1.6
Xiangjiang River
13
22
59
32
37
10.0
0.5
No. 2 Songhua River
17.7
24
119
22
17.3
South lake in
Changchun
38.0
13.8
69.6
25.8
8
45
34
118
Item
Cu
Pb
Zn
Ni
Cr
Yellow River
3.14
5.25
117.9
1.36
Xiangjiang River
4.0
5.0
7.0
Rivers of the world
5.0
3.0
Surface fresh water
1.8
0.2
No. 2 Songhua River
2.6
2.5
6.9
Changbaishan
Tianchi Lake
4.3
13.8
11.5
9
3
0.95
Non-polluted sediment
62
CHAPTER 7 — WATER QUALITY RELATED TO TRANSPORT OF SEDIMENT AND TOXIC MATERIAL
Inorganic
Biota and
and
Biota
Bioorgani
Bioorganic
Active
Contamination
transformation
microbial growth
exopolymers
anaerobic/aerobic
processes
modifying its structural, chemical and biological make-up. These
interactions and functions are summarized in Figure 7.2.
Water
Bound
Inactive
Free
7.2
EFFECTS OF SEDIMENT AND TOXIC ORGANIC
MATERIAL ON WATER QUALITY
7.2.1
Absorption of toxic organic material on sediment
particles
(1) Resolvability of toxic organic material. The resolvability of toxic organic material in water is closely related to the
partition coefficient of soil/sediment (Koc), the biological concentration coefficient (BCF), the partition coefficient of octanol-water
(Kow) and the rate of degradation of carcinogenic action.
Therefore, the solubility of the organic pollutant in water is an
important assessment parameter that forecasts its harmfulness
with regard to the environment. The major environmental parameters of organic compounds are shown in Table 7.4.
(2) Absorption of toxic organic material in sediment
particles. Most of the toxic organic compound, which is difficult to
degrade and easily adsorbed in sediment and layers of biological
fat, represents an accumulated and long-term toxic danger to
biology and the environment.
Nutrient and metabolic
and product transport
electrochemical and
diffusional gradients
Floc building
hydrodynamic
chemical and
biological
behaviour
155
Colonization sites
cation bridging contaminant adsorption/
desorption
Figure 7.2 — Conceptual model of floc form and function.
Table 7.4
The major environmental parameter of organic compound
Name of compound
Acrolein
Acrylonitrile
Benzene
Benzidine
Chlorobenzene
1,2,4-Trichlorobenzene
Hexachlorobenzene
1,2-Dichloroethane
1,1,1-Trichloroethane
Hexachloroethane
1,1,2-Trichloroethane
1,1,2,2-Tetrachloroethane
Chloroethane
2-Chloronaphthalene
1,2-Dichlorobenzene
1,3-Dichlorobenzene
1,4-Dichlorobenzene
3,3’-Dichlorobenzidine
1,1-Dichloroethylene
Trans-1,2-Dichloroethylene
1,2-Dichloropropane
Trans-1,3-Dichloropropene
2,4-Dinitrotoluene
2,6-Dinitrotoluene
Fluor-anthene
1,2-Diphenylhydrazine
Ethylbenzene
4-Chlorophenylphenyl ether
4-Bromophenylphenyl ether
Bis(2-Chloroethoxy)methane
Methylene chloride
Methyl chloride
Methyl bromide
Dichlorodifluoromethane
S
Kow
Koc
KB
Hc
Pv
BCF
2.1E5(20µ)
7.9E4(25µ)
1.78E3(25µ)
400(120µ)
488(25µ)
30(25µ)
6E-8(25µ)
5.5E3(20µ)
720(25µ)
50(22µ)
4.5E3(20µ)
2.9E3(20µ)
5.74E3(20µ)
6.74(25µ)
100(20µ)
123(25µ)
79(25µ)
4.0(22µ)
400(20µ)
600(20µ)
2.7E-3
2.7E3(25µ)
270(22µ)
180(20µ)
0.26(25µ)
1.84E3
152(20µ)
3.3(25µ)
4.8(25µ)
8.1E4(25µ)
2.0E4(20µ)
6.45E3(20µ)
900(20µ)
280(25µ)
1.02
1.78
135
21.9
690
1.9E4
2.6E6
63
320
4.2E4
117
245
30.9
1.0E4
3.6E3
3.6E3
3.6E3
3.236E3
135
123
105
100
95
190
7.9E4
871
2.2E3
1.2E5
8.7E4
10.7
18.2
8.9
12.3
120
0.49
0.85
65
10.5
330
9.2E3
1.2E6
30
152
2.0E4
56
118
14.9
4.8E3
1.7E3
1.7E3
1.7E3
1553
65
59
51
48
45
92
3.8E4
418
1.1E3
5.8E4
4.2E4
5.2
8.8
4.3
5.9
58
0.44
1.04
37
10.1
164
3.3E3
2.9E5
19
81
6.75E3
33
91
9.8
1.8E3
730
730
730
941
53
48
30
40
39
51
1.2E4
286
470
1.8E4
1.3E4
3.7
6.0
3.2
4.2
33
5.66E-5
8.8E-5
5.5E-3
3E-7
3.58E-3
2.3E-3
6.8E-4
4.26E-3
0.03
2.49E-3
7.42E-4
3.8E-4
0.148
5.4E-4
1.93E-3
3.61E-3
3.1E-3
8E-7
0.190
0.067
2.3E-3
1.33E-8
4.5E-6
7.9E-6
6.5E-6
3.4E-9
6.6E-3
2.19E-4
1.0E-4
2.8E-7
2.03E-3
0.04
0.197
2.98
220(20µ)
100(23µ)
95.2(25µ)
5E-4
11.7(20µ)
0.29(25µ)
1.09E-5(20µ)
180(20µ)
123(25µ)
0.4(20µ)
19(20µ)
5(20µ)
1E3(20µ)
0.017(20µ)
1.0(20µ)
2.28(25µ)
1.18(25µ)
1E-5(22µ)
591(25µ)
326(20µ)
42(20µ)
25(20µ)
5.1E-3(20µ)
0.018(20µ)
5E-6(25µ)
2.6E-5(25µ)
7(20µ)
2.7E-3
1.5E-3(20µ)
<0.1(20µ)
362.4(20µ)
3.76E3(20µ)
1.42E8(20µ)
4.87E3(25µ)
4.38
7.2
352.5
68.7
1.5E3
3.0E5
2.5E6
177.7
765.8
6.1E4
309.96
6.0E2
93.6
1.7E4
6.7E3
6.7E3
6.7E3
6.1E3
3.5E2
3.2E2
2.8E2
2.7E2
2.6E2
4.8E2
1.1E5
1.9E3
4.3E3
1.6E5
1.2E5
36.1
58.2
30.6
40.9
3.2E2
156
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
Table 7.4 (cont’d)
Name of compound
Trichlorofluoromethane
Isophorone
Hexachlorobutadiene
Hexachlorocyclopentadiene
Naphthalene
Nitrobenzene
2-Nitrophenol
4-Nitrophenol
2,4-Dinitrophenol
Benzo(a)anthracene
Benzo(b)fluoroanthene
Benzo(k)fluoroanthene
Benzo(g.h.i)perylene
Benzo(a)pyrene
Chrysene
Dibenzo(a,h)anthracene
Indeno(1,2,3-cd)pyrene
Fluorene
Vinylchloride
Trichloroethylene
Tetrachloroethylene
Toluene
Phenanthrene
Pyrene
Dieldrin
Chlordane
Aldrin
Alpha-Endosulfan
Beta-Endosulfan
Endosulfan sulfate
Endrin
Edrin aldehyde
Heptachlor
Heptachlor epoxide
Alpha-BHC
Beta-BHC
Delta-BHC
Gamma-BHC
PCB-1016
PCB-1221
PCB-1232
PCB-1242
PCB-1248
PCB-1254
PCB-1260
Toxaphene
Dimethyl phthalate
Diethyl phthalate
Di-n-butyl phthalate
Di-n-octyl phthalate
Bis(2-Ethylhexyl)phthalate
ButylBenzylphthalate
s:
Koc:
Kow:
KB:
S
Kow
Koc
KB
Hc
Pv
BCF
1.1E3(20µ)
1.2E4
2.0(20µ)
1.8(25µ)
31.7(25µ)
1.9E3(20µ)
2.1E3(20µ)
1.6E4(25µ)
290(25µ)
(20µ)
0.014(25µ)
4.3E-3(25µ)
2.6E-4(25µ)
2.8E-3(25µ)
1.8E-3(25µ)
5E-4(25µ)
5.3E-4(25µ)
1.69(25µ)
2.7E3(25µ)
1.1E3(20µ)
200(20µ)
534.8(25µ)
1.00(25µ)
0.13(25µ)
0.195(25µ)
0.056(25µ)
0.180(25µ)
0.53(25µ)
0.28(25µ)
0.22
0.25(25µ)
50(25µ)
0.18(25µ)
450(25µ)
1.63(25µ)
0.24(25µ)
31.4(25µ)
7.8(25µ)
0.42(25µ)
40.0(25µ)
407(25µ)
0.23(25µ)
0.054(25µ)
0.031(25µ)
(25µ)
0.50(25µ)
5.0E3(20µ)
896(25µ)
13(25µ)
3.0(25µ)
0.4(25µ)
2.9
331
180
6.0E4
1.0E4
1.95E3
74
56
93
500
4.1E5
1.15E6
1.15E6
3.2E6
1.15E6
4.1E5
6.9E6
3.2E6
1.5E4
17.0
263
759
620
2.8E4
8.0E4
3.5E3
3E5
2E5
0.02
0.02
0.05
3.5E3
1.4E3
2.6E4
2.2E2
7.8E3
7.8E3
1.4E4
7.8E3
3.8E5
1.2E4
1.6E3
1.3E4
5.75E5
1.1E6
1.4E7
2E3
3.63
295
3.6E5
7.4E9
4.1E9
3.6E6
159
87
2.9E4
4.8E3
940
36
27
35
240
2.0E5
5.5E5
5.5E5
1.6E6
5.5E6
2.0E5
3.3E6
1.6E6
7.3E3
8.2
126
364
300
1.4E4
3.8E4
1.7E3
1.4E5
9.6E4
9.6E-3
9.6E-3
0.024
1.7E3
670
1.2E4
1.1E2
3.8E3
3.8E8
6.6E3
3.8E3
1.8E5
5.8E3
771
6.3E3
2.77E5
5.3E5
6.7E6
964
17.4
142
1.7E5
3.6E9
2.0E9
1.7E3
84
48
1.3E4
1.8E4
420
22
17
27
122
5.3E4
1.4E5
1.4E5
3.5E5
1.4E5
5.3E4
6.9E5
3.5E5
3.8E3
5.7
97
252
148
4.7E3
1.2E4
710
4E4
2.8E4
0.012
0.012
0.029
710
310
4.4E3
3.9E-4
1.5E3
1.5E3
3.5E3
1.5E3
5.0E4
2.2E3
351
2.3E3
7.29E4
1.3E5
1.3E8
429
16.0
107
4.7E4
3.9E8
2.3E8
5.7E4
0.11
5.75E-6
0.0256
0.016
4.6E4
1.31E-5
7.56E-6
2.5E-5
4E-5
1E-6
1.22E-5
3.87E-5
1.44E-7
4.9E-7
1.05E-6
7.3E-8
6.95E-8
6.4E-5
8.14E-2
9.1E-3
0.0154
6.66E-3
2.26E-4
5.1E-6
4.57E-10
9.4E-5
1.6E-5
1E-5
1E-5
2.6E-5
4E-7
2E-9
4.0E-3
3E-4
6.0E-6
4.5E-7
2.07E-7
7.8E-6
3.3E-4
1.7E-4
1.13E-5
1.98E-3
3.6E-3
2.6E-3
0.74
0.21
2.15E-6
1.2E-6
2.8E-7
1.7E-5
3E-7
8.3E-6
667.4(20µ)
0.38(20µ)
0.15(20µ)
0.081(20µ)
0.087(25µ)
0.15(20µ)
0.151(20µ)
2.2(46µ)
5E-2(20µ)
2.2E-8(20µ)
5E-7
5E-7
1.03E-10
5.6E-9(25°C)
6.3E-9(25µ)
1E-10(20µ)
1E-10(20µ)
7.1E-4
2.66E3(25µ)
57.9(20µ)
14(20µ)
28.7(20µ)
9.6E-4(25µ)
2.5E-6(20µ)
1.78E-7(20µ)
1E-5(25µ)
6E-6(25µ)
1E-5(25µ)
1.9E-5(25µ)
1E-5(25µ)
2E-7(25µ)
2E-7(25µ)
3E-4(25µ)
3E-12
2.5E-5(20µ)
2.8E-7(20µ)
1.7E-5(20µ)
1.6E-4(20µ)
4E-4(25µ)
6.7E-3(25µ)
4.06E-3(25µ)
1.3E-3(25µ)
4.94E-4(25µ)
7.71E-5(25µ)
4.05E-5(25µ)
0.2-0.4(20µ)
4.19E-3(20µ)
3.5E-3(25µ)
1E-5(25µ)
1.4E-4(25µ)
2E-7(20µ)
6E-5
7.9E2
4.6E2
8.5E4
1.6E4
3.9E3
2.1E2
1.6E2
2.5E2
1.1E3
4.7E5
1.2E6
1.2E6
3.0E6
1.2E6
4.8E5
6.0E6
3.0E6
2.4E4
54.7
6.4E2
1.7E3
1.4E3
4.2E4
1.1E5
6.6E3
3.6E5
2.5E5
0.128
0.128
0.29
6.6E3
2.9E3
3.9E4
Resolvability in water (ppm)
Partition coefficient of soil/sediment
Partition coefficient of octanol-water
Partition coefficient of microbe-water (µg/g) (mg/L)
BCF:
Hc:
Pv:
Biological concentration coefficient
Constant of Henry (torr/mor)
Press of vapour (torr)
1.4E4
1.4E4
2.3E4
1.4E4
4.4E5
1.99E4
3.3E3
2.1E4
6.5E5
1.2E6
1.1E7
3.9E3
13.7
7.1E2
4.2E5
3.2E9
1.9E9
3.4E6
CHAPTER 7 — WATER QUALITY RELATED TO TRANSPORT OF SEDIMENT AND TOXIC MATERIAL
Table 7.5
The characteristics of adsorption and partition
Adsorption
Partition
High adsorption thermal
Low adsorption thermal
Non-linear isotherm
Linear isotherm
Competitive adsorption
Non-competitive adsorption
The comparison of partition and adsorption. The
adsorption mechanism of organic compounds in a water sediment
system is partition, and the adsorption mechanism of metals is
adsorption. Their reactions present differences in the mechanisms.
The differences between them relate to action, reaction thermal,
type of sorption isothermal formula and sorption competitiveness
(see Table 7.5).
The remaining concentration of organic compounds in
river sediment. The adsorption process is determined by the
hydrophile or hydrophobe nature of the compound and the composition of the adsorbent, of which the decisive factors are solubility,
partition coefficient of octanol-water and the organic carbon
content of adsorbent.
The remaining concentration of organic compounds
determined by experiments in river sediment is shown in
Table 7.6.
7.2.2
Effects of sediment particles absorbing toxic organic
material on water quality
As mentioned in section 7.1.2, the effects of sediment particles on
water quality are considerable and present a characteristic of a
dual nature. Toxic organic material can be adsorbed and kept by
sediment for a long time, representing the sediment-water
exchange. An example of this is the following case study: In Italy,
the use of pp’-DDT and lindane was forbidden in the 1970s, and
the application of lindane is currently restricted to agricultural use
and the application of pp’-DDT to floriculture. The presence of
pp’-DDT metabolites indicates that the pesticide is no longer used
in the catchment basin, and that DDT contamination is due to the
past usage of this pesticide.
7.3
WATER QUALITY MODEL OF SEDIMENT AND
TOXIC ORGANIC MATERIAL AND HEAVY
METAL
Water quality models are designed to simulate the responses of
aquatic ecosystems under varying conditions. They have been
applied to help explain and predict the effects of human activities
on water resources, such as lake eutrophication, dissolved oxygen
concentrations in rivers, the impacts of acid rain on natural water
bodies, and the fate, pathways, impacts and effects of toxic
substances in freshwater systems. Mathematical models are very
useful tools for water quality management because they allow:
(1) The identification of important variables in a particular
aquatic system, and help interpretation of the system’s
processes;
(2) Forecasting of the impacts of developments on water bodies;
and
(3) Policy testing and analysis.
The high degree of complexity, spatial and functional
heterogeneity, non-linearity, complex behavioural features (such
as adaptation and self-organization) and the considerable stochastic element of natural systems make model development a difficult
and highly skilled task. Data requirements for model calibration
and for model use pose additional constraints on their widespread
use. This complexity, and the limited knowledge of the processes
taking place in rivers and lakes, requires that a high degree of
simplification and a number of assumptions be built into any
model. The model user must be aware of the model’s limitations
and its assumptions in order to draw appropriate conclusions. At
present, highly predictive models are not general and general
models are not highly predictive.
Model types: Mathematical models belong to one of two
basic classes, namely theoretical (or deterministic) and empirical.
Theoretical models: If the physical, chemical and/or
biological mechanisms underlying a process are well understood,
a steady-state or dynamic model can be developed. Steady-state
models cannot be used for predicting system responses over time,
and they therefore have limited water management value. Timevariable models, on the other hand, can handle variable input
loads, and can be useful for establishing cause-effect relationships.
Table 7.6
Adsorption coefficient of PCBs and organic chloride pesticide
Compound
Soil/sediment
OC/OMµ%µ
Kd
2,2,4’-PCB
2,5,2’-PCB
Hexachlorinatedbiphenyls
Hexachlorinatedbiphenyls
Aroclor 1254
DDT
DDT
P,P’-DDE
µ-BHC(Lindane)
µ-BHC
µ–Chlordane
µ–Chlordane
Endrin
Kepone
Sandy soil
Suspended sediment of river
Sediment of Lake Michigan
Suspended sediment of river
Sediment of Lake Michigan
Sediment of ocean
Soil sample
Suspended sediment of river
Sandy soil
Creek sediment
Suspended sediment of river
Suspended sediment of river
Sand
Sediment of bay
1.9(om)
4.1(oc)
2.9(oc)
4.1(oc)
1.7(oc)
2.7(oc)
(om)
4.1(oc)
1.9(om)
2.8(om)
4.1(oc)
4.1(oc)
0.7(om)
1 700
460
10 000
9 000
13 000
7 000
48 000
140 000
41 000
14
24
13 000
1 000
58
oc: Organic carbon
157
om: Organic material
Koc/Kom
24 000
250 000
310 000
300 000
410 000
1 800 000
1 000 000
740
860
300 000
250 000
8 300
158
MANUAL ON SEDIMENT MANAGEMENT AND MEASUREMENT
When compared to empirical models, theoretical models are
generally more complex. They require a longer period of observation for calibration, and the number of variables and parameters to
be measured are greater. They also require a significant amount of
time for validation. Owing to their complexity, and because our
understanding of aquatic systems is usually incomplete, these
types of models are used less frequently than empirical models.
Empirical models: Empirical or statistically-based
models are generated from the data analysis of surveys at specific
sites. The relationships thus identified are then described in one or
more mathematical equations. These models can be built relatively
quickly when compared with theoretical models, and they are
easier to use because they have fewer data requirements.
Sometimes empirical models have to be generated from incomplete or scattered information about the aquatic system. For
example, the model may be supported by observations made over
a limited range of conditions or a relatively short time period. In
such cases, the model output should be interpreted with caution. It
is also important to remember that such models are not directly
transferable to other geographic areas or to different time scales.
Examples of water quality models: Hundreds of water
quality models have been developed. Some of them are specific to
a given site or problem, while others are more general, such as
multimedia models. There is no single model that can be applied
to all situations. Some examples of models are described below.
Water Analysis Simulation Programme (WASP): This
theoretical model is applicable to a wide variety of water quality
problems, and can be adapted for site-specific uses. It is a timevariable model that can be applied to one, two or three
dimensions. The input data consist of loads, boundary conditions,
mass transfer rate, kinetic rates and concentrations of organic
compounds, trace elements and phytoplankton. The output lists
variable concentrations.
REFERENCES
Chapman, D., 1992: Water Quality Assessments. Chapman and
Hall, London.
Evans, R.D., J.R. Wisniewski, and J. Wisniewski, 1997: The interactions between sediments and water. Proceedings of the
Seventh International Symposium, Baveno, Italy,
22–25 September 1996, Kluwer Academic Publishers.
Jin Xiangcan, 1990: Pollution Chemistry of Organic Compounds.
Qinghua University Publishe.
Jin Xiangcan, 1992: Pollution Chemistry of Sediment.
Environmental Science Publisher.
Osterkamp, W.R., 1995: Effects of Scale on Interpretation and
Management of Sediment and Water Quality. IAHS.
Zhao Peilun, 1998: The Effect of Sediment on Water Quality of the
Yellow River and Control of Water Pollution in Major Rivers.
Yellow River Hydropower Publisher.
Table 7.7
Main characteristics of the sampling stations and sediment samples
Sampling station
Stations and
samples
Subbasin
51 (0–10 cm)
51 (10–20 cm)
51 (38–48 cm)
52
53
54
43
56
45 (0–10 cm)
45 (10–20 cm)
45 (32–42 cm)
32A
13 (0–10 cm)
13 (10–20 cm)
13 (50–60 cm
1A
North
North
North
North
North
North
North
North
North
North
North
Central
South
South
South
South
Depth of
station (m)
71
71
71
33
19
30
89
58
143
143
143
90
47
47
47
11
Sediment samples
Type of
sample
Date of
samples (%)
Water
(%)
Organic
carbon
Core
Core
Core
Grab
Grab
Grab
Grab
Grab
Core
Core
Core
Grab
Core
Core
Core
Grab
1970–1992
1948–1970
1885–1970
1970–1992
1970–1992
1970–1992
1970–1992
1970–1992
1970–1992
1948–1970
1900–1922
1961–1992
1950–1992
1908–1950
1742–1784
1950–1992
65.4
64.4
58.5
59.6
75.1
67.0
39.0
42.5
25.0
31.0
24.0
42.2
49.0
41.0
28.0
74.6
9.96
6.15
4.15
4.87
11.65
9.84
7.44
4.53
3.20
4.11
3.20
5.83
6.84
4.78
5.80
4.97
OPERATIONAL HYDROLOGY REPORTS
WMO
No.
332
337
341
356
419
425
429
433
461
464
476
513
519
559
560
561
576
577
580
587
589
635
646
650
655
658
680
683
686
704
705
717
718
740
749
773
754
779
804
803
806
813
884
885
886
886
Report
No.
No. 1
No. 2
No. 3
No. 4
No. 5
No. 6
No. 7
No. 8
No. 9
No. 10
No. 11
No. 12
No. 13
No. 14
No. 15
No. 16
No. 17
No. 18
No. 19
No. 20
No. 21
No. 22
No. 23
No. 24
No. 25
No. 26
No. 27
No. 28
No. 29
No. 30
No. 31
No. 32
No. 33
No. 34
No. 35
No. 36
No. 37
No. 38
No. 39
No. 40
No. 41
No. 42
No. 43
No. 44
No. 45
No. 46
* Out of print
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
Manual for estimation of probable maximum precipitation (Second edition)
Automatic collection and transmission of hydrological observations*
Benefit and cost analysis of hydrological forecasts. A state-of-the-art report
Applications of hydrology to water resources management*
Meteorological and hydrological data required in planning the development of water resources*
Hydrological forecasting practices*
Intercomparison of conceptual models used in operational hydrological forecasting
Hydrological network design and information transfer
Casebook of examples of organization and operation of Hydrological Services
Statistical information on activities in operational hydrology*
Hydrological application of atmospheric vapour-flux analyses*
Applications of remote sensing to hydrology
Manual on stream gauging — Volume 1: Field work — Volume 2: Computation of discharge
Hydrological data transmission
Selection of distribution types for extremes of precipitation
Measurement of river sediments
Case studies of national hydrological data banks (planning, development and organization)
Flash flood forecasting
Concepts and techniques in hydrological network design
Long-range water-supply forecasting
Methods of correction for systematic error in point precipitation measurement for operational use
Casebook on operational assessment of areal evaporation
Intercomparison of models of snowmelt runoff
Level and discharge measurements under difficult conditions
Tropical hydrology
Methods of measurement and estimation of discharges at hydraulic structures
Manual on water-quality monitoring
Hydrological information referral service — INFOHYDRO Manual
Manual on operational methods for the measurement of sediment transport
Hydrological aspects of combined effects of storm surges and heavy rainfall on river flow
Management of groundwater observation programmes
Cost-benefit assessment techniques and user requirements for hydrological data
Statistical distributions for flood frequency analysis
Hydrological models for water-resources system design and operation
Snow cover measurements and areal assessment of precipitation and soil moisture
Remote sensing for hydrology — Progress and prospects
Hydrological aspects of accidental pollution of water bodies
Simulated real-time intercomparison of hydrological models
Applications of remote sensing by satellite, radar and other methods to hydrology
Land surface processes in large-scale hydrology
An overview of selected techniques for analysing surface-water data networks
Meteorological systems for hydrological purposes
Current operational applications of remote sensing in hydrology
Areal modelling in hydrology using remote sensing data and geographical information system
Contaminants in rivers and streams — Prediction of travel time and longitudinal dispersion
Precipitation estimation and forecasting