Handbook For Catchment Water Demand and

Handbook for the
Assessment of
Catchment Water
Demand and Use
in collaboration with the Department for Water Development

Ministry of Rural Resources and Water Development,
Zimbabwe, University of Zimbabwe, University of Zambia and
University of Zululand
Preface
The “Handbook for the Assessment of Catchment Water Demand and Use” has been
produced with funding from the Department for International Development (DFID) of the UK
Government under the Knowledge and Research (KAR) programme. The work has been
carried out as a collaborative venture between HR Wallingford (UK) and the Department for
Water Development, Ministry of Rural Resources and Water Development (Zimbabwe),
University of Zimbabwe, University of Zambia and the University of Zululand, South Africa.
The Handbook includes a number of case studies that have been undertaken by the various
members of the team. Pieter van der Zaag of the University of Zimbabwe has written chapter
2. Chapter 9 is based on the work of D. Kammer of the Department for Water Development
in Zimbabwe. This document was produced in May 2003.
The Handbook responds to the growing need to balance supply-side and demand-side
approaches to managing scarce water resources in catchments and river basins. It recognises
that a plethora of research and methodologies are readily available to assist planners and
managers to assess water resource availability in a catchment yet little is available to assist in
assessing water demand and use. The Handbook therefore aims to fill this gap by bringing
together a range of methodologies, examples of their application, supporting information and
key references.
The Handbook is aimed at professionals and practitioners in the southern African region. It
provides the user with a range of appropriate methods for estimating water demand and use
across a range of water uses including environmental, urban, industrial, rural domestic and
agricultural sectors. Guidance on the advantages and disadvantages of different assessment
techniques are provided and the texts supplemented by worked examples. Methods suitable
for forecasting long-term water demand and use are also included.
Acknowledgements
The following people and institutions have contributed to the preparation of this Handbook:
Peter Ashton (Council for Scientific and Industrial Research, Pretoria); Ruth Beukmann
(IUCN, Pretoria); Hannes Buckle (Rand Water); Emmanuel Dube (University of Zimbabwe);
Rumiana Hranova (University of Zimbabwe); Steve Gillham (Umgeni Water); Jeff Broome
(Ncube Burrow, Bulawayo); Bekithemba Gumbo (University of Zimbabwe); Hugo Maaren
(Water Research Commission, Pretoria); Bob Merry (Booker Tate UK); Sipho Mlilo
(University of Zimbabwe); Limpho Mutanya (University of Zimbabwe); Schalla Mulenga
(University of Zambia); Dr Muya (School of Civil Engineering, University of Zambia); Mr
Mutede (Department for Water Development Zimbabwe); Jerry Ndamba (Institute of Water
and Sanitation, Harare); Edwin Nyirenda (School of Civil Engineering, University of
Zambia); Dr Zebediah Phiri (Water Resources Action Programme, Zambia); Brian Rawlins
(Department of Hydrology, University of Zululand); Griphin Symphorian (University of
Zimbabwe); Tertia Uitenweerde (IUCN, Pretoria); Pieter van der Zaag (University of
Zimbabwe); Innocent Ziyambi (University of Zimbabwe).
This document is an output from a project funded by the UK Department for International Development (DFID)
for the benefit of developing countries. The views expressed are not necessarily those of DFID. The work is
being carried out under the Knowledge and Research (KAR) programme and the project details are:
Theme number W1 Water resource management – to improve the assessment, development and management
of water resources
Project title Integrated water information management (IWIM) system
Project No. R7135
HR Wallingford accepts no liability for the use by third parties of results or methods presented in this report.
The Company also stresses that various sections of this report rely on data supplied by or drawn from third party sources. HR
Wallingford accepts no liability for loss or damage suffered by the client or third parties as a result of errors or inaccuracies in
such third party data.
Handbook for the assessment of

catchment water demand and use
CONTENTS
Section Page no.
1. INTRODUCTION.......................................................................... 1
1.1 Structure of the Handbook 1
1.2 Scope of the Handbook 1
1.3 Definitions of water demand and use 2

1.3.1 Water use 2
1.3.2 Water demand 3
1.4 Overview of water use in southern Africa 3
1.4.1 Background 3
1.4.2 Shared river basins within the SADC region 4
1.4.3 Water use by sector within the SADC region 4
1.4.4 Water scarcity in the SADC region 5
1.5 References 6
2. AN INTRODUCTION TO THE PRINCIPLES OF MANAGING

WATER AT A CATCHMENT LEVEL ........................................... 7
2.1 Background 7
2.2 Roles and responsibilities of catchment managers 7
2.3 Balancing supply and demand 9
2.3.1 The supply side 10
2.3.2 The demand side 10
2.4 The legal framework 11

2.4.1 Regulation of water for particular uses 11
2.4.2 The use of water permits or water right 11
2.4.3 Hierarchy of water use 11
2.5 The value of water 12
2.6 Scales and boundary conditions 13
2.7 Issues in water allocation 14

2.7.1 Defining key concepts 14
2.7.2 Uncertainty 14
2.8 Efficiency and equity 15

2.8.1 Equity 15
2.8.2 Efficiency 15
2.8.3 Trade-offs 16
Handbook for the assessment of i
2.9 Water losses 16
2.10 Water allocation between sectors 17
2.11 Recommendations on water allocation 18
2.12 References 18
3. ENVIRONMENTAL WATER DEMAND AND USE..................... 20

3.1 Background 20
3.2 Environmental flow assessment methods 20
3.3 Hydrological index methods 22

3.3.1 The Tennant method 22
3.3.2 The Texas method 24
3.3.3 Flow duration curve analysis 24
3.3.4 Aquatic base flow method 25
3.3.5 Range of variability approach (RVA) 25
3.3.6 Unsuitable historical discharge techniques. 26
3.3.7 Advantages of hydrological index methodologies 27
3.3.8 Disadvantages of hydrological index methodologies 27
3.4 Hydraulic methods 27
3.4.1 The wetted perimeter technique 27
3.5 Holistic methods 32
3.5.1 The building block methodology 32
3.6 Habitat methods 34
3.6.1 Instream Flow Incremental Methodology 35
3.7 Comparison of the instream flow methods 39
3.8 Lake level requirements 40
3.9 Methodologies for assessing ecological flow requirements of

wildlife 41
3.10 References 41
4. AGRICULTURAL WATER DEMAND AND USE ....................... 44

4.1 Introduction 44
4.2 Overview of irrigated agriculture in the SADC region 44

4.2.1 Angola 45
4.2.2 Botswana 45
4.2.3 Democratic Republic of the Congo 45
4.2.4 Lesotho 45
4.2.5 Malawi 46
4.2.6 Mauritius 46
4.2.7 Mozambique 47
Handbook for the assessment of ii

4.2.8 Namibia 47
4.2.9 Seychelles 47
4.2.10 South Africa 47
4.2.11 Swaziland 48
4.2.12 Tanzania 48
4.2.13 Zambia 49
4.2.14 Zimbabwe 49
4.3 Use of empirical equations to calculate irrigation water
demand and use 49
4.4 Approximate method for calculating irrigation water demand

and use 52
4.5 Detailed method of estimating irrigation water demand and

use 53
4.5.1 Methods of estimating reference crop evapotranspiration 53
4.5.2 Estimation of the crop coefficient 58
4.5.3 Estimation of effective rainfall 61
4.5.4 Groundwater contribution 62
4.5.5 Stored soil water 63
4.5.6 Estimation of leaching requirements 63
4.5.7 Irrigation efficiencies 64
4.5.8 Detailed estimation of irrigation water demand and use 68
4.6 Measurement of irrigation water use 69
4.6.1 Measurement of conveyance losses 69
4.6.2 Ponding tests 69
4.6.3 Inflow-outflow studies 70
4.7 Irrigation water use productivity 70
4.8 Improving rain fed agricultural production 75

4.8.1 In-situ water conservation 76
4.8.2 Supplementary irrigation 77
4.9 Deficit irrigation 77
4.10 Case study of improved water use efficiency for an irrigation

scheme in Swaziland 78
4.11 Livestock water use 81
4.12 References 82
5. RURAL DOMESTIC WATER DEMAND AND USE ................... 85

5.1 Background 85
5.1.1 Definition of water demand, consumption and use 85
5.1.2 Typical rural domestic water sources 85
5.1.3 Background to rural domestic water supply and sanitation schemes 86
5.2 Typical rural domestic water use figures for southern Africa
87
5.2.1 Angola 88
5.2.2 Botswana 88
Handbook for the assessment of iii
5.2.4 Swaziland 89
5.2.5 Tanzania, Kenya and Uganda 89
5.2.6 Zimbabwe 91
5.3 Indirect and direct methods of estimating rural water demand
and use 92
5.4 Factors affecting rural domestic water demand and use 93

5.4.1 Population 93
5.4.2 Household occupancy rates 93
5.4.3 Level of service 94
5.4.4 Tariff levels 97
5.5 Indirect methods of estimating rural domestic water demand
and use 99
5.6 Direct methods of estimating rural domestic water demand

and use 101
5.6.1 Direct interviews with individuals 102
5.6.2 Community discussions and focus groups 102
5.6.3 Seasonal calendars and diaries 103
5.6.4 Transect walks and direct observation 103
5.7 References 105
6. INDUSTRIAL WATER DEMAND AND USE ............................ 107

6.1 Background 107
6.2 Industrial processes water use 107
6.3 Industrial water consumption statistics 109
6.4 Participatory methods for obtaining industrial water use

consumption data 109
6.5 Water audits for industry 110

6.5.2 Measuring in-plant industrial water use 112
6.5.3 Conducting a water audit 112
6.5.4 Measuring leakage 113
6.6 The effect of improving water use efficiency on effluent
water quality 114
6.7 Case studies of industrial water demand management for

three industries in Zimbabwe 114
6.7.1 Background to case studies 114
6.7.2 Results from the galvanised wire manufacture 114
6.7.3 Results from the soft drink manufacturer 116
6.7.4 Results from the sugar refinery 117
6.7.5 Conclusions 119
Handbook for the assessment of iv

6.8 References 120
7. URBAN WATER DEMAND AND USE..................................... 121

7.1 Background 121
7.2 Typical urban water demand figures for southern Africa 122
7.2.1 Angola 124
7.2.2 Botswana 124
7.2.3 Democratic Republic of the Congo 125
7.2.4 Lesotho 125
7.2.5 Malawi 125
7.2.6 Mauritius 125
7.2.7 Mozambique 126
7.2.8 Namibia 126
7.2.10 Swaziland 127
7.2.11 Tanzania 127
7.2.12 Zambia 128
7.2.13 Zimbabwe 128
7.3 Estimation of water use where records are available 128
7.3.1 Estimation of water use using meters 128
7.3.2 Estimation of water use from pumping records 129
7.3.3 Estimation of water use for unmetered conumers using test metering 131
7.4 Estimates of unmetered urban water demand 132
7.4.1 Estimates of unmetered urban domestic demand 132
7.4.2 Estimate of the urban domestic demand of the population without access to
piped water supply 134
7.4.3 Estimating unmetered urban institutional and commercial water use and
demand 134
7.4.4 Estimating unmetered public water demand and use 135
7.4.5 The effect of the number of occupants in a household on water use 135
7.5 Unaccounted for water 137
7.5.1 Background 137

7.5.2 Methods for reducing unaccounted for water 139
7.5.3 Estimation of water losses in an urban water supply system 140
7.5.4 Location of water losses in an urban water supply scheme 141
7.5.5 The level of metering required within an urban water supply reticulation
scheme 142
7.5.6 Methods for locating leaks 144
7.6 Water demand management measures 145
7.6.1 Pressure reduction as a demand management measure 146
7.6.2 Physical measures to reduce water demand 147
7.6.3 The effect of tariff levels on water demand 148
7.7 Urban water demand case study for the city of Windhoek in
Namibia 154
7.8 References 156
8. FORECASTING WATER DEMAND AND USE........................ 158

Handbook for the assessment of v
8.1 Background 158
8.2 Influences on water demand and use 158
8.3 Criteria for assessing forecasting methods 158

8.3.1 Consistency and transparency of method 159
8.3.2 Logical/theoretical appeal 159
8.3.3 Incorporates and explains historical trends 159
8.3.4 The treatment of factors not taken into account in the past 159
8.3.5 Empirical validation 159
8.3.6 Acceptance by the regulatory body 160
8.3.7 Cost and feasibility 160
8.4 Choice of forecasting method 160
8.4.1 Checklist for determining the forecasting method 161
8.4.2 Checklist for applying the chosen forecasting method 161
8.6 Details of forecasting methods 162
8.6.1 Judgmental forecasts 163
8.6.2 Extrapolation of historical data and trend analysis 163
8.6.3 Forecasts based on population growth rate 165
8.6.3 Forecasts based on component analysis 167
8.6.4 Multiple linear and non-linear regression analysis 169
8.7 Forecasting agricultural water demand and use 170
8.9 Forecasting industrial water demand and use 173

8.9.1 Checklist of data for forecasting industrial water use 174
8.10 Forecasting urban water demand 175
8.11 The effect of the HIV/AIDS virus on future water demand in

southern Africa 176
8.12 Case study for demand forecasting for city of Masvingo in

Zimbabwe 177
8.12.1 Background 177
8.12.2 Use of a multiple linear regression equation to forecast demand 178
8.12.3 The effect of water demand management measures on future water demand
181
8.12.4 Potential and short-term reduction of water demand for Masvingo 182
8.13 References 184
9. RIVER TRANSMISSION LOSSES........................................... 185

9.1 Background 185
9.2 Nature of river transmission losses 185
9.3 Example of transmission losses from southern Africa 186

9.3.1 Zimbabwe 186
9.3.2 Kenya 190
Handbook for the assessment of vi

9.4 References 190
APPENDICES
A1 INDUSTRIAL WATER CONSUMPTION LOOK UPTABLES...A-1
A1.1 Background to look up tables A-1
A1.2 Examples of how to use the industrial water consumption

look up tables A-1
A1.3 Industrial water consumption look up tables A-2

A1.3.1 Beverages (soft drinks) A-2
A1.3.2 Breweries A-3
A1.3.3 Sorghum beer and sorghum malting A-4
A1.3.4 Brick production A-5
A1.3.5 Cement and concrete products A-6
A1.3.6 Ceramics manufacture A-6
A1.3.7 Manufacture of various specialist chemical products A-6
A1.3.8 Chipboard and medium density fibreboard (MDF) production A-8
A1.3.9 Cosmetics manufacture A-9
A1.3.10 Electronic goods assembly A-9
A1.3.11 Fibreglass production A-9
A1.4.12 Fish processing A-10
A1.4.13 Food processing – dairy produce A-10
A1.3.14 Food processing – flour products A-12
A1.3.14 Food processing – fruits A-13
A1.3.15 Food processing – miscellaneous A-14
A1.3.16 Food processing – multi-product confectionary plant A-14
A1.3.17 Food processing – vegetable products A-15
A1.3.18 Fresh meat production A-17
A1.3.19 Commercial laundries A-19
A1.3.19 Lead acid battery production A-19
A1.3.20 Leather tanning A-20
A1.3.21 Light industrial estate water consumption A-21
A1.3.22 Metal finishing A-21
A1.3.22 Metal processing A-22
A1.3.23 Mining A-23
A1.3.24 Plastic manufacturing A-24
A1.3.25 Power generation A-24
A1.3.26 Poultry processing A-28
A1.3.27 Pulp and papermaking A-28
A1.3.28 Quarries A-30
A1.3.29 Semiconductor wafer fabrication A-30
A1.3.30 Steel manufacturing A-31
A1.3.31 Sugar cane refining A-31
A1.3.32 Textile manufacturing A-31
A1.3.33 Vehicle manufacturing A-33
A1.3.34 Wallpaper manufacturing A-34
A1.3.35 Wine production A-34
A1.4 References A-35
FIGURES
FIGURE 1.1 MAP OF THE SOUTHERN AFRICAN DEVELOPMENT COMMUNITY (SADC) REGION ..... 2
FIGURE 2.1 VARIATION OF WATER AVAILABILITY AND DEMAND, AND RELIABILITY OF SUPPLY ...... 17
FIGURE 3.1 FLOW DURATION CURVE METHOD ........................................................................ 24
Handbook for the assessment of vii
FIGURE 3.2 PHOTOGRAPH OF A RIFFLE ON A SOUTH AFRICAN WATERCOURSE ......................... 28
FIGURE 3.3 USE OF THE WETTED PERIMETER METHOD TO ESTIMATE INSTREAM FLOWS ............ 29
FIGURE 3.4 SURVEYED CROSS-SECTION OF THE HWADZI RIVER ............................................. 30
FIGURE 3.5 NON-DIMENSIONAL WETTED PERIMETER VERSUS DISCHARGE RELATIONSHIP .......... 31
FIGURE 3.6 TYPICAL CURVES TO ESTABLISH THE MAINTENANCE OF LOW FLOWS ....................... 33
FIGURE 3.7 EXAMPLES OF THE FLOW BUILDING BLOCKS USED IN THE BUILDING BLOCK
METHODOLOGY ................................................................................................... 34
FIGURE 3.8 CONCEPTUALIZATION OF HOW PHABSIM CALCULATES HABITAT VALUES AS A
FUNCTION OF THE DISCHARGE ............................................................................. 36
FIGURE 3.9 EXAMPLE OF WEIGHTED USABLE AREA VERSUS DISCHARGE CURVES FOR VARIOUS
LIFE STAGES OF A SALMON .................................................................................. 37
FIGURE 3.10 SPECTRUM OF INSTREAM FLOW METHODOLOGIES................................................. 39
FIGURE 4.1 APPROXIMATE METHOD FOR ESTIMATING IRRIGATION DEMAND AND USE ................ 50
FIGURE 4.2 DETAILED METHOD FOR ESTIMATING IRRIGATION WATER DEMAND AND USE ............ 51
FIGURE 4.3 DIAGRAM OF A CLASS A EVAPORATION PAN ......................................................... 55
FIGURE 4.4 DIAGRAM OF A COLARADO EVAPORATION PAN ...................................................... 56
FIGURE 4.5 REFERENCE CROP EVAPOTRANSPIRATION CALCULATED USING THE PENMAN-
MONTEITH METHOD ............................................................................................ 57
FIGURE 4.6 TYPICAL CROP COEFFICIENT CURVE ..................................................................... 59
FIGURE 4.7 TYPICAL RANGES EXPECTED IN KC FOR FOUR GROWTH STAGES ............................. 60
FIGURE 4.8 CROP COEFFICIENT CURVE TOGETHER WITH CROP WATER REQUIREMENTS ............ 61
FIGURE 4.9 THE RELATIVE MAGNITUDE OF QUANTITIES OF WATER FLOWING THROUGH AN
“AVERAGE” IRRIGATION SCHEME .......................................................................... 67
FIGURE 4.10 YIELDS AND WATER REQUIREMENTS OF IRRIGATED AND RAIN FED AGRICULTURE .... 70
FIGURE 4.11 W ATER USE VERSUS YIELD RELATIONSHIP FOR IRRIGATED WHEAT IN SOUTHERN
ZIMBABWE FOR 1995 TO 1999 ............................................................................ 72
FIGURE 4.12 RELATIONSHIP BETWEEN NET WATER USE AND YIELD FOR MAIZE FOR NYANYADZI IN
ZIMBABWE.......................................................................................................... 74
FIGURE 4.13 RELATIONSHIP BETWEEN NET IRRIGATION WATER AND YIELD FOR MAIZE FOR
NYANYADZI IN ZIMBABWE .................................................................................... 74
FIGURE 4.14 METHODS OF IMPROVING RAINFED AGRICULTURAL PRODUCTION ........................... 75
FIGURE 4.15 EFFECTS OF SUB-SOILING IN TANZANIA ................................................................ 76
FIGURE 4.16 GROWTH IN SUGARCANE AREA AND THE USE OF DRIP IRRIGATION FOR THE SIMUNYE
SUGAR ESTATE IN SWAZILAND ............................................................................. 79
FIGURE 4.17 SUCROSE PRODUCTIVITY BY IRRIGATION TYPE FOR THE SIMUNYE SUGAR ESTATE IN
SWAZILAND ........................................................................................................ 80
FIGURE 4.18 RELATIVE VALUE OF THE PROJECT BENEFITS FOR THE SIMUNYE SUGAR ESTATE IN
SWAZILAND ........................................................................................................ 80
FIGURE 5.1 SETTLEMENTS SURVEYED FOR THE DRAWERS OF W ATER II STUDY ....................... 90
FIGURE 5.2 W ATER USE FIGURES FOR PIPED AND UNPIPED SOURCES FROM THE DRAWERS OF
W ATER II STUDY ................................................................................................. 91
FIGURE 5.3 W ATER USE PER HOUSEHOLD VERSUS HOUSEHOLD OCCUPANCY .......................... 94
FIGURE 5.4 W ATER USE VERSUS DISTANCE FROM SOURCE ..................................................... 97
FIGURE 5.5 EFFECT OF TARIFF LEVELS AND LEVELS OF SERVICE ON WATER USE ...................... 98
FIGURE 5.6 ELASTICITY OF WATER DEMAND FOR DIFFERENT USES .......................................... 99
FIGURE 5.7 INDIRECT METHOD OF CALCULATING TOTAL WATER DEMAND AND USE .................. 100
FIGURE 6.1 W ATER SUPPLY, USE AND TREATMENT FOR A TYPICAL INDUSTRIAL PLANT ............ 108
FIGURE 6.2 SIMPLIFIED WATER BALANCE FOR A MANUFACTURING FACILITY ............................ 111
FIGURE 6.3 SIMPLIFIED FLOW DIAGRAM FOR THE WIRE GALVANISING PROCESS PLANT ............ 115
FIGURE 6.4 SIMPLIFIED FLOW DIAGRAM FOR THE SOFT DRINK MANUFACTURING PLANT ........... 117
FIGURE 6.5 EXISTING COOLING WATER AND CONDENSATE SYSTEM FOR THE SUGAR REFINERY 118
FIGURE 6.6 PROPOSED COOLING WATER AND CONDENSATE SYSTEM FOR THE SUGAR REFINERY
........................................................................................................................ 119
FIGURE 7.1 W ATER SUPPLY IN LARGE AFRICAN CITIES: SOURCE OF WATER........................... 122
FIGURE 7.2 DUTY POINT OF A PUMP ..................................................................................... 130
FIGURE 7.3 EFFECT IN THE CHANGE IN STATIC HEAD ON DISCHARGE ..................................... 130
FIGURE 7.4 AVERAGE INTERNAL WATER CONSUMPTION FOR MIDDLE INCOME GROUP DWELLINGS
IN SOUTH AFRICA IN 1987................................................................................. 136
FIGURE 7.5 GENERAL SCHEME OF A WATER SUPPLY SYSTEM ................................................ 141
Handbook for the assessment of viii

FIGURE 7.6 DIAGRAM OF DIFFERENT USES OF DOMESTIC WATER AND THEIR ELASTICITIES OF
DEMAND ........................................................................................................... 148
FIGURE 7.7 W ATER CONSUMPTION FOR HIGH AND LOW DENSITY SUBURBS FOR THE TOWN OF
RUWA IN ZIMBABWE .......................................................................................... 149
FIGURE 7.8 MONTHLY BILLED WATER CONSUMPTION FOR AFFLUENT AND LESS AFFLUENT
HOUSEHOLDS IN MASVINGO .............................................................................. 150
FIGURE 7.9 BLOCK TARIFF STRUCTURES FOR W INDHOEK, GABORONE AND HERMANUS ......... 152
FIGURE 7.10 THE EFFECT OF WATER CONSERVATION MEASURES ON THE CONSTRUCTION OF
INFRASTRUCTURE ............................................................................................. 155
FIGURE 8.1 FORECASTING WATER USE FOR THREE POPULATION GROWTH SCENARIOS ........... 162
FIGURE 8.2 RESULTS OF TREND ANALYSES USING DIFFERENT FITTING TECHNIQUES ............... 164
FIGURE 8.3 THE DANGERS OF USING EXTRAPOLATION TECHNIQUES FOR FORECASTING WATER
DEMAND FOR MASVINGO IN ZIMBABWE............................................................... 165
FIGURE 8.4 W ATER USE FORECAST BASED ON POPULATION GROWTH .................................... 167
FIGURE 8.5 COMPONENTS OF AN URBAN WATER SUPPLY SYSTEM ......................................... 168
FIGURE 8.6 HOUSEHOLD LEVEL COMPONENTS ..................................................................... 169
FIGURE 8.7 THE EFFECT OF A CHANGE IN IRRIGATION TECHNIQUE ON WATER DEMAND AND USE
........................................................................................................................ 171
FIGURE 8.8 EXAMPLE OF A RURAL DOMESTIC DEMAND FORECAST ......................................... 173
FIGURE 8.9 EXAMPLE OF A RELATIONSHIP BETWEEN GDP AND INDUSTRIAL WATER DEMAND .. 174
FIGURE 8.10 FORECASTING URBAN WATER DEMAND............................................................... 176
FIGURE 8.11 TREATED WATER PRODUCTION FOR THE CITY OF MASVINGO 1977 TO 2001 ........ 178
FIGURE 8.12 ACTUAL AND MODELLED WATER USE FOR MASVINGO 1977 TO 2001 ................... 180
FIGURE 8.13 PREDICTED WATER USE FOR MASVINGO UP TO 2021.......................................... 181
FIGURE 9.1 ILLUSTRATION OF TRANSMISSION LOSSES .......................................................... 186
Handbook for the assessment of ix

Handbook for the assessment of x
1. INTRODUCTION
The aim of the Handbook is to support professionals and practitioners in sub-Saharan Africa
responsible for the management of water resources at a catchment and sub-catchment level.
The Handbook provides practical guidance for assessing and forecasting water demands and
use for the following sectors:
• Environment;
• Agriculture;
• Rural domestic;
• Urban;
• Industry.
Each of these sectors is covered in a separate chapter as detailed below.
1.1 Structure of the Handbook

The Handbook is structured as follows:
• Chapter 1 provides background on the scope of the Handbook together with important
definitions and overview of water resources and water demand in southern Africa;
• Chapter 2 provides an introduction to the principles of managing water at a catchment
level;
• Chapter 3 details methods to estimate instream environmental water demands;
• Chapter 4 provides methods of estimating agricultural water demand and use based
primarily on techniques recommended by the United Nations (UN) Food and Agriculture
Organization (FAO);
• Chapter 5 provides an outline of methods of assessing rural domestic water demand and
use;
• Chapter 6 outlines methods to assess industrial water demands and use. Appendix A
contains look up tables that provide typical specific water consumption figures for a
variety of industries;
• Chapter 7 details methods to assess urban water demand and use including commercial
and institutional demands but excluding industrial water demands. Typical per capita
water consumption figures for urban areas throughout southern Africa are provided;
• Chapter 8 provides brief details on demand forecasting methods;
• Chapter 9 outlines methods for assessing river transmission losses. Although river
transmission losses are not strictly a demand it is important that they are taken into
account when allocating water at a catchment level.
1.2 Scope of the Handbook

The driving principle behind the Handbook is the requirement that catchment and water
resources managers need to have access to simple and effective tools to estimate water
demand and use on a catchment and sub-catchment basis. In the past water resources
planning has been supply driven. However, with water resources in arid and semi-arid areas
becoming increasing scarce it is important that water demand and use is managed efficiently
before new sources of water are developed. To manage water resources effectively, current
and future water demand and use for all sectors should be estimated as accurately as possible.
It should be noted that the Handbook is limited to assisting catchment and water resources
managers estimate water demand and use at a sub-catchment and catchment level. The
Handbook for the assessment of 1
Handbook outlines techniques that are primarily aimed for use in the countries that form the
Southern African Development Community (SADC). However, many of the techniques
described can be applied in other parts of Africa and the world. A map of the 14 states from
which SADC is comprised is shown in Figure 1.1.
Kinshasa
Democratic Republic
of the Congo Dodoma
Seychelles
Dar Es Salaam
Tanzania
Luanda
Angola
Malawi
Zambia Lilongwe
Lusaka Blantyre
Harare
Mutare
Zimbabwe
Namibia Mauritius
Beira
Windhoek Botswana Bulawayo
Mozambique
Walvis Bay Gaborone
Pretoria
Mbabane
Maputo
Johannesburg
Swaziland
Maseru Capital cities
South Durban Major cities
Lesotho
Africa
0 500 km
Cape Town Port Elizabeth
Figure 1.1 Map of the Southern African Development Community (SADC)

region
1.3 Definitions of water demand and use

The terms water use and water demand are often used interchangeably. However, these terms
have different meanings. In the context of the Handbook the definitions of these terms are
given below.
1.3.1 Water use
Water use can be distinguished into three different types. These are:
• Withdrawals or abstractions where water is taken from a surface or groundwater

source, and after use returned to a natural water body, e.g. water used for cooling in

industrial processes that is returned to a river. Such return flows are particularly
important for downstream users in the case of water taken from rivers;
• Consumptive water use or water consumption that starts with a withdrawal or an
abstraction but in this case without any return flow. Water consumption is the water
abstracted that is no longer available for use because it has evaporated, transpired, been
incorporated into products and crops, consumed by man or livestock or otherwise
removed from freshwater resources. Water losses during the transport of water between
the points of abstractions and the point of use, (e.g. resulting from leakage from
distribution pipes), are excluded from the consumptive water use figure. Examples of
consumptive water use include steam escaping into the atmosphere and water contained
in final products i.e. it is water that is no longer available directly for subsequent uses;
• Non-consumptive water use i.e. the in situ use of a water body for navigation, instream
flow requirements for fish, recreation, effluent disposal and hydroelectric power
generation.
1.3.2 Water demand
Water demand is defined as the volume of water requested by users to satisfy their needs. In
a simplified way it is often considered equal to water consumption, although conceptually the
two terms do not have the same meaning. This is because in some cases, especially in rural
parts of southern Africa, the theoretical water demand considerably exceeds the actual
consumptive water use.
1.4 Overview of water use in southern Africa

There are 14 members of SADC: Angola, Botswana, Democratic Republic of the Congo,
Lesotho, Malawi, Mauritius, Mozambique, Namibia, Seychelles, South Africa, Swaziland,
Tanzania, Zambia and Zimbabwe. The SADC region covers an area of almost 6.8 million
km2. Regional estimates put the renewable freshwater resources at an annual average of
650,000 million m3 distributed in the region’s rivers, lakes and aquifers. The sections below
give an overview of water use in the SADC region.
1.4.1 Background
The southern African climate varies from tropical rain forests in the north of the region to
desert conditions in the south-west. These climatic conditions make rainfall one of the most
important climatological elements in the SADC region. Large areas within the SADC region
are very dry and cannot support sustainable human existence and agriculture, especially in the
south-western parts of the region. A very dry period in the region can result in crop failures,
food deficits and, in the extreme, starvation. Rainfall is nearly non-existent in some parts of
Namibia. In 1999, for instance, 85% of the country received below average rainfall. This
meant that the country could not produce the food it needed. The eastern coastal zones and
northern sub-areas (Angola, Malawi, Zambia, Mozambique, Tanzania and parts of
Zimbabwe) are relatively wet. In these regions the mean annual precipitation varies from
1,000 mm to 1,600 mm per year, with isolated areas receiving more than 2,000 mm. The high
annual evaporation rates, (in some areas as high as 4,000 mm), means that only some 3% to
15% of all rainfall flows to rivers and lakes. Only 360 km3, out of 600 km3, are suitable for
use in the agricultural, industrial and domestic sectors. Water, therefore, plays an important
role in the socio-economic development of the SADC, especially in terms of food production,
health, energy and the environment.

1.4.2 Shared river basins within the SADC region
There are fifteen major rivers shared among the continental SADC member-states. All the
continental SADC countries share one or more river basins, as noted in Table 1.1. Namibia,
for instance, one of the most arid countries in the region, has access to five international river
basins. Mozambique shares nine of its rivers with other countries. The Zambezi basin, the
largest and probably most crucial in southern Africa, crosses eight countries.
Table 1.1 Shared river basins within the continental SADC states
Country Number of basins Name of river basins
Angola 5 Cunene, Cuvelai, Okavango, Congo, Zambezi

Botswana 4 Limpopo, Okavango, Orange, Zambezi
Democratic Republic 2 Congo, Nile
of the Congo
Lesotho 1 Orange
Malawi 2 Ruvuma, Zambezi
Mozambique 9 Buzi, Incomati, Limpopo, Ruvuma, Save,
Maputo, Pungue, Umbeluzi, Zambezi
Namibia 5 Cunene, Cuvelai, Okavango, Orange, Zambezi
South Africa 4 Incomati, Limpopo, Maputo, Orange
Swaziland 3 Incomati, Maputo, Umbeluzi
Tanzania 3 Nile, Ruvuma, Zambezi, Congo
Zambia 2 Zambezi, Congo
Zimbabwe 6 Buzi, Limpopo, Okavango, Pungwe, Save,
Zambezi
Source: Reference 1.1
1.4.3 Water use by sector within the SADC region
Table 1.2 gives broad patterns of water use in southern Africa. Whilst the absence of data on
the total volumes of water used in each country prevents detailed comparisons from being
made, agricultural water use in each country clearly dominates when compared to the domestic
and industrial water use sectors. The high proportion of water used for agriculture suggests that
each SADC country rely heavily on food grown within its borders to meet national goals of
food security.

Table 1.2 Water use by sector for continental SADC states
Country Agriculture (%) Industry (%) Domestic (%)
Angola 76 10 14
Botswana 48 20 32
Democratic Republic 23 16 61
of the Congo
Lesotho 56 22 22
Malawi 86 3 10
Mozambique 89 2 9
Namibia 68 3 29
South Africa 62 21 17
Swaziland 71 8 21
Tanzania 89 2 9
Zambia 77 7 16
Zimbabwe 79 7 14
1.4.4 Water scarcity in the SADC region
Table 1.3 presents categories of water scarcity associated with varying levels of water supply
per person per year, the typical scales of problems encountered in each category in southern
Africa.
Table 1.3 Water scarcity
Water scarcity category and associated problems Volume of water available

3
(m /person/year)
Beyond the “water barrier”: continual, wide-scale
< 500
water supply problems, becoming catastrophic during
droughts.
Chronic water scarcity: continual water supply
problems, worse during annual dry seasons; frequent 500 to 1,000
severe droughts.
Water stressed: frequent seasonal water supply and
quality problems, accentuated by occasional 1,000 to 1,666
droughts.
Moderate problems: occasional water supply and
quality problems, with some adverse effects during 1666 to 10,000
severe droughts.
Well-watered: very infrequent water supply and
quality problems, except during extreme drought
> 10,000
conditions.
Table 1.4 provides estimates of the total water available for each of the continental SADC
states on a per capita basis for the year 2000 and 2025. The estimated populations for the
year 2025 take into account the likely effects of HIV/AIDS on population growth.

Table 1.4 Water availability for continental SADC states
Water per Water per

Total Population Estimated
person in person in
water in 2000 population
Country 2000 2025
available 3 in 2025 3
3 (m /person/ (m /person/
(km ) * (millions) (millions)**
year) year)
Angola 205.0 12.9 15,888 22.0 9,335

Botswana 1.6 1.6 976 2.0 808
Democratic Republic
1,019.0 52.0 19,579 114.1 8,930
of the Congo
Lesotho 5.2 2.2 2,412 3.2 1,602
Malawi 17.5 10.8 1,624 16.1 1,089
Mozambique 17.0 20. 0 5,856 28.8 4,066
Namibia 2.7 1.7 1,553 2.6 1,052
South Africa 52.8 43.3 1,220 49.0 1,077
Swaziland 2.8 0.9 3,017 1.3 2,228
Tanzania 80.0 33.7 2,371 63.6 1,257
Zambia 127.0 9.2 13,818 14.9 8,526
Zimbabwe 15.5 13.1 1,182 14.0 1,108
Notes: *This is the surface plus ground water that is generated within the geo-political boundaries of
the country each year and excludes water that flows in from neighbouring states. Minor
volumes of recycled water are included in the values for water available in South Africa.
**Population growth rates in each country used to estimate the population in 2025 have been
adjusted to account for the current prevalence of HIV/AIDS in that country.
Table 1.4 indicates that five of the 12 continental SADC countries are water stressed and a
further one country (Botswana) is facing a chronic scarcity of water. Table 1.4 also indicates
that the water availability per person will significantly decrease by 2025.
1.5 References
1.1 Southern African Development Community (SADC) (1999) Round Table Conference
on the theme "Integrated Water Resources Development and Management in the
Southern African Development Community."
http://www.sadcwscu.org.ls/programme/rtc/rtc-II.htm
1.2 IUCN The World Conservation Union (1996) Water in southern Africa IUCN – The
World Conservation Union (Regional office for southern Africa).

2. AN INTRODUCTION TO THE PRINCIPLES OF
MANAGING WATER AT A CATCHMENT LEVEL
2.1 Background
This chapter provides an introduction to some of the issues and principles that need to be
addressed when allocating water at a macro-level. The chapter has been written by Pieter van
der Zaag of the University of Zimbabwe.
An important purpose of water management is to match or balance the demand for water with
its availability, through suitable water allocation arrangements. As detailed in the previous
chapters at a catchment and sub-catchment level there is a large number of often conflicting
of water uses including:
• Irrigation;
• Domestic use in urban centres;
• Domestic use in rural areas;
• Livestock;
• Industrial use;
• Commercial use;
• The environment (e.g. instream flow requirements for aquatic life and wildlife);
• Institutions (e.g. schools, hospitals);
• Hydropower;
• Cooling (e.g. for thermal power generation);
• Waste and wastewater disposal;
• Fisheries;
• Recreation;
• Navigation.
In many southern African countries there have been significant reforms in the way in which
water is managed. One aspect of these water reforms in southern Africa is increased
stakeholder participation in water management through catchment management organisations.
The roles and responsibilities of catchment managers are discussed below.
2.2 Roles and responsibilities of catchment managers

Catchment planners are mainly concerned with the economic and social development of the
catchment and are not normally directly involved with water management. However, in
planning the development of a catchment the potential changes in water use and demand, as
well as the changes in the water resource, must be considered. Any change in the land use or
activities in the catchment affects the water situation and requires decisions to be made about
whether and how those changes can be incorporated into the water management strategy for
the catchment.
Increasingly, water has become the limiting factor to development either on a catchment level
or a national level in many SADC countries. Future development and certainly further
growth will, in many cases, rely on the location of a new source (such as through inter-basin
transfers) or water saving either through increases in water use efficiency or a change in the
catchment development strategy toward less water-intensive economic activities.
Those responsible for catchment planning are usually not directly involved with, or expert in,
water management and require support information to make informed development decisions.

Catchment planning is also a dynamic process where development priorities are constantly
changing. Consequently, the impact of the development plans on water use are also
constantly changing and affect water demands.
The primary components of catchment planning where water management is concerned are:
• To ensure that planned developments do not have an adverse effect on either the
hydrology or water quality of the catchment
• To ensure that adequate water resources are available to all water users in all sectors over
the development horizon
To implement these primary components, the catchment planner requires knowledge in

several areas common to water managers. These include:
(i) Catchment water resources
Catchment development is only feasible if there is sufficient water to support it. Water
resources are often now the limiting factor to future development. The potential for new
sources or increased storage either within or outside the catchment may also be considered to
improve the resource potential if warranted. The schedule for the development of these and
other water resources infrastructure must also be considered to ensure that the infrastructure is
in place when needed. This should be part of the planning process.
(ii) Water demand forecasts
Water demand forecasts are tied directly with catchment planning and their relationship
should be cyclical, as the forecasts are based on expected development and growth in the
catchment. As part of demand management, the potential for improving efficiency in
distribution or otherwise reducing demands must be assessed as this can make significant
differences in the resource potential of the catchment.
(iii) Water allocation policies and strategies
Water allocation policies are directly linked to development planning and may also be
cyclical because water allocation policies may result from development policies or vice-versa.
Development plans may highlight the need to reconsider allocation policies. For example,
where catchment plans lean toward industrial development this may conflict with agricultural
water demands and may lead to a reassessment of water allocation to these sectors.
(iv) Reliability requirements
Different economic activities have varying reliability requirements and both the reliability
needs of the industries and the ability of the catchment and its water infrastructure to meet the
required levels of reliability. The impact of drought and drought alleviation strategies are an
aspect of this.
(v) Costs of supplying water and expected returns of the planned

development
Especially where water is becoming a limiting factor, the cost of supplying additional water
for a development may be prohibitive. These costs may be offset if the value of the industry
to the catchment is high. Such costs must be included in the overall catchment plan. Water
quality must also be considered within these costs both in terms of supplying water of
adequate quality and ensuring that the resulting effluent is acceptable. Water quality
standards must also be met.
(vi) Hydrological impacts of planned developments
The hydrological impacts of the development must be fully considered, not just looking at the
impacts of a single development, but for the combined effect of all planned development over
the entire planning horizon period and including the impact of drought.
(vii) Environmental considerations
The basic flow requirements for rural communities and other unregulated use must be
considered in terms of meeting these demands as well as ensuring that the hydrological
impacts on these are acceptable. Similarly, minimum flow requirements for environmental or
aesthetic needs are also important. Sustainability is the key.
(viii) Source and quality of the information used
The source of the information used and its quality is of great importance. Planning decisions
must be based on the best information available and the risks inherent in using that
information must be well understood and incorporated into the decision making process.
Future monitoring needs should also be considered at this time.
(ix) Water permits
Whatever the planned development for the catchment, water users will need permits for water
use and the plans must fit with current permitting policy (though permit policy can also be
flexible to accommodate changes in development plans).
The background to many of these components and the importance that estimating water
demands accurately plays in managing water at a catchment level are discussed below.
2.3 Balancing supply and demand

There are various ways in which water can be allocated. The challenge is to find an optimal
allocation that, firstly, adheres to laid down regulations, and secondly, satisfies the water
demand of all users as much as possible. The problem of water allocation can be said to be
"to balance properly between a whole set of obligations:
• To international conventions;
• To human basic rights for wellbeing of both upstream and downstream societies;
• For protection of land productivity;
• For delivery of ecological goods and services from both terrestrial and aquatic
ecosystems; and
• For resilience of ecosystems to both natural and man-made disturbances." (Reference
2.1).
Water allocation is not generally an issue when water availability far surpasses water demand.
In such situations all demands can be satisfied, and in there may be no need for a regulated
allocation of water. However, this is not the case for the majority of the catchments in
southern Africa. In most catchments in southern Africa water availability is frequently less
than the demand for it. It is then necessary to find a suitable allocation of the scarce water.
Water allocation is not only concerned with the physical allocation of water. More broadly it
is about satisfying conflicting interests depending on water. These may be functions derived
from issues such as the following:

• Water such as navigation (e.g. navigability is often reliant on minimum water levels);
• Hydropower (e.g. a minimum head difference may be required);
• Environment (e.g. many aquatic species will require a certain flow regime or seasonal
fluctuations in water level to live and breed);
• Cultural and recreation (availability of water is often necessary for cultural and recreation
needs).
Although many of these functions are only to a certain extent consumptive, they can conflict
in both their timing and spatial distribution. Flood protection is also a function of the water
resources system that relates to the water resources. Flood protection through the
construction of storage dams can have a positive impact on water availability for other
functions (e.g. hydropower), but can have negative impacts on others (e.g. on the
environment). Finding a suitable allocation key for water can be quite complex, since a large
number of parameters have to be considered both on the supply and the demand side.
2.3.1 The supply side
On the supply side the generation of water in a catchment area naturally fluctuates, both
within years and between years. Water also occurs in different forms that often have different
uses. Special reference is made to rainfall and its use as "green water" in agriculture. Green
water cannot be allocated in the same way as "blue" water occurring in rivers and aquifers.
However, dryland agriculture and other types of land use do influence the partitioning of
rainfall into groundwater recharge, surface runoff and soil moisture (i.e. evaporation and
transpiration), and hence their availability.
2.3.2 The demand side
There are various parameters that affect demand at a catchment level.
(i) The demand for water fluctuates. However, fluctuations in demand are normally
much less than those on the supply side. For many types of uses, water demand
increases when water availability decreases, such as during the dry season.
(ii) Many water uses are (partially) consumptive, meaning that the water abstracted
will not return to the water system in the form of "blue water"; consumptive water use
typically converts blue or green water into water vapour, which in this form cannot be
allocated to other users. Water uses that are non-consumptive allow others to use the
water afterwards. Recreational water uses are a typical example. However, some
non-consumptive uses alter the time when this water becomes available for other
users. A typical example is water used for the generation of hydropower: electricity
is needed also during the wet season, and thus water has to be released from dams for
this purpose, when demand for it from other sectors may be low. As a result, this
water used for electricity generation is unavailable to these potential uses when they
need it. The environment is another (partially) non-consumptive user of water; its
requirements are frequently out of sync with the needs of other users. (That is
precisely why these environmental water requirements are now increasingly being
recognised.).
(iii) Many uses of water generate return flows that in principle, are available for other
uses. However, return flows normally have a lower quality than the water originally
abstracted. This may severely limit their re-use. Sometimes the quality of return flows
is a hazard to public health and the environment.
(iv) Different types of water use require different levels of assurance. For arable
(non-perennial) irrigated crops, levels of assurance of 80% (i.e. failure to produce the
maximum yield owing to water shortages in one out of five years) may be acceptable.

For urban water supply assurance levels of 96% or higher are the norm (failing in one
out of 25 years).
2.4 The legal framework

In many countries water is considered a public good i.e. the water is owned by the citizens of
a country, and the government manages this public good on their behalf. Laws and
regulations will therefore provide the rules pertaining to the use of this public resource. In
countries where water is considered a public good, water allocation may be viewed as the
process of converting a public good into a private one. An irrigator, for instance, will apply
the water to his/her privately owned crop. The crop will consume a large part of the water,
converting it into water vapour and increasing its yield. The irrigator derives direct and
private benefit from using a public good, but in so doing denies another person the
opportunity to use that water and deriving similar private benefits.
Balancing supply and demand must be done within the established legal framework. A
country's water law and subsidiary government regulations will prescribe many aspects of
water allocation. Amongst these are:
• Regulation of water for particular uses;

• The use of permits or water rights;
• Prioritisation of hierarchy of different water uses.
2.4.1 Regulation of water for particular uses
Water laws usually prescribe the types of water use that are regulated and those that are not
regulated. Regulated water uses require some kind of permit, concession or right.
Unregulated water use does not require licences to abstract from or discharge to body of
water. Water used for “primary” uses such as the environment or human beings often does not
require a permit or water right.
2.4.2 The use of water permits or water right
A water permit or water right typically defines:
• The source of from which the water is abstracted e.g. groundwater, watercourse,
reservoir;
• The point of abstraction of the water;
• The purpose for which the water will be used (e.g. irrigation of 500 ha of land).
A permit or right specifies certain conditions under which water use is permitted. A typical
condition is that the permit or right is limited in that it does not permit the use of water that
infringes on similar rights of others. Another condition frequently specified is that the water
should be used beneficially and not be wasted, and that return flows should adhere to certain
quality standards. Restrictions may also be placed on the permit during periods of drought.
2.4.3 Hierarchy of water use
Water laws often stipulate the hierarchy of different types of water use; distinguishing
between various sectors e.g. primary use (e.g. human consumption), environmental use,
industrial use, agricultural use, water for hydropower. In most countries water use for
primary purposes has priority over any other type of water use. Some countries also specify a
hierarchy of the remaining uses, whereby the most important economic use in that country
normally receives a high priority of use. In other countries all uses of water other than for
primary (and sometimes environmental) purposes have equal standing. In times of water
shortage the amount of water allocated to all non-primary uses will be decreased
proportionally, so that all these user share the shortage equally.
The law may provide more detailed stipulations with a direct bearing on the allocation of
water. The law may stipulate, for instance, that the allocation of water should be equitable. In
some countries, in contrast, the law directs that junior rights may not affect senior rights. In
most cases, however, the legal framework does not provide a detailed "recipe" of how the
water should be allocated. Water managers therefore have to interpret the more general
principles as laid down in the law, and translate these into operational rules for day-today
allocation decisions. In many countries the water manager may not even do this without
consulting all relevant stakeholders.
2.5 The value of water

The various uses of water in the different sectors of an economy add value to these sectors.
Some sectors may use little water but contribute significantly to the Gross National Product
(GNP) of an economy. Other sectors may use a lot of water but contribute relatively little to
that economy. Table 2.1 gives the contribution of the various sectors of the Namibian
economy to its GNP, and the amount of water each sector uses.
Table 2.1 Contribution of various sectors of the Namibian economy to GNP

and the quantity of water used by each sector
Sector Water uses Contribution to GNP

(%)
3
Million m /year Percentage (%)
Irrigation 107 43.0 3

Livestock 63 25.3 8
Domestic 63 25.3 27
Mining 8 3.2 16
Industry and commerce 7 2.8 42
Tourism 1 0.4 4
Total 249 100 100
Industry and commerce uses less than 3% of all water used in Namibia, but contribute 42% to
the Namibian economy. In contrast, irrigated agriculture uses 43% of all water used, but
contributes only 3% to the economy. Care should be taken when interpreting the above data.
For instance, it is well known that the agricultural sector typically has a high multiplier effect
in the economy, since many activities in other sectors of the economy depend on agricultural
output, or provide important input services. The "real" value added by water may thus be
underestimated by the type of data given in the table.
The added value of some uses of water are very difficult, if not impossible, to measure. For
instance the value of domestic use of water is very difficult to quantify. The value of irrigated
use of water is general at least a factor ten less than other types of industrial and commercial
uses.
The damage to an economy by water shortage may be immense. It is well known, for
instance, that a positive correlation exists between the Zimbabwe stock exchange index and
rainfall in Zimbabwe. The drought of 1991/92 had a huge negative impact on the
Zimbabwean economy. During the drought of 1991/92, the country’s agriculture production

fell by 40% and 50% of its population had to be given relief food and emergency water
supplies, through massive deep drilling programmes, since many rural boreholes and wells
dried up. Urban water supplies were severely limited with unprecedented rationing.
Electricity generation at Kariba fell by 15% causing severe load shedding. As a result
Zimbabwe's economy contracted by 11%.
Conversely, floods, though often beneficial, can sometimes be devastating. The February
2000 floods killed 800 persons in southern Mozambique. One million people required some
form of emergency assistance. About 20,000 cattle drowned and 140,000 hectares of crops
were destroyed. Road, rail and irrigation infrastructure was severely damaged. Health
centres as well as water supply and sanitation infrastructure in many towns and villages
suffered extensive damage, exposing many people to water-borne diseases such as cholera,
malaria and diarrhoea. The destruction caused by the floods is estimated at US$ 600 million.
Mozambique’s economic growth went down from 10% in 1999 to 2% in 2000.
2.6 Scales and boundary conditions

Any allocation decision potentially has third party effects: it may affect those not immediately
involved in the allocation process, either beneficially or detrimentally. A special case, and a
very important one, is where downstream users are affected that are located outside the
jurisdiction of a given water allocation institution.
Any allocation process that does not encompass the entire catchment runs the risk of being
affected by upstream uses and in turn impacting on downstream uses. Since most catchments
in southern Africa are simply too large in extent, and often shared by more than one country,
the water allocation processes is normally fragmented into sub-catchment areas which form
part of the larger catchment. In such cases the allocation process must include boundary
conditions; i.e. a specification of water requirements at the inlet and at the outlet of the
catchment area under consideration. Even for a catchment area, with its downstream
boundary being an estuary, will have to set such boundary conditions so as to minimise salt
intrusion, and/or ensure the health of the estuary for environmental, social and/or economic
purposes (e.g. for mangrove forests and prawn fisheries).
Boundary conditions are especially important in river basins that are shared by more than one
country. If an upstream water allocation institution does not consider the requirements of the
downstream country, it may even affect the bilateral relations of the two neighbouring
countries. It would be advisable to formalise such boundary conditions in writing and to get
them endorsed by all water allocation institutions involved; in a similar manner as how claims
of individual water users are formalised in water permits or rights.
The water allocation process should ideally consider both the detailed allocation decisions
between individual water users at the local level, as well as the "big picture" allocation
decisions covering the entire river basin. Obviously, these different spatial scales require
different levels of accuracy and specificity. However, they are both required, since decisions
at these different spatial scales affect each other.
Historically, the decision-making process has been iterative, with an initial focus on the
smaller spatial scales, especially in heavily committed parts of a basin. With the steadily
increasing pressures on our water resources, the interconnectedness between the various parts
of the basin have become apparent in many river systems. This has inevitably led to widening
the scope of the water allocation process also to the largest spatial scale.

2.7 Issues in water allocation
In this section some important issues directly related to water allocation are briefly discussed.
These issues typically cannot be solved overnight. Any stakeholder involved in water
allocation, however, must be aware of them.
2.7.1 Defining key concepts
Key concepts used in a country's water allocation system must be very precisely and clearly
defined, and be known and understood by the water users. Such key concepts may include:
• The ownership of water;

• Water use;
• Primary uses of water;
• Equity;
• Efficiency;
• The precise rights and obligations conferred with a water permit.
A particularly important issue is the definition of water use, since this basically defines the
point where water converts from a public to a private good. Lack of clarity about where
exactly this conversion occurs will create confusion, which will directly impact on the
effectiveness of the water allocation process. For instance, if a permit holder has lawfully
stored water in their dam, has this water already been used and hence is owned by the permit
holder or not yet? The South African Water Act (1998) defines water use as taking and
storing water, activities which reduce stream flow, waste discharges and disposals, controlled
activities (declared activities which impact detrimentally on a water resource), altering a
watercourse, removing underground water for certain purposes, and recreation.
2.7.2 Uncertainty
Generally speaking, if a user does not know how much water he or she is entitled to, and how
much water is likely to be available at a future time, he or she tends to over-use or hoard
water often incurring considerable losses. The allocation of water over different uses should
therefore aim to deal effectively with uncertainty and increase the predictability of water
available to the various uses. Increased predictability is an important condition that will allow
users to use water more efficiently. Even a better understanding of how unpredictable water
availability is will improve a user's ability to deal with this.
Two types of uncertainty may be distinguished:
• Physical uncertainty;
• Institutional uncertainty.
Physical uncertainty
Physical uncertainty does not so much refer to the stochastic nature of hydrological processes
(which is normally quite well understood), but more to the impact of human activities on the
hydrological cycle. At the global level, human-induced climate change is a possibility and
may have wide-ranging effects, but the specific effects are not yet well understood. At a
smaller spatial scale, the effects of land use change on the availability of blue water are
difficult to predict. Will a more efficient use of soil moisture for rain fed crop production
indeed translate into decreased blue water flows? The link between groundwater and surface
water abstraction is more straightforward. However, it is still difficult to predict the precise

effect of groundwater abstraction in a given location on the surface water availability
somewhere downstream.
The physical uncertainties mentioned here must be acknowledged. If a proper understanding

of such processes is lacking, in the first instance conservative estimates should be made on
possible impacts of certain interventions. The organisation responsible for water management
should then put in place a programme of data collection meant to gradually improve the
understanding of these dynamic processes.
Institutional uncertainty
A different type of uncertainty is created by the institutions that are involved in water
allocation. If the manner in which such institutions allocate water is unknown to the users or
not well understood by them, or seen as haphazard, then users may distrust the allocation
process. They will receive the wrong incentives and may, for instance, overstate their water
requirements, hoard water or even over-use it.
The institutional system of water allocation should therefore be transparent to users. All users
should know the principles and procedures guiding the allocation of water. Moreover, the
allocation process must treat all users in the same way. It must also be transparent, and
information on permits granted or permits refused must be freely accessible, not only to all
water users, but to the wider public as well. A fair and transparent allocation process will
enhance the individual users' trust in the process, and will increase their confidence in the
worth of their permits/rights to use water. Trust in the allocation process will enhance users
willingness to invest in water related infrastructure, and desist from "free-rider behaviour" in
times of water scarcity.
2.8 Efficiency and equity

It could be argued that equity, efficiency and ecological integrity should form the pillars of
any water management activity. Since water allocation is a major water management activity,
and so should inform water allocation decisions. Supposing that the environmental/ecological
water requirements are adequately taken care of, by assigning to the environment rights to
sufficient water with an acceptable ecological regime, then equity and efficiency remain.
Some people believe that there is a trade-off between the principles of equity and efficiency;
i.e. a more efficient allocation system may ignore certain issues of equity, and vice versa, a
more equitable allocation system may be less efficient. This is not necessarily true for all
situations. Some tentative definitions and some implications for water allocation are briefly
explored below.
2.8.1 Equity
Equity can be defined as affording everyone a fair and equal opportunity in the utilisation of
the resource according to one’s needs. Equitable access does not necessarily mean access to
equal quantities but rather equal opportunity to access water (Reference 2.9). Equity deals
with the distribution of wealth or resources among sectors or individuals of society.
2.8.2 Efficiency
Different definitions of efficiency can be used, depending on the objective to be achieved.

The reason why efficiency is important is that water is a finite and often scarce resource.
Generally, efficiency measures how much one can do with one unit of water. Economic

efficiency would then measure the benefits derived from a unit volume of water used. Water
use efficiency measures the amount of water actually used for a given use.
At a more abstract level, efficiency can also indicate to what extent the ensemble of technical,
legal, institutional, economic and other measures induce efficient use of the scarce water. For
instance, certain legal and institutional arrangements may enhance people's willingness to
privately invest in water infrastructure, or induce them to waste less water, or pollute less.
This will eventually lead to increased water use efficiency as well as increased economic
efficiency.
This wider definition of efficiency calls for pricing arrangements that ensure cost recovery of
water services. This will not only give the correct signal to water users, namely that water is
valuable and should not be wasted, but will also lead to the sustainability of infrastructure and
institutions. The wider definition of efficiency also calls for suitable legal arrangements that
provide users with sufficient security of water tenure, such that they are willing to invest in
water-related infrastructure.
2.8.3 Trade-offs
The principle of economic efficiency is often translated into proper pricing of water services.
This may obviously jeopardise the equity principle, in that poorer households may not be able
to buy such a service. The fact that poorer households are thus denied access to a basic
amount of water may however be extremely costly to society (e.g. in terms of disease, ill
health). From a societal perspective it may therefore be highly efficient to provide all
households with a very cheap (subsidised) lifeline quantity of water, and to make up the
financial shortfall through cross-subsidies. In this manner efficiency and equity in water
allocation systems may be achieved.
2.9 Water losses

Reducing water losses often has a high priority in attempting to balance demand with supply.
However, water losses should always be carefully and precisely defined. This is because it
depends on the scale and the boundaries whether water is considered a loss or not. At the
global scale no water is ever lost. At the scale of an irrigation scheme, a water distribution
efficiency of 60% indeed means that slightly less than half of the water is lost. Part of this
water, however, may return to the river and be available to a downstream user. At the scale of
the catchment, therefore, it is the transpiration of crops (40% in this example) that can be
considered a loss!
In many situations, and especially in irrigated agriculture, a reduction of water losses may not
free up the "saved" water. Even "real" water losses, such as when water is released from a
dam through the river bed for a downstream user, may provide an important service (e.g.
recharge of aquifers and water for the environment). Once such services are recognised and
formalised into permits (or in a "Reserve", as done in South Africa), the water manager may
sometimes be able to find solutions that are advantageous to a number of different parties. In
other cases, of course, this may not be possible. Analysing water losses should therefore
always:
• clarify the scale and boundaries at which the analysis is done

• acknowledge both the consumptive and non-consumptive parts of the water use under
consideration
• consider any other type of use (including the environment) that may benefit from the
water "lost".

2.10 Water allocation between sectors
Some types of water use add more value than other types. The classic case is the different
values attained in the agricultural and urban sectors: the value attained in urban sectors is
typically an order of magnitude higher than in agriculture (Reference 2.2). If water is
currently used in the agricultural sector, the opportunity cost, i.e. the value of the best
alternative use, may be ten times higher, subject of course of "location and the hydraulic
connections possible between users" (Reference 2.2). Thus a shift towards the higher value
use is often promoted.
Whereas the opportunity cost of water for domestic water use may be highest, the moment
availability is higher than demand, the opportunity cost of the water will fall to the next best
type of use. It is just not possible to consume all the water at the highest value use. The
proper opportunity cost for irrigation water may therefore be only half, or less, than the best
alternative use (Reference 2.5). Even then the reliability of supply acceptable to irrigated
agriculture is much lower than that for urban water supply: a storage dam yielding x m3 of
water supplied to irrigation at 80% reliability, may yield only 0.5x m3 (or less, depending on
hydrology) for urban water supplied at 95% reliability. The effective opportunity cost of
water used for irrigation should therefore again at least be halved. The resulting opportunity
cost is thus only a fraction of that some neo-classical economists claim it to be.
Figure 2.1 shows the variation of supply and demand in an imaginary case. It shows that, in
general, primary (domestic) and industrial demands, with the highest ability and willingness
to pay, require a high reliability of supply, which is normally achieved through relatively
large storage provision. Environmental demands are also not the most demanding on the
resource. Agricultural water requirements tend to be much higher, fluctuate strongly but also
accept a lower reliability of supply.
0
Water availability and demand
Reliability of supply
Agricultural demand
Urban and industrial
demand
Environmental demand
Primary demand
Water availability
1
Time
Figure 2.1 Variation of water availability and demand, and reliability of

supply
The emerging picture is that the sectors with highest value water uses should have access to
water. In many countries these sectors require only 20% to 50% of average water availability,

and these demands can easily be satisfied in all but the driest years. In most years much more
water will be available, and this water should be used beneficially, for instance for irrigation.
There is therefore no need for permanent transfers from agriculture to other sectors, except in
the most heavily committed catchment areas of the world. What is needed is a legal and
institutional context that allows temporary transfers of water between agriculture and urban
areas in extremely dry years. No market is required to cater for such exceptional situations.
A simple legal provision would suffice, through which irrigators would be forced to surrender
stored water for the benefit of urban centres against fair compensation of (all) benefits
forgone. In those heavily committed catchment areas where permanent transfers of water out
of the agricultural sector are required, normally voluntarily negotiated solutions can be
agreed, provided the laws allow this to happen.
2.11 Recommendations on water allocation

An important purpose of water management is to match or balance the demand for water with
its availability, through suitable water allocation arrangements. The balancing of water
demand with water availability is catchment specific and hence there is no one particular
method that can be recommended. The balancing of supply with demand will often involve a
process of decision making where difficult compromises have to be made. In all cases, the
water allocation process requires a sound quantitative understanding of both water availability
and water demand. Moreover, the following aspects should receive careful attention, and
possible win-win combinations sought:
• The constitutional obligation to provide a basic amount of fresh water to the population;
• The legal (or treaty) obligation to consider downstream requirements beyond the area
being considered for water allocation;
• The legal obligation to provide for environmental water requirements;
• Water losses should be analysed considering different spatial scales, and the unintended
functions these losses may serve;
• Allocation principles should include clear provisions for (extreme) drought situations;
• Allocation principles should promote water users' willingness to invest in water
infrastructure and to improve efficiencies.
2.12 References
2.1 Falkenmark, M. and Folke, C. (2002) The ethics of socio-ecohydrological catchment
management: Towards hydrosolidarity Hydrology and Earth System Sciences Volume
6 pp 1-9
2.2 Briscoe, J. (1996) Water as an economic good: the idea and what it means in practice.
Paper presented at the World Congress of the International Commission on Irrigation
and Drainage. September 1996 Cairo Egypt.
2.3 Pallett, J. (1997) Sharing water in Southern Africa. Desert Research Foundation of
Windhoek, Namibia
2.4 Pazvakawambwa, G., and van der Zaag, P. (2000) The value of irrigation water in
Nyanyadzi smallholder irrigation scheme, Zimbabwe. Proceedings of the 1st
WARFSA/WaterNet Symposium 'Sustainable Use of Water Resources'. Maputo
Mozambique November 2000
2.5 Rogers, P. (1998) Integrating water resources management with economic and social
development. Paper presented at the Expert Group Meeting on Strategic Approaches to

Freshwater Management on behalf of Department of Economic and Social Affairs,
United Nations. 27-30 January 1998 Harare, Zimbabwe.
2.6 Rogers, P. Bhatia, R. and Huber, A. (1997) Water as a social and economic good: how
to put the principle into practice. TAC Background Papers No.2. Global Water
Partnership, Stockholm.
2.7 Rosegrant, M.W. and Gazmuri Schleyer, R., (1996) Establishing tradable water rights:
implementation of the Mexican water law. Journal of Irrigation and Drainage Systems
Volume 10: pp 263-279.
2.8 Savenije, H.H.G. and van der Zaag, P. (2001), “Demand Management” and “Water as
an economic good”; paradigms with pitfalls. Paper presented at the international
workshop on non-structural measures for water management problems, London,
Ontario, 18-20 October 2001.
2.9 WRMS Zimbabwe (1999) Water Resources Management Strategy. Ministry of Rural
Resources and Water Development, Harare, Zimbabwe

3. ENVIRONMENTAL WATER DEMAND AND USE
3.1 Background
The environment is increasingly being considered a legitimate water user in many SADC
countries. As a consequence the water requirement of the environment needs to be estimated.
The amount of water that will be allocated to the environment is a decision made by society,
and is to some extent arbitrary. The quantity of water allocated to the environment will
always be less than what the environment ideally would require, namely the natural,
undisturbed, flow regime of a river. Society, therefore has to weigh the potential costs and
benefits to the environment and to all other water users, of allocating (or not) a certain amount
of water to the environment. In so doing, society accepts a certain modification of the natural
environment. This accepted level of modification may differ from river to river, and is
sometimes defined in terms of "ecological management classes".
The environmental or instream flow requirement is often defined as how much of the original
flow regime of a river should continue to flow down it in order to maintain the riverine
ecosystem in a prescribed state (e.g. pristine, good, satisfactory). However, an environmental
instream flow often fulfils a number of different functions. In addition to the ecology of a
watercourse there may be a need to recommend instream flow requirements for the following
reasons:
• Protection of the rights of other abstractors;

• Navigation;
• Prevention of saline intrusion;
• Dilution of effluent;
• Maintenance of the flood carrying capacity of the channel;
• Cultural and social reasons;
• Prevention of invasive plant species;
• Maintenance of the channel diversity.
There are numerous methods available for the assessment of environmental flows. These are
outlined below.
3.2 Environmental flow assessment methods

The environmental flow assessments are used as a method for estimating the quantity of water
required. An environmental flow assessment produces one or more descriptions of possible
future flow regimes for a river each linked to an objective relating to the condition or the
health of the riverine ecosystem. For example the requirement may be stated as “a water
depth of at least 50 cm is required throughout the year to provide adequate wetted perimeter
for a particular fish species”. Alternatively it may be more complex detailing a
comprehensive flow regime that specifies magnitudes, timing and duration of low flow and
floods at a number of temporal scales.
There is a range of methods available for assessing instream flow requirements based on:
• Simple hydrological indices;

• Hydrological simulations;
• Consensus and discussion based approaches;
• Historical data analysis;
• Biological response simulation techniques often referred to as habitat simulation methods.

Few, if any, of the approaches available provide a complete solution and hence a wide range
of approaches may be appropriate, especially for different levels of planning. The
environmental or instream flow requirement for a watercourse is the minimum flow required
to enhance or maintain aquatic and riparian life. There are several assessment procedures for
determining environmental flows. The decision on which method to use is dependent on the
following:
• Type of river (e.g. perennial, seasonal, high base flow, flashy);

• Perceived environmental importance;
• Complexity of the decision to be made;
• Increased cost and difficulty of collecting large amounts of information;
• Severity of different resource developments.
The level of detail required will be case dependent. In many countries a two-tier system is
used comprising catchment wide and scoping method for “level-one” studies and more
detailed methods for “level-two” studies. Level two studies move away from standard setting
(i.e. setting a single minimum flow) and towards an incremental approach (i.e. quantification
of varying instream requirements) that enable various management options to be assessed.
Stages in determining the minimum flow requirement may be as follows:
• Outlining of requirements;
• Data collection method;
• Modelling and analysis process, and the use of this information to set an instream flow
requirement in a rational manner;
• Use of tools in an active manner (e.g. reservoir releases);
• Follow up monitoring of success and revision of goals.
Knowledge concerning the environmental requirements of rivers is likely to remain

incomplete for the foreseeable future. As a consequence there will always be a danger that an
instream flow requirement will be set too low, resulting in damage to the riverine
environment or too high resulting in potential waste of resources or exploitation of other of
other more sensitive water resources. It should also be noted that too much water during
natural low flow periods could lead to undesirable changes especially in the arid and semi-
arid areas that exist in southern Africa.
The methods discussed in the Handbook are shown below.

Environmental flow assessment methods
Hydrological index methodologies discussed include:
• Tennant method;
• Texas method;
• Flow duration curve method;
• Aquatic base flow method;
• Range of variability approach.
Hydraulic rating methodologies discussed include:
• Wetted perimeter method.
Habitat simulation methodologies discussed include
• Building block methodology
Holistic methodologies discussed include:
• Instream flow incremental methodology.
Each method differs in its data requirements, procedures for selecting flow requirements,
ecological assumptions and effects on river hydraulics.
3.3 Hydrological index methods

Hydrological index methods are the simplest type of environmental flow assessment and rely
on the use of historical hydrological data for making flow recommendations. These data are
usually in the form of long-term, historical monthly or daily discharge records. These are
used to determine environmental flow requirements. Hydrological index methodologies are
the simplest and least data intense methods estimating instream flows. There are of the order
of 15 frequently referenced, hydrological based methods for environmental flow
requirements. However, many of these are specific to regions outside of southern Africa.
The most commonly used methods include:
• The Tennant (or Montana) method;

• Texas method;
• Annual minima;
• Flow duration curve analysis;
• Range of variability approach.
These methods are discussed below.
3.3.1 The Tennant method
The Tennant method is based on discharge statistics and historical flows. The minimum flow
requirement for a watercourse is expressed as a percentage of the mean annual naturalised
flow at a specified site. The naturalised flow regime is the hydrological regime of the
watercourse with the man-made influences (e.g. abstractions of water, changes in runoff
resulting from urbanisation) removed from the flow series. To produce a naturalised flow
series flow records are required. This series is then modified to remove man-made influences

thus giving a naturalised flow record. There are a number of hydrological textbooks that
describe in detail how to go out naturalising a flow series.
The Tennant method was developed to specify minimum flows for watercourses in the mid-
western USA. Percentages of the mean annual naturalised flow are specified to maintain the
riparian habitat in a particular state e.g. 10% for survival, 30% for a satisfactory healthy
ecosystem, 60% to 100% for a pristine ecosystem. It was developed using calibration data
from hundreds of watercourses in the USA (Reference 3.24).
There have been several modifications to the Tennant method by various practitioners since it
was first used in the USA in 1976. These include the following:
• Modifications for spring runoff;

• Equations to take into account existing flow modifications;
• Modifications to incorporate monthly minimum flow levels.
Where the Tennant method is used the following should be noted:
• The basic method takes no account of flow fluctuations and seasonal effects;
• The method is more suitable to large, perennial watercourses where flow variability is
less than for seasonal watercourses;
• No account is taken of the stream geometry;
• Recommendations should be compared to other flow statistics e.g. mean 10 and 30 day
naturalised low flows.
The Tennant method could provide a “model” for the development of minimum flow levels at
a catchment level for southern Africa. However, to modify the Tennant method so that it can
be used in the southern African context would require extensive fieldwork to be undertaken in
the region. This would entail both the collection of biological and hydrological data
throughout southern Africa to enable relationships between discharge and physical habitat
availability and suitability for aquatic biota to be established. The method would also not be
applicable for semi-arid and arid regions where watercourses are dry for several months of the
year as it is likely to result in flows that are too high.
Advantages of the Tennant method
The main advantages of the Tennant method are:
• It is simple to use;
• Once relationships between discharge and the aquatic environment have been established
it requires relatively little data;
• It does not require costly fieldwork to be carried out.
Disadvantages of the Tennant method
The Tennant method has the following disadvantages:
• It does not preserve the natural variability of the watercourse by taking account of daily
and yearly variation of flows i.e. the method only prescribes a minimum environmental
base flow;
• The naturalised flow regime (i.e. the regime before any anthropogenic influences on the
watercourse have occurred) has to be established;

• The method never produces a zero flow recommendation. However, in semi-arid regions
where watercourses are naturally dry for some months of some years a zero flow may be
appropriate;
• The method is based on fieldwork carried out in the USA. This fieldwork is not
applicable to semi-arid regions of the world such as southern Africa. The Tennant
method could act as a model for southern Africa, however, this would entail the collection
and correlation of both biological and discharge data for the region;
• The relationship between flow and the state of the aquatic ecosystem is poorly
established;
• The method is site specific.
3.3.2 The Texas method
This method uses variable percentages of the monthly median flows. The percentages are
calibrated to regions with characteristic fauna taking into account results from previous fish
inventories and known life history requirements. The Texas method is an advancement over
existing preliminary planning methods. This is because it is the first such technique to treat
the recommended flow percentage for each month as a variable along with the biological
characteristics (e.g., spawning/incubation periods) and regional hydrological characteristics
(e.g. highly variable monthly flows with positive skewness) (Reference 3.7).
3.3.3 Flow duration curve analysis
In flow duration curve analysis naturalised or present-day historical flow records are analysed
over specific durations to produce curves displaying the relationship between the range of
discharges and the percentage of time each of them is equalled or exceeded. For example in
some cases the 90 percentile flow (Q90) may be set as the minimum environmental flow. This
is the flow that is exceeded 90% of the time. A typical example of a flow duration curve is
shown in Figure 3.1. However, to apply such a flow duration curve technique, hydrological
flow data are required.
500
450
400
350
Discharge (m3/s)
300
250
200
150
100
50
Ninety percentile flow Q90= 18 m3/s
0
0 10 20 30 40 50 60 70 80 90 100
Percentage time
Figure 3.1 Flow duration curve method

3.3.4 Aquatic base flow method
The aquatic base flow method is based on the assumption that the median flow for the lowest
flow month is adequate throughout the year for fisheries, unless additional flow is required to
meet the needs of spawning and incubation. An example of the method is given below.
Example of the aquatic base flow method
The Lukozi River in Zimbabwe has the following monthly median flows:
Month Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Flow 14 12 9 8 5 2 1.5 2 3 4 7 13
3
(m /s)
3
Using the aquatic base flow method the minimum environmental flow requirement is 1.5 m /s.
It should be noted that many rivers in southern Africa are ephemeral and will have no flow
during the dry season. For such watercourses the aquatic base flow method will yield a flow
of zero. Hence the aquatic base flow method is unlikely to be applicable for many
watercourses in the SADC region.
3.3.5 Range of variability approach (RVA)
The range of variability approach is the most sophisticated form of the hydrological index
methodologies. It is aimed at providing a comprehensive statistical characterisation of the
ecologically relevant features of the flow regime, recognising the crucial role of hydrological
variability in maintaining ecosystems. The method is intended to be applied to rivers where
protection of the natural ecosystem functioning and conservation of the natural biodiversity
are the primary management objectives. The methodology comprises six basic steps, the first
of which is the characterisation of the natural range of hydrological variation using a number
of ecologically relevant hydrological indices, termed Indicators of Hydrologic Alteration
(IHA). These are summarised in Table 3.1.

Table 3.1 Summary of hydrological parameters used in the Indicators of
Hydrologic Alteration
IHA statistics group Regime Parameters

characteristics
Group 1: Magnitude of Magnitude Mean value for each calendar month

monthly water Timing
conditions
Group 2: Magnitude and Magnitude Annual minima and minima 1 day means
duration of annual Duration Annual minima and minima 7 day means
extreme water Annual minima and minima 30 day means
conditions Annual minima and minima 90 day means
Group 3: Timing of Timing Julian date of each annual 1 day minimum
annual extreme water and maximum
conditions
Group 4: Frequency and Frequency Number of high and low pulses each year
duration of high and low Duration Mean duration of high and low pulses
pulses
Group 5: Rate/frequency Rates of change Means of all positive differences between
of consecutive water daily values
condition changes Means of all negative differences between
daily values
The second step is to select management targets for each of the IHA parameters. The
fundamental concept is that the river should be managed so that the annual value of each IHA
parameter falls within the range of natural variation of that parameter. The management
targets should be based on available ecological information. In the absence of adequate
ecological information it is recommended that ±1 standard deviation is used as the default for
the initial setting of targets.
Step 3 is to use the flow based management targets, known as the Range of Variability (RVA)
to set up management rules that will enable the targeted flow conditions in most, if not all,
years. Step 4 involves implementing a monitoring programme to assess the ecological effects
of the new management system. The fifth step is to characterise the actual stream flow
variation using the same hydrologic parameters and compare then to the RVA targets. The
final step is to repeat the first five steps incorporating the results of the preceding years’
management and any new ecological research or monitoring information to revise either the
management system or the RVA targets.
The RVA method requires at least 20 years of flow data. If 20 years worth of data are not
available it is usually necessary to extend the record. In some cases hydrological simulation
models may be used. The RVA approach was designed to bridge the gap between applied
river management and current aquatic ecology theories (Reference 3.20).
3.3.6 Unsuitable historical discharge techniques.
A hydrological technique that is not acceptable for establishing instream flows for fish is the 7
day 1 in 10 year low flow (expressed as 7Q10). This statistic was developed in the USA to
ensure wastewater treatment plants did not violate water quality standards during droughts. It
does not address instream environmental requirements.

3.3.7 Advantages of hydrological index methodologies
The advantage of hydrological methodologies are as follows:
• They are relatively simple to use;

• They are an appropriate method to use for use at a catchment level for establishing low-
resolution estimates of the quantity of water required
• Relatively little site-specific data are needed, in many cases only historical flow records
are required;
• It does not require costly fieldwork to be carried out;
• Hydrological indices that are derived from such methods can be incorporated into as sub-
components of holistic type methods.
3.3.8 Disadvantages of hydrological index methodologies
The disadvantage of hydrological index methodologies are as follows:
• Although the hydrological index methodologies require a relatively low level of resources
per site they require a large amount of field work to be carried out in order to set the
various standards and parameters required;
• These methods should only be applied for high level scoping studies.
3.4 Hydraulic methods

Hydraulic methods relate various parameters of the hydraulic geometry of a watercourse
channel to discharge. The most commonly used hydraulic method is the wetted perimeter
technique.
3.4.1 The wetted perimeter technique
Background
The wetted perimeter method is the simplest of the field survey-based, site-specific
techniques that allows the minimum instream flow of a watercourse to be calculated
(Reference 3.25). It should be noted that the wetted perimeter technique includes no explicit
representation of the aquatic habitat. To establish the minimum environmental flow a wetted
perimeter-discharge relationship is generated. The wetted perimeter of a watercourse is
defined as the length of the line of intersection of the channel wetted surface with a cross-
sectional plane normal to the direction of the flow.
The wetted perimeter-discharge relationship should be generated for watercourse cross-

sections that are at riffle sites or at sites where fish passage is likely to be limited. A riffle is
an area of shallow rapids in an open stream where a turbulent water surface is induced by
obstructions that are wholly or partly submerged. A photograph of a riffle on a South African
watercourse is shown in Figure 3.2.

Figure 3.2 Photograph of a riffle on a South African watercourse
The method assumes that preserving the wetted perimeter in critical habitat areas such as
riffles, adequate flow will be available to maintain aquatic life. The wetted perimeter method
is illustrated in Figure 3.3 and applied as follows:
• The relationship between the wetted perimeter and the discharge of the watercourse at a
riffle, or where the passage of fish is limited, is established;
• A non-dimensional graph of wetted perimeter versus discharge is plotted. The values of
wetted perimeter and discharge are expressed as a proportion of their maximum value;
• The breakpoint of the curve is established. The breakpoint indicates where small
decreases in the flow result in increasingly greater decreases in the wetted perimeter. The
breakpoint on the wetted perimeter-discharge curve can be mathematically defined as the
point where the curvature of the curve is 45o. It should be noted that it is not possible to
reliably assess the breakpoint by eye. The breakpoint should be taken at the point where
the slope of the curve is one (dy/dx = 1) (Reference 3.25);
• Compound cross-sections with multiple benches may produce an irregular relationship
between wetted perimeter and discharge, and there may be more than one breakpoint. In
these cases the lowest breakpoint is usually the most relevant to minimum flow
determination;
• Once the breakpoint has been established the minimum instream flow requirement can be
estimated.

P1 Dimensionless wetted
P2 P3 P4
perimeter P/Pmax
Breakpoint
Minimum
environmental
flow
Dimensionless discharge Q/Qmax
Figure 3.3 Use of the wetted perimeter method to estimate instream flows
An important consideration in site selection is the ease with which the flow through the site
can be measured or calculated. Discharge through a riffle is relatively difficult to measure
directly with any confidence by manual flow gauging. It is often better to measure the
discharge at a nearby site suitable for manual gauging or use readings from a nearby gauging
station. These surrogate discharge measurements should be sufficiently close to the site of
interest that any losses or inflows between the two sites could be ignored.
It should be noted that compound cross-sections with multiple benches will produce an
irregular relationship between wetted perimeter and there may be more than one breakpoint
where the slope is unity. The lowest breakpoint is probably the most relevant to minimum
flow determination.
Data requirements
• Surveyed watercourse cross-sections at appropriate locations e.g. riffles;

• Measured or generated stage-discharge curve for the surveyed cross-section.

Example of the wetted perimeter method
Current metering has been carried out for a number of discharges for the Hwadzi River in
Zimbabwe at a point in the watercourse where the passage of fish is limited. A surveyed
cross-section of the Hwadzi River at the point, shown in Figure 3.4, where the discharge
measurements have been made is available.
1.0
0.8
0.6
Depth
0.4
0.2
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Chainage
Figure 3.4 Surveyed cross-section of the Hwadzi River
An estimate of the minimum instream environmental flow requirement is made as follows:
(i) A graph of wetted perimeter versus discharge is plotted.

(ii) The wetted perimeter and discharge data are non-dimensionalised with respect to
their maximum values.
(iii) A relationship is established between the dimensionless wetted perimeter and
discharge. This is shown in Figure 3.5. In this example a curve was fitted and the
wetted perimeter was found to be related to the discharge by the equation P =
0.08lnQ + 1.

1.0
0.9
Non-dimensional wetted perimeter P/Pmax
0.8
0.7
0.6
P = 0.08lnQ + 1
0.5
R2 = 0.972
0.4
0.3
0.2
0.1
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Non-dimensaional discharge Q/Qmax
Figure 3.5 Non-dimensional wetted perimeter versus discharge relationship
The breakpoint is calculated. The breakpoint is defined by the point on the curve where the
dy 0.08
gradient is one (i.e. = 1 ). Hence given that P = 0.08lnQ + 1 differentiating gives 1 = .
dx Q
Hence the minimum value of flow to meet the instream flow requirement is:
3
Qinstream = 0.08Qmax = 0.08 x 16 = 1.3 m /s
In cases where it is not possible to fit a simple relationship to the data, the slope of the
relationship should be calculated for each plotted point until the point where the gradient
equals one is established.
Source: Adapted from Reference 3.25
Advantages and disadvantages of the wetted perimeter method
The main advantages of the wetted perimeter method are:
• It is relatively simple to use;

• It requires relatively little data.
Disadvantages of the wetted perimeter method
The wetted perimeter method has the following disadvantages:
• It recommends only a minimum environmental base flow;

• The method is site specific.

3.5 Holistic methods
The most widely used holistic method in southern Africa is the building block methodology.
This methodology for determining instream flow requirements is outlined below.
3.5.1 The building block methodology
The building block methodology was developed in South Africa by the Department of Water
Affairs and Forestry and various academic institutions (Hughes and Münster 1999). The
building block methodology requires the following:
• The total flow volume of the following four building blocks components:
- Low flows;
- Habitat maintenance floods;
- Channel maintenance/Flushing floods;
- Spawning migration flows;
• Monthly distribution of the four building block components;
• Establishment of the present ecological state (A to F) and future management category (A
to D). These are given in Table 3.2.
Table 3.2 Ecological states used in the building block methodology
Class Description of ecological state
A Unmodified
B Largely natural with few modifications
C Moderately modified
D Largely modified
E Natural habitat loss extensive
F Modifications at a critical level
The major objective of the method is to estimate the values of the four building block
components as a percentage of the mean annual runoff of the natural flow regime. A building
block instream flow study would be carried out as follows:
(i) The monthly naturalised flow series for the site of interested must be established.
(ii) The ecological management class of the site is established with A being an
unmodified site and F representing a site that had been modified to a critical level.
There are methods for estimating the ecological management status of the site based
upon various habitat integrity indices developed in South Africa.
(iii) The flow variability has to be established to summarise the variability within the wet
and dry seasons. This is based on the average coefficient of variation (i.e. standard
deviation/mean) for the three main wet season months and the three main dry season
months (excluding those that have zero mean monthly flows). The actual coefficient
of variation (CV) is the sum of these two means. The assumption is that rivers with a
high degree of variability in their flow regime will require a lower proportion of their
natural mean annual runoff because they are used to experiencing such conditions.
Rivers with more reliable flows and less flow variation are assumed to be ecological
less well adjusted to frequent extremes in the flow regime.

(iv) The base flow is calculated. The base flow index (BFI) is the proportion of the total
flow occurring as the base flow. There are various hydrological methods to assess the
base flow.
(v) A combined variability index is calculated by dividing the coefficient of variation by

the base flow index.
(vi) For particular sub-catchments, curves can be constructed for maintenance low flow
estimation and maintenance high flows versus the variability index (CV/BFI) for the
four future ecological management classes. A typical set of curves is shown in Figure
3.6.
(vii) The drought low and drought high flows are established.
(viii) The monthly distribution of flows is then produced. It should be noted that one of the
basic principles of the approach is that a higher proportion of the natural monthly
flow is required during the dry months than during the wet months.
Percentage of natural MAR to meet
maintenance low flow
Management
class
B
C
D
CV
BFI
Figure 3.6 Typical curves to establish the maintenance of low flows
Rhodes University in South African has produced a piece of software to automate the
building block methodology. However, this software can only be applied in South Africa,
Swaziland and Lesotho. A hypothetical instream flow requirement created using the Building
Block Methodology is shown in Figure 3.7.

Channel maintenance/
Flushing flood
(Third building block)
Habitat maintenance flood
(Second building block)
Discharge
Spawning migration flows

(Fourth building block)
Low flow
(First building block)
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Month
Figure 3.7 Examples of the flow building blocks used in the building
block methodology
Advantages of the building block method
The building block methodology has the following advantages:
• It takes into account the monthly flow variability for both high and low flows;
• It has been developed specifically for use in southern Africa;
• The low flow building block can be used to assess preliminary instream flow
requirements.
Disadvantages of the building block method
The building block methodology has the following disadvantages:
• It is site specific;
• It requires an estimation of base flow, the natural mean annual runoff and naturalised
flows.
3.6 Habitat methods

Habitat rating methods provide the most complex and the most flexible approach to
environmental flow assessments. They provide information on how habitats change with flow
for instream uses, either biological or recreational. No prior assumptions are made about the
state of the natural ecosystem. Changes of physical habitats with stream flow are accounted
for and combined with the habitat preferences of a given species to determine the amount of
habitat available over a range of stream flow conditions. The result is a curve relating
available habitat area and stream discharge. Optimum stream flows for a certain number of
species can be ascertained from these curves, and the results can be used as a guide for
recommending environmental flows. The most commonly used habitat method is the
Instream Flow Incremental Methodology (IFIM). This is discussed below.
3.6.1 Instream Flow Incremental Methodology
The Instream Flow Incremental Methodology (IFIM) is a conceptual framework for assessing
the effect of water resources development or management activities on aquatic and riverside
ecosystems, and for solving water resources management problems and conflicts that involve
the definition of an ecological flow to minimise impacts on ecosystems. IFIM is a collection
of analytical procedures and computer models that allows the development of a different
approach for each problem and situation. The goal of this method is to relate fish and wildlife
parameters to stream discharge in equivalent terms to those used to estimate other beneficial
uses of water.
IFIM is based on the assumption that living organisms in running water have their distribution
(longitudinally and laterally) controlled by the hydraulic conditions. The decision variable
generated by IFIM is the total habitat area with suitable conditions for a species at a particular
life stage or for a particular activity (e.g. spawning), computed as a function of discharge.
The environmental flow is usually the highest value of a range of minimum flows computed
for several species, assuming that this value will be adequate for the preservation of the
ecosystem. The target species, one or more, are usually game, commercial, endangered or
indicator fish species (Reference 3.19).
IFIM relates changes in the extent of habitats that are available to aquatic species to changes
in discharge. This allows instream flow demands to be expressed in the same terms as other
water resource demands. The IFIM methodology is usually coupled with the Physical Habitat
Simulation System (PHABSIM) model to generate a habitat-discharge relationship. IFIM
coupled with PHABSIM can be used to predict changes in almost any environmental
parameter that can be quantified in the form of a flow dependent relationship (Reference
3.19).
PHABSIM is a collection of computer programs that combine aquatic organisms preferences

for velocity, depth and channel conditions to predict habitat availability at various discharge
levels. It simulates the relationship between stream flow and available physical habitat
(defined by depth, velocity, substrate and cover). For each life stage of the target species, the
model requires expressions of the relative suitability for that species of the full range of
values taken by these variables. These univariate curves are called habitat suitability indices;
they may be derived from existing literature, expert opinion or by sampling techniques such
as electro-fishing or snorkelling.
PHABSIM simulates suitable habitats for aquatic species by using depth, velocity and stream
channel characteristics to describe local physical niches that are occupied by aquatic species.
This method develops a measure of physical habitat called Weighted Usable Area (WUA).
The PHABSIM methodology can be summarised as follows:
• The cross-section of the watercourse is divided up into a number of cells;

• Depth, velocity, flow area and cover conditions are measured or simulated for a given
discharge;
• Suitability Index Criteria (SI) are used to weight the area of each cell for the discharge;
• The habitat values for each of the cells in the study reach are summed to obtain a single
habitat value for the discharge;

• The procedure is repeated for a range of discharges and a graph of Weighted Usable Area
(WUA) versus discharge is produced for a particular life stage of the indicator species;
• The WUA is the aggregate product of the Suitability Use Index Criteria (SI) and the area
of each stream cell developed using stream gauging techniques. It is calculated as the
summation, over all stream cells in a study reach, of the product of cell area and SI for
depth, velocity and stream channel characteristic index (cell area x SI depth x SI velocity
x SI channel index). It represents a description of the amount of equivalent optimal
habitat occurring within a standard length of stream (Reference 3.19);
• Calculating the WUA for numerous discharges allows the relative habitat suitability of a
watercourse under different flow conditions to be assessed.
PHABSIM allows changes in habitat resulting from changes in instream flow to be quantified
and thus provides answers to “what if” water management questions. Figure 3.8 shows how
PHABSIM calculates habitat values as a function of discharge.
Site-specific microhabitat data 1.0 Habitat suitability data
Suitability index
Suitability index
1.0
0.8 0.8
Cross-section B 0.6 0.6
0.4 0.4
0.2 0.2
Flow
0.0 0.0
Cross-section A 0 1 2 3 4 5 0 0.2 0.4 0.6 0.8 1.0
Velocity (m/s) Depth (m)
Suitability index
1.0
0.8
0.6
V1 V2 V3 …………………………..velocity
D1 D2 D3…………………………..depth 0.4
C1 C2 C3…………………………..cover
0.2
A1 A2 A3………………………..….area
0.0
0 0.2 0.4 0.6 0.8 1.0
Cover
Seasonal relationship
100,000
between discharge and
microhabitat
for each life stage
usable area
Weighted
0
0 100
Discharge
Figure 3.8 Conceptualization of how PHABSIM calculates habitat values

as a function of the discharge
PHABSIM contains a number of standard, one-dimensional, steady-state, open channel

hydraulic models to predict values of depth and velocity at different simulation discharges.
These are combined with habitat models to describe Weighted Usable Area (WUA) under a
variety of channel configurations and flow management conditions. The use of hydraulic
simulation models allows physical habitats to be described for unmeasured discharges. The
models require calibration using field data collected at two or more calibration discharges and

hence is fairly data intensive. Observations of substrate and cover are recorded using a
coding system and are assumed to be independent of discharge. Once calibrated the model
can simulate values of microhabitat variables over the full range of discharge within a river
reach, combining the results with habitat suitability data to produce WUA versus discharge
relationships.
Simulated values of microhabitat variables (from the calibrated hydraulic model) are
combined with habitat preference data for each target species for each stage of life.
Combining this with a time series of historical flows yields a time series of available physical
habitat for each life stage of the target species. An example is shown in Figure 3.9. Using
PHABSIM it is possible to simulate habitat curves relating to season and complete life cycles
of target species. By relating habitat to discharge, PHABSIM provides a quantitative entity,
allowing river ecologists to negotiate prescribed flows in equivalent terms to other water
resource demands.
12
10
Weight usable area (m2/m)
6
Adult
Juvenile
4 Fry
Environmental flow required
for various life stages of salmon
2
0 5 10 15 20 25 30 35 40
Discharge (m3/s)
Figure 3.9 Example of weighted usable area versus discharge curves for
various life stages of a salmon
In summary, PHABSIM predicts the impact of changing flows on fish, invertebrates and
macrophytes and predicts physical habitat change and quantifies it in respect of the ecological
value of habitat losses/gains. Hydraulic modelling techniques such as PHABSIM require
detailed hydraulic and morphological surveys and knowledge of the habitat preferences of the
species of interest.

Data requirements
The data requirements to carry out an IFIM study using PHABSIM are dependent upon the
level of detail to which the analysis is carried out. A minimum of the following data is
required:
• Historical stream flow data;

• Stream flow measurements;
• Channel geometry measurements;
• Indicators of habitat suitability for each species (from literature, expert opinion, field
sampling);
• Critical depth of flow per species;
• Critical velocities of flow per species.
In addition the following information may be required:
• Differences in critical indicators over lifetime of species;

• Flow effects on inter-species relations;
• Water quality standards and dilution;
• Local land use;
• Location of potentially polluting industries;
• Pollution types;
• Dilution standards;
• Proximity of agriculture;
• Seasonal variations (emphasis on dry season flows and floods);
• Aesthetic needs and standards;
• Cultural needs and regulations;
• Recreational needs.
Advantages
• If employed correctly IFIM allows the values of every legitimate stakeholder to be taken
into account;
• The method takes into account the flow requirements of the indicator species over its
entire life cycle;
• An assessment of the natural flow requirement can be made independently of the
naturalised flow data.
Disadvantages
• It is time consuming and costly to carry out an IFIM study;

• The IFIM techniques have been mainly applied at micro- and meso-habitat levels
focusing on one or a few river reaches. There is little experience of using IFIM
techniques to assess environmental water demands at a catchment or even a sub-
catchment level;
• The method is site specific;
• To implement IFIM requires a multi-disciplinary team with expertise in hydrology, river
morphology, water quality, aquatic and terrestrial ecology, carrying out biological field
surveys and hydraulic engineering.

3.7 Comparison of the instream flow methods
Historical and hydraulic techniques such as the Tennant Method and wetted perimeter
technique are applicable for establishing minimum environmental demands for high level
water resources management. It should be noted that these techniques provide an initial “low
confidence” estimate. These techniques can be applied rapidly at a large number of sites to
provide a first estimate of the likely quantities of water required to maintain the ecology in a
given condition. The building block methodology can also be used for rapid assessments.
However, in order to use this method for rapid environmental flow appraisals, monthly
naturalised flow series are required and a country specific piece of software (possibly based
on the one produced by Rhodes University in South Africa) needs to written.
The IFIM method utilising PHABSIM is a commonly used method for more complex
decisions e.g. the construction of a hydropower plant or the setting abstractions limits from an
ecologically sensitive watercourse. In conclusion there is no one methodology that should be
used for establishing the instream flow demand. Figure 3.10 indicates that the choice of the
method used is a function of the complexity of the decision to be made and the complexity of
the system. To establish the necessary instream flow the practitioner should consider the
history and purpose of the various techniques available and must use this knowledge to make
an informed choice of the best method to use.
Habitat
rating
Complexity of the decision
methods
Building
block
methodology
Historical
discharge
methods
Complexity of the system
Figure 3.10 Spectrum of instream flow methodologies

3.8 Lake level requirements
There are many catchments in southern Africa where freshwater lakes form an important part
of the hydrological system. In many cases it will be important to define water level
requirements for lakes. A recent study carried out for the Mhlatuze catchment in northern
KwaZulu-Natal in South Africa utilised the following method to establish water level
requirements for freshwater lakes.
• A Present Ecological State (PES) was allocated to each lake. For each PES an Ecological
Management Class (EMC) was prescribed. The PES was described in terms of six
classes with A being near pristine and F being irreversibly changed;
• Each lake’s importance from a social/cultural and ecological point of view was then
established. This was considered when determining the EMC;
• After a process of public consultation the EMC was allocated to each lake. The EMC
ranges in classes from A (near pristine) to D (largely modified). Unlike the PES the EMC
does not extend to Classes E and F. This is because water resources that are currently in
Classes E and F are not considered to represent sustainable systems and must therefore be
protected and managed for improvement
For the freshwater lakes to establish a sustainable water level the following is required. A
range of water levels needs to be determined for:
• The drought year minimum;

• The maintenance year dry season minimum;
• The management maximum.
Other factors that also need to be determined include:
• Duration and frequency of water levels;

• Maximum rates of drawdown;
• Longer term fluctuations in water level.
The following should then be carried out for the freshwater lake:
• Identify the reference conditions;

• Determine the present operation of the lakes for the provision of water;
• Assess the present state for each of the ecological components;
• Assess the habitat integrity of the system for water body and the littoral/riparian zone;
• Determine the ecological importance;
• Determine the social importance;
• Assess the Attainable Ecological Management Class (AMEC);
• Recommend the EMCs either side of the AMEC and list the flow related and non-flow
related activities that would be required to meet these classes;
• Prioritise and list the objectives to attain the AMEC;
• Recommend the water levels required to achieve the AMEC and motivate these levels
based on ecological grounds backed up by hydrological records where available;
• Specify the degree of confidence in the recommendations and identify further work
required to firm up the recommendations;
• Link the results to the system hydrology to determine realistic assurances of supply for
the recommended drought and maintenance year water levels and to extrapolate these
levels, through the use of hydrology, to the EMCs on either side of the AEMC.

3.9 Methodologies for assessing ecological flow
requirements of wildlife
There has been little research carried out on producing techniques for assessing the ecological
flow requirements of wildlife. There are no Stage I or reconnaissance level or guidelines for
the assessment of environmental flows. There have been some limited studies carried out to
assess the influence of environmental flows on the availability and quality used by birds.
There have also been limited cases where the IFIM has been used to assess the flow
requirements of semi-aquatic mammals.
3.10 References
General
3.1 Alves, H.L. (1994) Instream flow determination for mitigating environmental impact
on fluvial systems downstream of dams and reservoirs, Advances in Water Resources
Management, Balekema, The Netherlands,
3.2 Cambray, J.A. (1991) The effects of fish spawning and management implications of
impoundment water releases in an intermittent South African river, Regulated Rivers:
Research and Management, Vol. 6, pp39-52
3.3 Dunbar, M.J., Gustard A., Acreman, M.C.and Elliott, C.R.N (1998) Overseas
approaches to setting river flow objectives Environment Agency R&D Technical
Report W145
3.4 Gherke, P.C., Brown, P., Schiller, C.B., Moffat, D.B. and Bruce, A.M. (1995) River
regulation and fish communities in the Murray-Darling River system Australia,
Regulated Rivers: Research and Management, Vol. 11, pp363-375
3.5 Hughes, D. Watkins, D., Münster, F. and Cobbing (1998) A contribution to the
preliminary reserve methodology for South African Rivers, Institute of Water
Research, Rhodes University, South Africa
3.6 Jordanova, A.A., Rowlston, W.S., Birkhead, A.L. and James C.S. (1998) The role of
hydraulics in holistic instream flow requirement assessment, University of
Pietermaritzburg, KwaZulu-Natal, South Africa
3.7 King, J. Tharme, R., Brown, C., Definition and implementation of instream flows
Final Report for the World Commission on Dams, September 1999
3.8 Moseley, P.M. (1982) Analysis of the effect of changing discharge on channel
morphology and instream uses in a braided river, Ohau River, New Zealand, Water
Resources Research Vol. 18 No. 4 pp800-812
3.9 Orth, D.J. and Leonard, P.M. (1990) Comparison of discharge methods and habitat
optimisation for recommending instream flows to protect fish habitat, Regulated
Rivers: Research and Management, Vol. 5, pp129-138
3.10 Ward, J.V. and Stanford, J.A. (1993) Research needs in regulated river ecology,

3.11 Walker K.F. and Thomas, M.C. (1993) Environmental effects of flow regulation on
the Lower River Murray, Australia Regulated Rivers: Research and Management,
Vol. 8, pp10-119
Building block methodology
3.12 Bullock, A., Maviya, J., Topping, C., and Tobani (1998) Environmental flow
requirements in Zimbabwe, Water Resources Management Strategy background
paper
3.13 Hughes and Münster, F. (1999) A decision support system for an initial ‘low
confidence’ estimate of the quantity component for the reserve of rivers, Institute of
Water Research, Rhodes University, South Africa
3.14 King, J.M. and Louw, D. Instream flow assessment for regulated rivers in South
Africa using the Building Block Methodology Aquatic Ecosystem Health and
Management 1 pp 109-124
3.15 Tharme, R.E. King, J.M. (1998) Development of the building block methodology for
instream flow assessments and supporting research on the effects of different
magnitude flows on riverine ecosystems Water Research Commission, South Africa
Instream flow incremental methodology and Physical Habitat Simulation

System (PHABSIM)
3.16 Gan, G. and McMahon (1990) Variability of results from the use of PHABSIM in
estimating habitat area, Regulated Rivers: Research and Management, Vol. 5, pp233-
239
3.17 Gore, J.A., King, J.M. and Hamman, K.C.D. (1991) Application of the instream flow
incremental methodology to Southern African rivers: Protecting endemic fish of the
Olifants River, Water SA Vol 17 No 3 pp225-236
3.18 Hearne, J., Johnson and Armitage, P. (1994) Determination of ecologically

acceptable flows in rivers with seasonal changes in the density of macrophyte,
3.19 US Department for the Interior (1995) The Instream Flow Incremental Methodology
– A Primer for IFIM, Biological Report 29, National Biological Service, US
Department of the Interior, USA
Lake levels
3.20 Department for Water Affairs and Forestry South Africa (2000) Report on lake water
level requirements supporting report no.2 (unpublished)
Range of variability approach
3.21 Richter, B., Baumgartner, J. and Baun, D. (1997) How much water does a river need?
3.22 Richter, B. and Baun, D. (1998) A method of assessing hydrologic alteration within
ecosystems
3.23 Richter, B., Baumgartner, J. Baun, D. and Powell, J. (1998) A spatial assessment of
hydrologic alteration within a river network
Tennant method
3.24 Tennant, D. L. (1976) Instream flow regimes for fish, wildlife, recreation and related
environmental resources Fisheries 1:6-10
Wetted perimeter method
3.25 Gippel, C.J. and Stewardson M.J. (1998) Use of the wetted perimeter in defining the
minimum environmental flows, Regulated Rivers: Research and Management, Vol. 14,
pp53-67

4. AGRICULTURAL WATER DEMAND AND USE
4.1 Introduction
Agriculture is an important activity in southern Africa in terms of food security, economic
activity and water use. Irrigated agriculture, moreover, plays a disproportionately important
role because it is generally two to three times more productive than rain-fed agriculture, and
because irrigation uses approximately 70% of the region's water supplies. It has been estimated
that about 1.8 million hectares in the SADC region are under irrigation. This is approximately
7% of arable land in the region. Irrigation is largely reserved for high value crops such as fruits
and vegetables, but sugar cane occupies 292,000 ha alone. Other irrigated crops include wheat,
rice, cotton, maize, coffee, tea and tobacco. SADC countries rely on irrigation to varying
degrees, from 0.5% of agriculture in Botswana to 36% in Swaziland. Table 4.1 provides details
of the percentage of irrigated land and water use for continental SADC countries.
Table 4.1 Irrigated land and water use for continental SADC countries
Country Irrigated area as a Percentage of total water use

percentage of arable (%)
area in 1993 (%)
Angola 2.5 76
Botswana 0.5 48
Lesotho 0.9 56
Malawi 1.7 86
Mozambique 4.0 89
Namibia 0.9 68
South Africa 10.3 62
Swaziland 35.8 71
Tanzania 5.0 89
Zambia 0.9 77
Zimbabwe 7.0 79
As Table 4.1 illustrates the agricultural sector of SADC has the largest water demand and use
owing to crops being grown under irrigation. The irrigation water demand and use can be
established from:
• Estimates using empirical formula;

• Measurements of water consumption from flow gauging devices;
• Field measurements of the consumptive use of crops.
Field measurements of water use by crops are complex, time consuming and expensive.
Generally irrigation water demands and use are estimated from empirical equations or
calculated from readings from flow measurements located on irrigation schemes. Methods
for estimating agricultural water demand and use are given in Sections 4.3 and 4.4.
4.2 Overview of irrigated agriculture in the SADC region

Sections 4.2.1 to 4.2.14 provide an overview of irrigation in each of the SADC countries.
This overview is based on work carried out by the Food and Agriculture Organization. The
land on which water is used primarily for the purpose of agriculture production has been
called in the text water managed areas. The term irrigated areas has been limited to that part
of the water managed areas equipped with hydraulic structures: full or partial control
irrigation, equipped wetland or valley bottoms and areas equipped for spate irrigation. The
difference between the two categories comprises cultivated wetland and valley bottoms
without irrigation equipment and recession cropping areas (Reference 4.4).
4.2.1 Angola
The potential for irrigation in Angola is immense and may be in excess of several million
hectares. Owing to the civil war in Angola there is a paucity of data on irrigation in Angola.
At independence in 1974 it was estimated that there was between 75,000 ha and 80,000 ha
irrigated using modern techniques. However, in 1991 this area was estimated to be between
10,000 ha and 30,000 ha (Reference 4.4).
4.2.2 Botswana
Total irrigation potential in Botswana has been estimated to be 39,300 ha, however, once
water availability is taken into account this drops to 20,000 ha. This figure includes the area
used for flood recession cropping. The total water managed area is 7,881 ha which is 39% of
the irrigation potential. Full or partial water control is provided on 1,381 ha which is only
17% of the total water managed area. Of this, 612 ha is irrigated from groundwater and
769 ha from surface water, either by pumping from the Limpopo and other rivers or from
storage reservoirs. The major irrigated crops are vegetable, maize, pasture and fodder crops,
citrus and cotton. There are some 6,500 ha of flood recession cropping around the Okavango
and Chobe enclave, on which mostly maize and millet are planted.
Full and partial control irrigation techniques include 218 ha of surface irrigation (e.g.
handwatering, furrow and basin and drag hose) and 1,163 ha under overhead sprinkler (e.g.
centre pivot, sprinkler) and 271 ha under drip and micro-sprinkler. Future irrigation
development is likely to be on a very limited scale the main constraint being shortage of water
and its high opportunity cost (Reference 4.4).
4.2.3 Democratic Republic of the Congo
The irrigation potential in the Democratic Republic of the Congo varies between 4 million
hectares and 20 million hectares, depending on the source of the estimate. Owing to the
political situation in the Democratic Republic of the Congo over the past two decades it is
difficult to obtain reliable figures on the irrigation infrastructure. It is estimated that only
some 8,000 ha is actively irrigated.
4.2.4 Lesotho
In the early 1970s the potential for irrigation development in Lesotho was estimated to be
some 12,500 ha. Since this time no other survey has been undertaken to assess the total
potential in Lesotho. The total water managed area is about 2,722 ha and corresponds to the
total equipped area for full or partial irrigation. This area can be divided as follows:
• 203 ha of small schemes (less than 100 ha). Surface and sprinkler irrigation is mainly
practised in these schemes which are generally donor sponsored. The main crops are
vegetables;
• 2,519 ha of large schemes (more than 100 ha) developed from 1986 onwards. These
schemes were equipped for sprinkler irrigation. However, because the schemes never
managed to make a profit they are no longer irrigated.

Irrigation development has not been very successful in Lesotho and many schemes have been
converted into dryland farming systems. Reasons for the poor performance of irrigation in
Lesotho include:
• Using a top-down approach, whereby farmers were informed that their land had been
chosen for irrigation development and their plots were consolidated into blocks without
prior consultation, resulting in opposition to irrigation development;
• Farmers being expected to provide free labour to the scheme, regardless of the size of the
landholding, while their profit share was based on the size of the holding resulting in poor
labour productivity;
• Low produce prices.
The most successful irrigation projects in Lesotho have been based on an individual approach
to communally owned irrigation schemes where farmers control the on-field crop production
activity (Reference 4.4).
4.2.5 Malawi
The total water managed area is about 89,900 ha which is about 56% of the total potential
area for irrigation estimated at 161,900 ha. At present 28,800 ha is equipped for full or partial
control irrigation. Almost all irrigation is from surface water, either from weirs or by
pumping from rivers. There are some small areas (15 ha to 20 ha) along Lake Malawi that are
irrigated using groundwater. Irrigation techniques include:
• 15,700 ha of surface irrigation (e.g. furrow and basin);

• 9,000 ha of sugar cane under sprinkler irrigation at Sucoma;
• 2,300 ha of tea, coffee and other crops under sprinkler irrigation;
• 1,000 ha under micro-irrigation.
Some 1,100 ha of surface irrigation are in need of rehabilitation. The cropped area in these
full of partial control schemes is 31,500 ha per year. There are three basic categories of
farming in the full and partial control irrigation sub-sector:
• Private estates (18,300 ha). These include Sucoma (9,000 ha) and Dwangwa (6,000 ha)
sugar estates and Kawalazi estate that have been developed as joint ventures between
Government and local and foreign investors;
• Government run small-holder schemes (3,200 ha). These were established by the
government to give irrigation opportunities to local farmers.
• Self-help smallholder schemes (6,500 ha). These have usually been designed and
constructed by the government with the full support and participation of farmers in each
stage of development.
There are also some 61,900 ha of dambo (wetland) areas under rice cultivation (Reference
4.4).
4.2.6 Mauritius
The irrigation sector in Mauritius is primarily focused on sugar cane, which has one of the
highest yields in the world. In 1970 the full or partial control irrigated area was 12,000 ha, all
for sugar cane. In the late 1980s, half of the full or partial control area was irrigated by
sprinkler. The full or partial control irrigated area was estimated to be 16,720 ha in 1987 and
17,500 ha in 1995. The majority of this area was devoted to sugar cane on estates under
sprinkler irrigation. In the past groundwater was used as a major source for irrigation of sugar

cane plantations. This practice has now ceased owing to the high cost of electricity. There is
also a limited area of irrigated land used to produce market garden vegetables.
In recent years there has been an emphasis to move towards water-saving techniques and non-
labour intensive system. Surface irrigation systems that have poor efficiency are being
replaced by drip or sprinkler (Reference 4.4).
4.2.7 Mozambique
The irrigation potential in Mozambique has been estimated to be some 3,300,00 ha of which
61% lies in the Zambezi Valley, which represents 9% of the cultivable area. In 1993 the area
equipped for full or partial irrigation was estimated to be 106,700 ha, only 3.2% of the
irrigation potential. This area is divided into small schemes (less than 30 ha), medium
schemes, (30 ha to 200 ha) and large schemes. The total area actually irrigated is estimated to
cover 45,000 ha i.e. 42% of the equipped area. The main irrigated crops are rice, sugar cane,
maize and citrus fruit. Most irrigation schemes are fed by water abstracted from rivers. Basin
irrigation for rice and furrow irrigation for maize and vegetables are practised. Sprinkler
irrigation is widespread, especially in the sugar and cotton areas.
Irrigation schemes in Mozambique vary from large multi-user schemes (30,000 ha) to small
individual schemes (1 ha). In the large schemes, where commercial and family sector
irrigation co-exist, water management is almost non-existent. Hydraulic structures do not
function, large administrative management bodies have been set up to run the schemes and
there is no physical possibility of water control (Reference 4.4).
4.2.8 Namibia
The total water managed area is 8,142 ha. This figure includes 2,000 ha of recession
agriculture in the floodplains of the Okavango and Zambezi Rivers mainly used for
cultivating maize. In 1992 of the 6,142 ha equipped for irrigation with full or partial control
some 886 ha (14%) were irrigated from groundwater and 5,256 ha (86%) utilised surface
water either by pumping from the Okavango and Orange Rivers or from storage reservoirs.
The major irrigated crops are maize, lucerne and pasture, wheat and cotton. The irrigation
area increased by some 350 ha between 1991 and 1994. Full or partial irrigation techniques
comprise 2,950 ha under surface irrigation (basin and furrow), 1,845 ha under centre pivot
and sprinkler and 1347 ha under drip and micro-sprinkler.
Although some 3.6 million hectares has soil classified as highly suitable for irrigationbut
taking into account water availability the potential for irrigation development is only
45,000 ha. Of this area 8,142 ha is already equipped, 10,000 ha is available for a sugar
project in Caprivi and about 27,000 ha is available for development in the rest of the country
(Reference 4.4).
4.2.9 Seychelles
Irrigation has been developed in recent years in the Seychelles as a supplementary supply to
rain fed crops in cases of water shortages or droughts. Irrigation projects are focused on high
value crops such as orchids (Reference 4.4).
4.2.10 South Africa
The potential for full or partial control irrigation development, based on soil and water
availability and suitability, is estimated at 1.5 million hectares. The total water managed area

is estimated at 1.27 million hectares, consisting only of full or partial control irrigation. Three
irrigation techniques are used:
• Surface irrigation comprising furrow, border and basin irrigation are used for 396,000 ha;
• Sprinkler irrigation is used for 660,000 ha;
• Micro-irrigation is applied on 144,000 ha.
Four types of irrigation scheme can be distinguished in South Africa according to the type of
management. These are:
• Government Water Schemes that in 1994 covered 329,000 ha and are operated by the
Department of Water Affairs and Forestry;
• Irrigation Boards covering 155,000 ha in 1994 that were developed on the basis of
legislation that enables groups of private individuals to obtain statutory powers;
• Private schemes that covered 660,000 ha in 1992 and that are irrigated by private farmers;
• Schemes of the Rural Development Programme, located in the former homelands, that
cover 70,000 ha.
The main irrigated crops are pasture, wheat, lucerne and irrigated sugar cane. It has been
reported that some 110,000 ha of irrigated land is affected is affected by waterlogging or
salinization (Reference 4.4).
4.2.11 Swaziland
The potential for irrigation development based on water and soil availability and suitability is
estimated to be 90,000 ha. This comprises 47% of the cultivated land. The irrigation sector
in Swaziland consists mainly of large private farmers and internationally-owned enterprises.
The total water managed area is estimated at 67,400 ha of full or partial control irrigation
schemes. These schemes a re of two types:
• Micro- and small schemes (1,000 ha in 1980). These communal, smallholder projects are
characterised by individual family holdings of around 0.5 ha. The total size of schemes
averages 20 ha;
• Large schemes (66,400 ha estimated in 1990). Furrow irrigation of sugar cane is the most
practised system, followed by sprinkler and micro-irrigation of pineapple and citrus fruits.
In general water is provided from reservoirs or river diversions (Reference 4.4).
4.2.12 Tanzania
Based on soil and water availability the potential for irrigation development in Tanzania has
been estimated to be 828,000 ha which is approximately 2% of the cultivable area. Exact
figures for the water managed area are not known. It is estimated to be somewhere between
120,000 ha and 200,000 ha. Most of this is in traditional, small holder schemes. Medium to
large schemes cover approximately 20,000 ha to 50,000 ha.
Almost all irrigation schemes are fed by abstraction from rivers. In a few cases storage
reservoirs have been constructed. On some of the larger projects sprinkler irrigation is used.
The main crops in the large scale irrigation projects is rice. Sugarcane occupies about
10,000 ha of the irrigated area. In some limited areas maize is cultivated (Reference 4.4).

4.2.13 Zambia
The irrigation potential based on water and soil resources has been estimated at 1.4 million
hectares. Of this 520,000 ha, including dambo (wetland) and irrigated areas could be
economically developed in the future. The total water managed area is some 146,400 ha.
At present some 46,400 ha are equipped for full or partial control irrigation. A total of
61,900 ha of crops are grown annually in these schemes of which only 46,400 are irrigated.
The majority of the irrigated area is supplied by surface waters i.e. reservoirs or rivers. There
are some 2,500 ha of the irrigated area are supplied by groundwater. Surface irrigation
(furrow and basin) is practised on 28,400 ha, there are 17,200 ha under sprinkler and centre
pivot and 800 ha fed by drip. Irrigation schemes in Zambia fall into three main categories:
• Commercial estates (18,000 ha). These include both large and small scale companies and
commercial farms growing a wide range of high-value crops (e.g. coffee, fruit,
vegetables);
• Parastatal and semi-parastatal (12,400 ha). These include large scale farming
developments with varying degrees of government control;
• Smallholder (16,000 ha). Of this area 3,669 ha has been established by the government as
formal schemes in areas of traditional land tenure. Plots usually occupy 0.5 ha to 1 ha
(Reference 4.4).
4.2.14 Zimbabwe
Taking into account soil suitability and water availability the area of land suitable for
irrigation has been estimated to be 331,000 ha. Between 1968 and 1990 the irrigated area
tripled from 60,000 ha to 191,000 ha, including double cropped areas. Most of the areas that
have been developed for full or partial control irrigation are large scale commercial private
and parastatal schemes. Small scale schemes cover 9,421 ha, while cultivated wetlands and
gardens cover some 20,000 ha. Most water used for irrigation is pumped from regulated
rivers or from farm reservoirs. In 1990 it was estimated that 40% of the nine principal crops
were produced under irrigation. Major irrigated crops included wheat, cotton, sugarcane,
soybean, tobacco and maize.
4.3 Use of empirical equations to calculate irrigation water

demand and use
There are two main methods that can be used to establish irrigation demand and use using
empirical equations. These are:
• Approximate estimates;
• Detailed estimates.
For most water resources management applications at a catchment and sub-catchment level
approximate estimates of irrigation water demand and use should suffice. However, there
may be some cases that warrant more detailed estimates. The approximate and detailed
procedures for estimating irrigation water demand and use are shown in Figures 4.1 and 4.2
respectively.

Calculation of reference
crop evapotranspiration ETo
Estimation of effective
rainfall Pe
Calculation of irrigation demand

IR = ETo - Pe
Estimate overall irrigation

efficiency Ei
Calculate irrigation water

use = IR
Ei
Figure 4.1 Approximate method for estimating irrigation demand and use

Calculation of reference
crop evapotranspiration ETo
Calculation of crop
coefficient Kc
Estimation of effective
rainfall Pe
Estimation of leaching
requirement LR
Estimation of groundwater
contribution Ge
Calculation of irrigation demand

IR = (KcETo - Pe - Ge)
(1 - LR)
Estimation of conveyance,
field and application
efficiencies Ec, Eb, Ea
Calculate overall irrigation

efficiency Ei = EcEbEa
Calculate irrigation water

use = IR
Ei
Figure 4.2 Detailed method for estimating irrigation water demand and
use
It should be noted that often the greatest uncertainty in determining irrigation water use is the
estimation of irrigation efficiency. The overall irrigation efficiency is defined as the ratio of
water consumed by crops to the water diverted from the source (e.g. a river or reservoir). The
overall irrigation efficiency can vary from 10% to 90% and is heavily dependent on the
irrigation technology used, and the operation and maintenance of the irrigation scheme.
Figures for irrigation schemes can be estimated from literature or from similar schemes where
there are measuring devices. Irrigation efficiency is discussed in Section 4.5.7.
The approximate and detailed methods for estimating irrigation water demand and use are
discussed below.
4.4 Approximate method for calculating irrigation water

demand and use
The irrigation water demand and use can be estimated approximately at a catchment and sub-
catchment level using the following equations:
Irrigation water demand IR = (ETo − Pe )A
Where: ETo is the reference crop evapotranspiration for grass;

Pe is the effective rainfall;
A is the area under irrigation.
IR
Irrigation water use =
Ei
Where: IR is the irrigation water demand;

Ei is the overall irrigation efficiency.
Using the potential evapotranspiration for a reference crop (grass) may cause an upward bias
in the estimates of irrigation water demand and use. However, for estimating irrigation water
demand and use at a catchment and sub-catchment level it provides a rapid and simple
method. Values of reference crop evapotranspiration estimated using the Penman-Monteith
method and effective rainfall are available as part of CLIMWAT. CLIMWAT is a climatic
database produced by the Food and Agriculture Organization (FAO) of the United Nations
that provides data for a total of 3262 meteorological stations from 144 countries throughout
the world (References 4.6 and 4.7). An example of an approximate method for estimating
irrigation water demand and use is given below.

Example of an approximate method for estimating irrigation water demand and use
An estimate of the water demand and use is required for a 1000 ha sprinkler irrigation scheme
close to the town of Karoi in Zimbabwe. Crops are grown throughout the year on the scheme. A
similar irrigation scheme in an adjacent sub-catchment is known to have an overall irrigation
efficiency of 70%.
An approximate estimate can be made using the reference crop evapotranspiration and the
effective rainfall for Karoi taken from the FAO CLIMWAT database. The results of the calculation
are shown below.
Month Reference crop Effective Irrigation Irrigation Irrigation

evapotranspiration rainfall requirement water water use
3
(mm) (mm) (mm) demand (million m )
3
(million m )
January 124 143 0 0 0
February 109 133 0 0 0
March 127 86 41 0.41 0.59
April 117 34 83 0.83 1.19
May 115 8 107 1.07 1.53
June 102 3 99 0.99 1.41
July 112 0 112 1.12 1.60
August 143 1 142 1.42 2.03
September 177 4 173 1.73 2.47
October 202 14 188 1.88 2.69
November 159 76 83 0.83 1.19
December 133 129 4 0.04 0.06
Total 10.32 14.74
4.5 Detailed method of estimating irrigation water demand

and use
In order to carry out a detailed estimate of irrigation water demand and use, using empirical
formulae, the following information is required:
• Reference crop evapotranspiration;

• Crop type and crop evapotranspiration;
• Cropped area;
• Effective rainfall;
• Soil type and leaching requirements;
• Irrigation efficiencies.
The FAO has produced a piece of software known as CROPWAT which in conjunction with
CLIMWAT can be used to carry out detailed calculations to estimate irrigation water demand
and use (Reference 4.8). The various components required to carry out a detailed estimation
of irrigation water demand and use are discussed below.
4.5.1 Methods of estimating reference crop evapotranspiration
The reference crop evaporation is based on a hypothetical, well-watered grass reference crop
with specific characteristics. There are four main methods of estimating reference crop
evapotranspiration (ETo). These are:
• Blaney-Criddle method;

• Radiation method;
• Pan evaporation method;
• Penman-Monteith method.
These methods are briefly discussed below.
Blaney-Criddle method
This method is recommended where the only climatic data available is air temperature
(Reference 4.9).
ETo = c[p(0.46T + 8)]
Where: T is the mean daily temperature;

c is the empirical adjustment factor depending on minimum relative humidity,
sunshine hours and daytime wind estimates;
p is the mean daily percentage of annual daytime hours.
Advantages of the Blaney-Criddle method
The main advantages of the Blaney-Criddle method are:
• It is simple to use;
• It can be used where minimal climatic data is available.
Disadvantages of the Blaney-Criddle method
The main disadvantages of the Blaney-Criddle method are:
• Its validity is limited to areas where temperature is a good indicator of general weather
conditions;
• The results can be unreliable for the following conditions:
- Equatorial regions where temperatures remain fairly constant but other weather
parameters change;
- Small islands and coastal areas where air temperature is affected by the sea
temperature having little response to seasonal radiation;
- High altitudes where mean daily temperatures are relatively low but day time
radiation levels are high;
- Climates with a wide variability in sunshine hours during the transition months
(e.g. monsoons);
• It requires extensive local calibration to estimate the value of the empirical factor ‘c’;
• The method should only be used to estimate ETo for periods of at least one month.
Radiation method
The radiation method is recommended where air temperature and sunshine, cloudiness or
radiation is measured, but measurements of wind and humidity are not available (Reference
4.9).
ETo = c( WRs)
Where: Rs is the solar radiation in mm/day;

W is a weighting factor dependent on temperature and altitude;

c is an adjustment coefficient dependent on mean humidity and daytime wind
conditions.
Advantages of the radiation method
The main advantages of the radiation method are:
• The radiation method generally produces more reliable results than the Blaney-Criddle
method;
• It is relatively simple to use;
• For equatorial zones, small islands or high altitudes the method may be more reliable
than the Blaney-Criddle method even if measured sunshine or cloudiness data are not
available.
Disadvantages of the radiation method
The main disadvantages of the radiation method are as follows:
• The adjustment coefficient c needs to be estimated accurately in order to produce

reasonable results;
• The method tends to underestimate evapotranspiration in arid climates;
• The method should only be used to calculate ETo for periods of at least one month.
Pan evaporation method
Evaporation pans provide a method by which the integrated effect of radiation, wind,
temperature and humidity on the evaporation from an open water surface can be measured.
The pan evaporation is related to the reference evapotransipration by an empirically derived
pan coefficient as shown in the equation below.
ETo = KpEpan
Where: Epan is the pan evaporation in mm/day;

Kp is the pan coefficient.
The value of ETo produced by the above equation should be compared with results from other
evaporation pans and results from the Penman-Monteith equation. There are several different
types of pan. The Class A and Colorado pans are two that are most commonly used, these are
shown in Figures 4.3 and 4.4 respectively.
1207 mm
255 mm
150 mm Wooden support

Ground level
Figure 4.3 Diagram of a Class A evaporation pan

Water level 5cm to 7.5 cm Ground level
below the pan rim 5 cm
46 cm
92 cm
Figure 4.4 Diagram of a Colarado evaporation pan
Advantages of pan evaporation methods
The pan evaporation method has the following advantages:
• Under the correct circumstances the pan evaporation method can be used to predict values
of ETo for periods of ten days or longer;
• Evaporation pans are relatively cheap and simple to set up.
Disadvantages of pan evaporation methods
The pan evaporation method has the following disadvantages:
• The accuracy of the results is dependent upon the type and location of the pan;
• The results of the pan evaporation may be affected by birds drinking from the pan;
• The colour and material of the pan affect the evaporation rate. As a consequence the
value of Epan is not constant with time owing to ageing, deterioration and repainting;
• The method is susceptible to microclimatic conditions under which the pans are
operating;
• The reliability of the results is dependent upon the frequency of the pan maintenance;
• The method should only be used to calculate ETo for at least ten days.
Penman-Monteith method
This method is recommended for use wherever possible, providing there is sufficient
meteorological data. The reference crop evapotranspiration is calculated from the following
equation (Reference 4.9).
900
0.408 ∆ (R n - G) + γ U 2 (e a - e d )
ETo = T + 273
∆ + γ (1 + 0.34 U 2)
Where: Rn is the net radiation at crop surface;

G is the soil heat flux;
T is the average temperature;
U2 is the wind speed measured at 2 m height;
(ea-ed) is the vapour pressure deficit;
∆ is the slope vapour pressure curve;
γ is the psychrometric constant.
Reference crop evapotranspiration rates in mm/day calculated using the Penman-Monteith

method for four African capital cities are shown in Figure 4.5.
7
Reference crop evapotranspiration ET o (mm/day)
0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Month
Harare Zimbabwe Luanda Angola
Gaborone Botswana Kinshasa Democratic Republic of the Congo
Figure 4.5 Reference crop evapotranspiration calculated using the

Penman-Monteith method
Advantages of the Penman-Monteith method
The Penman-Monteith method has the following advantages:
• The Food and Agriculture Organization (FAO) of the United Nations recommend the use
of the Penman-Monteith equation as the sole method for the estimation of reference crop
evapotranspiration;
• Climatic data and tools to assist with the calculation of ETo using the Penman-Monteith
method are available from the FAO;
• The method can be used to estimate ETo on a daily basis.
Disadvantages of the Penman-Monteith method
The Penman-Monteith method has the following disadvantages:
• The method requires a considerable amount of climatic data that may not be readily
available. However, mean monthly climatic data is available digitally for approximately
3250 climate stations worldwide from the FAO.

Recommended method to estimate reference crop evapotranspiration
The FAO recommends the Penman-Monteith method as the sole method standard method for
the computation of ETo. The use of the Penman-Monteith method does require a reasonable
quantity of climatic data. However, advice on the infilling of missing data and the setting up
of climate stations is given in FAO Irrigation and Drainage Papers Nos. 56 and 27
respectively. Climate data for over 3000 climate stations in 144 counties is available in a
digital format as part of FAO Irrigation and Drainage Paper No. 49. There are also a number
of software packages that use the Penman-Monteith method equation to assess reference crop
evaporation. The FAO computer program CROPWAT utilises the Penman-Monteith method
and a spreadsheet is also available from the FAO to calculate reference crop evaporation.
4.5.2 Estimation of the crop coefficient
Crop evapotranspiration is the amount of water utilised by a crop under standard conditions
i.e. a disease free, well fertilised crop grown in a large field under optimum soil conditions
(References 4.9 and 4.10). The crop evapotranspiration ETcrop in mm/day is calculated by
multiplying the reference crop evapotranspiration ETo in mm/day by a dimensionless crop
coefficient Kc:
ETcrop = K c ETo
The crop coefficient is dependent upon the following factors:
• Crop type;
• Climate;
• Soil evaporation;
• Crop growth stage.
A typical crop coefficient curve is shown in Figure 4.6 below.

1.2
Kcmid
1.0
Crop coefficient Kc
0.8
Kcend
0.6
0.4 Kcinitial
0.2
Initial Crop
Mid-season Late season
development
0.0
0 20 40 60 80 100 120 140 160 180
Time of season (days)
Figure 4.6 Typical crop coefficient curve
As Figure 4.6 illustrates, the crop coefficient varies over the growing period of the crop. As
the crop develops the ground cover, crop height and the leaf area change. As a consequence
the evapotranspiration of a crop will change during its growing period. The majority of crops
have four growth stages:
• Initial stage when the leaf area is small and the majority of the evapotranspiration is due
to soil evaporation;
• Crop development stage when leaf area grows from approximately 10% to full ground
cover;
• Mid-season stage when the crop is reaching maturity;
• Late season runs from the start of maturity to the harvest.
Typical ranges of crop coefficient Kc together with the factors that affect the Kc during the life
of a crop are shown in Figure 4.7.

1.2
Sugar cane
1.1 Cotton
Maize
Cabbage, onions
1.0
Frequent Apples
Ha
rve
0.9
st e
Crop coefficient Kc
df
res
0.8
h
0.7
Wetting
0.6 events
0.5
0.4
Dr
ied
0.3
Infrequent
0.2 25% 40% 60%
Ground cover
0.1 (short)
Initial Crop Mid-season Late-season
development (long)
Main factors affecting Kc in the four growth stages
Soil Ground cover Crop type

(Humidity) Crop type
evaporation plant Harvesting date
development (Wind speed)
Figure 4.7 Typical ranges expected in Kc for four growth stages
A typical crop coefficient curve can be constructed by establishing the values of Kc for the
initial, middle and end of the curve. Typical values of the crop coefficients are given in the
FAO Irrigation and Drainage Papers Nos. 24 and 56. Crop water requirements can be
calculated using CROPWAT, a piece of software available as part of the FAO Irrigation and
Drainage Papers No. 46.

Example of a crop water requirement calculation
An example of the construction of a crop coefficient curve is shown in Figure 4.8 below. The
calculation has been carried out using ten day periods. The mean crop coefficient for the
middle of each of the ten day periods has been calculated. The crop water requirement ETcrop
in mm/day is calculated by multiplying the crop coefficient by the reference crop
evapotranspiration for each ten day period.
1.2
1.10
1.1
1.0
0.9
0.8
Crop coefficient Kc
0.7
0.6
0.5
0.4 0.35
0.3
0.2 0.15
0.1
0.0
1 2 3 1
2 3 1 2 3 1 2 3
May June July August
Kc 0.15 0.15 0.20 0.56 0.91 1.10 1.10 1.10 0.90 0.52
ETo (mm/day) 3.1 3.5 3.7 4.0 4.2 4.1 4.0 3.9 3.8 3.6
ETc (mm/day) 0.5 0.5 0.7 2.2 3.8 4.5 4.4 4.3 3.4 1.9
Figure 4.8 Crop coefficient curve together with crop water requirements
4.5.3 Estimation of effective rainfall
Crop water requirements may be fully or partly met by precipitation. However, it should be
noted that not all rainfall is effective as some may be lost to surface runoff, deep percolation
or evaporation. The effective rainfall may be defined as the proportion of the rainfall that is
useful for crop production. There are several factors that affect effective rainfall including:
• Rainfall characteristics;
• Soil type;
• Groundwater characteristics;
• Management practices.
There are several empirically based methods available to calculate effective rainfall including:
• Renfro equation;
• US Bureau of Reclamation method;
• US Department of Agriculture Soil Conservation Service method.
The US Department of Agriculture Soil Conservation Service method is often used. Table
4.1 below allows the mean monthly effective rainfall to be estimate provided the crop
evapotranspiration and mean monthly rainfall are known (Reference 4.10).
Table 4.1 Mean monthly effective rainfall related to mean monthly ETcrop and mean
monthly rainfall
Mean monthly 12.5 25 37.5 50 62.5 75 87.5 100 112.5 125 137.5 150
rainfall (mm)
Mean monthly effective rainfall (mm)

Mean monthly ETcrop (mm)
25 8 16 24
50 8 17 25 32 39 46
75 9 18 27 34 41 48 56 62 69
100 9 19 28 35 43 52 59 66 73 80 87 94
125 10 20 30 37 46 54 62 70 76 85 92 98
150 10 21 31 39 49 57 66 74 81 89 97 104
175 11 23 32 42 52 61 69 78 86 95 103 111
200 11 24 33 44 54 64 73 82 91 100 109 117
225 12 25 35 47 57 68 78 87 96 106 115 124
250 13 25 38 50 61 72 84 92 102 112 121 132
An example of an effective rainfall estimation is given below.
Example of an effective rainfall requirement estimation
A tomato crop has a mean monthly evapotranspiration of ETcrop = 137 mm/month

The mean monthly rainfall is 95 mm/month
Interpolating using the US Department of Agriculture Soil Conservation Service method

outlined in Table 4.1 the mean monthly effective rainfall is approximately 69 mm.
4.5.4 Groundwater contribution
The contribution from groundwater to the crop water requirement is a factor of the following:
• Depth of the groundwater below the root zone;

• Capillary properties of the soil;
• Soil capillary capacities.
For heavy soils the distance of the movement of the groundwater is high and the rate low, and
for coarse soils the distance of movement of the groundwater is small and the rate is high. It
should be noted that detailed field experiments are usually needed to ascertain the
groundwater contribution to the crop water requirement. For most catchment planning
purposes it is unlikely that it is necessary to calculate groundwater contributions.
4.5.5 Stored soil water
Stored soil water at the start of the growing season resulting from rain can mean that the soil
is close to or at its field capacity. This stored soil water at the start of the growing season can
be deducted from the crop water requirement. However, for the purposes of catchment
planning it is usually not necessary to take account of stored soil water as it will vary
significantly on a seasonal basis.
4.5.6 Estimation of leaching requirements
The leaching requirement is the minimum amount of irrigation water that must be drained
through the root zone to control soil salinity (Reference 4.10). For sandy loam to clay loam
soils with good drainage, in locations where the rainfall is low, the leaching requirement can
be estimated from the following:
For surface irrigation methods including sprinklers the leaching requirement is:
ECw 1
LR = x
5ECe − ECw Le
ECw 1
For drip and high frequency sprinklers LR = x
2MaxECe Le
Where: ECw is the electrical conductivity of the irrigation water (mmhos/cm);

ECe is the electrical conductivity of the soil saturation extract for a given crop
appropriate to the tolerable yield reduction;
MaxECe is the maximum tolerable electrical conductivity of the soil extract for a
given crop;
Le is the leaching efficiency.
Values of ECe, ECw and MaxECe are tabulated for different crops for a variety of yield
potentials in FAO Irrigation and Drainage papers numbers 24 (Reference 4.10) and 29
(Reference 4.23). The leaching efficiency (Le) varies with soil type and can be as low as 30%
(0.3) for cracking and swelling clays and as high as 100% (1.0) for sandy soils. The water
needed to satisfy both the crop water requirements (ETcrop) and the leaching requirements is:
ETcrop − Pe
Water to meet crop and leaching requirements =
(1 − LR )
Where: Pe is the effective rainfall;

LR is the leaching requirement.

Example of a leaching requirement estimation
The seasonal crop water requirement ETcrop for sugarbeet is 850 mm. The effective rainfall
during the growing period is 150 mm. A water quality analysis has shown the water to have
an electrical conductivity of 6.8 mmohs/cm. The soil is a sandy-loam with a leaching
efficiency (Le) of 90% (0.9). For sugarbeet the ECe is 7.0.
ECw 1 6. 8 1
Leaching requirement LR = x = x = 0.27
5ECe − ECw Le 5 x7 − 6.8 0.9
To meet the crop water requirement and leaching requirement to obtain a 100% yield the
following quantity of water is required:
ETcrop − Pe 850 − 150
Water to meet crop and leaching requirements = = = 960 mm
(1 − LR ) (1 − 0.27 )
4.5.7 Irrigation efficiencies
The net irrigation requirements can be calculated from the crop water requirements, effective
rainfall, the groundwater contributions and the stored soil water. However, in order to
calculate the total irrigation requirements the quantity of water required for leaching together
with the irrigation efficiency should be taken into account. The irrigation efficiency is
defined as the ratio of water consumed by crops to the water diverted from the source (e.g. a
river or reservoir). The irrigation efficiency can be calculated from the equation:
Vc
Overall irrigation efficiency E i =
Vw
Where: Ei is the irrigation efficiency;

Vc is the water consumed by the crops;
Vw is the water diverted from the source (e.g. a river or reservoir).
The overall irrigation efficiency (Ei) is made up of the conveyance efficiency (Ec), the field
canal efficiency (Eb) and the application efficiency (Ea). These are defined as follows.
The conveyance efficiency is defined as follows:
Vfarm
Conveyance efficiency E c =
VW
Where: Ec is the application efficiency;

Vfarm is the quantity of water delivered to the farm or a block of fields;
Vw is the quantity of water diverted from the source (e.g. a river or a reservoir).
The conveyance efficiency is dependent upon on the method of delivery (e.g. piped, lined or
unlined canal), the size of the irrigation scheme and the rotation of the water supply. The
value of the conveyance efficiency can vary from 0.6 to 0.9 (Reference 4.9).
The field canal efficiency (Eb) is defined as follows:
Vfield
Field canal efficiency E b =
Vfarm

Where: Eb is the application efficiency;
Vfarm is the quantity of water delivered to the farm or a block of fields;
Vfield is the quantity of water delivered to the field.
The field canal efficiency is dependent on the method of delivery and the size of the field.
The value of the field canal efficiency can vary from 0.6 to 0.9 (Reference 4.9).
The application efficiency (Ea) is defined as the ratio of the water made available directly to
the crop and the water received at the inlet to the field. The application efficiency is defined
as follows:
Vcrop
Application efficiency E a =
Vfield
Where: Ea is the application efficiency;

Vfield is the quantity of water delivered to the field;
Vcrop is the quantity of water available to the crop.
Often considerably more water is applied to the soil than it can possibly hold. Irrigation
application efficiencies can vary from extremely low values to values approaching 1.0
(100%). In normal irrigation practices, surface irrigation efficiencies are usually around 0.6
(60%), whereas well-designed sprinkler irrigation systems are generally considered to be
approximately 0.75 (75%). An example of an irrigation efficiency calculation is given below.
Example of irrigation efficiency calculation
A 2000 ha irrigation scheme is divided into 50 ha blocks fed by unlined canals and utilising
furrow irrigation. The scheme has the following irrigation efficiencies:
Conveyance efficiency Ec = 0.80

Field canal efficiency Eb = 0.75
Application efficiency Ea = 0.70
Overall irrigation efficiency Ei = EcEbEa = 0.80 x 0.75 x 0.70 = 0.42 i.e. 42%
Typical irrigation efficiencies for the SADC region
Irrigation water demand is expected to more than double by 2020, but its share of total water
use in the SADC region is expected to decrease from 70% to 63% as urban demand is expected
to outpace all other sectors. The agricultural water sector holds even greater potential for
savings than the urban sector because it uses three times as much water and is even more
inefficient than urban use. Only 40% of water abstracted from water surface and groundwater
sources are believed to reach crop root system. Table 4.2 gives indicative values for overall
irrigation efficiencies.

Table 4.2 Typical overall irrigation efficiencies
Irrigation system Typical Percentage of Percentage of irrigated

overall irrigated area in area in Zimbabwe in 1990
efficiency South Africa in (%)
(%) 1990(%)
Surface systems (flood) 55 38 25

Conventional sprinkler 75 22 75
Mechanical (centre pivot) 80 24 -
Micro jet 85 16 -
Drip 90 - -
Table 4.2 shows, the least efficient irrigation method is flood irrigation at 55%, while drip
irrigation achieves 90%. In both South Africa and Zimbabwe (the only countries for which
data were available for this study), a significant percentage of land is irrigated with flood
methods, 38% and 25% respectively. The remainder of Zimbabwe's irrigated land is watered
by conventional sprinkler method, while South African agriculture in 1990 had a relatively
even distribution of the other methods except drip. Both these methods are inherently
inefficient. It is therefore not surprising to see water usage figures of 15,000 m3/ha in
Zimbabwe as opposed to 5,000 m3/ha for the same crop in a more water efficient Israel
(Reference 4.16). Clearly, significant savings in water could be realised by retrofitting
inefficient flood and conventional sprinkler irrigation systems with more efficient mechanical,
micro jet and drip irrigation methods.
Figure 4.9 shows the irrigation water supply process and the inflows and outflows to a typical
irrigation scheme (Reference 4.16). The black part of the figure indicates the quantity of
water diverted from the source of water, in this case a river. The width of the arrows
downstream of the river diversion illustrates the relative magnitude of water quantities in an
“average” irrigation scheme in southern Africa. Figure 4.9 also provides an indication of the
relative irrigation efficiencies. This figure also illustrates that water losses resulting from
evaporation are relatively small compared with operational losses to groundwater and surface
water. These operational losses eventually return to the river. As a result of this the river
discharge downstream of the project is higher than would be expected just by looking at the
flow that is downstream of the diversion.
The downstream river discharge can be re-used by a downstream irrigation system. Hence
the efficiency of irrigation water use at a catchment level can be considerably higher than the
overall efficiency of a single project in the catchment. When managing water at a catchment
level it has to be appreciated that a significant quantity of water is re-cycled between the
sources and sinks. As a consequence water supply can be conceptualised in terms of two
distinct components. The primary water supply from precipitation, inter-catchment transfers
and desalination. The secondary water supply that derives from re-cycling the primary water
supply (Reference 4.16). However, it should be recognised that return flows or re-cycled
water can be quite saline and polluted by pesticides and fertilisers. Surface and sub-surface
flows of usable water may also drain to salt sinks such as oceans, inland seas or saline
aquifers. This will also prevent the re-use of the water. This concept is more fully
investigated in Seckler’s paper entitled “The new era of water resources management: From
“dry” to “wet” water savings” (Reference 4.21).

River upstream of project - flow into reservoir
Storage reservoir
Reservoir seepage Evaporation

Flow released from reservoir
Surface water or groundwater inflow

River reach used to convey irrigation water Evaporation
Seepage from river
Undiverted
river flow
River diversion for irrigation
Surface runoff Operational Conveyance system

losses to Evaporation
from non- mains and laterals
groundwater Conveyance system
irrigated land mains and laterals
Surface water
Industrial and Pumped from
sewage inflow groundwater Off take to tertiary
distribution system
Natural
inflow
Operational
losses to Distribution system
Evaporation
groundwater tertiary units
Field water supply

Surface
through field inlets
water
Bypass water
Deep Precipitation
percolation Fields
plus leaching
Evapotranspiration
Tailwater from crops
Drainage from
River downstream groundwater Other evapotranspiration from irrigated land
of project
Figure 4.9 The relative magnitude of quantities of water flowing through

an “average” irrigation scheme

4.5.8 Detailed estimation of irrigation water demand and use
The irrigation water demands for an irrigation scheme with n different crops being grown on
it can be calculated as follows:
n
é A( ETcrop − Pe − G e ) ù
Irrigation water demand = å êëê
i =1
1 − LR
ú
ûú
The irrigation water use is estimated by taking the irrigation efficiency into account as
follows:
1
n
é A( ETcrop − Pe − G e ) ù
Irrigation water use =
Ei å êêë
i =1
1 − LR
ú
úû
Where: Ei is the overall irrigation efficiency.
It should be noted that the groundwater table and stored soil water at the beginning of a
growing season can also make a contribution to the agricultural water demand and use.
However, these contributions are not easily measured and for the purposes of assessing
agricultural water demand and use on a catchment basis it is often usual to ignore them.
In many parts of the world the growing season is long enough that double and some times
triple cropping can take place on the same irrigated area. In these cases the irrigated area
should reflect the number of crops grown in a year. An example of estimating irrigation
water demand and use is given below.
Example of the estimation of irrigation water demand and use
The annual agricultural water demand and use for a 1500 ha irrigation scheme in southern
Africa can be calculated as follows. The following crops together with their associated areas,
crop water requirements groundwater contribution and leaching fraction are given below. The
overall irrigation efficiency of the scheme has been estimated to be 0.55 (55%). The mean
annual rainfall for the scheme is 600 mm. A water quality analysis has been carried out that
indicates that from the electrical conductivity of the irrigation water ECw is 7 mmhos/cm. The
soil is sandy with a leaching efficiency (Le) of 100%.
Crop type Crop area Crop water Effective Groundwater Leaching
requirement rainfall Pe contribution requirement
(ha) ETcrop (mm/year) Ge (mm/year) fraction LR
(mm/year)
Maize 500 1140 420 50 0.21

Cotton 250 1210 420 - 0.22
Soybeans 750 870 408 - 0.38
é æ 1140 − 420 − 50 ö æ 1210 − 420 ö æ 870 − 408 öù
Irrigation water demand = 10 ê500ç ÷ + 250ç ÷ + 750ç ÷ú
ë è 1 − 0 .21 ø è 1 − 0 . 22 ø è 1 − 0.38 øû
3
= 12.4 million m per annum
3
Irrigation water use = Agricultural water demand = 12.4 = 22.5 million m
per annum Irrigation efficiency Ei 0.55

4.6 Measurement of irrigation water use
Measurements of irrigation water use focus on determining the following:
• Withdrawals from surface water resources (e.g. watercourses and reservoirs) and
groundwater abstractions;
• Deliveries from surface water delivery systems;
• Conveyance losses owing to evaporation and seepage;
• Return flows from groundwater and surface water.
Consumptive use during application and use of water by the plants are usually estimated. It is
important to determine which water use processes are critical to the objectives of the study,
particularly where surface water delivery systems are used. Withdrawals may be considered
to have occurred when the water leaves a reservoir, with delivery occurring when water is
diverted from the canal into the irrigation field. Conveyance losses (evaporation and canal
seepage), as opposed to the volume of water applied to the field, can be a major component of
the withdrawals, especially in arid areas and areas with low water tables and porous surface
material. In other areas, surface water withdrawals may be considered to occur as the water is
diverted from the canals adjacent to the fields so that conveyance losses are negligible.
Similarly, withdrawals can be from rivers, streams, or wells adjacent to the irrigation field
with negligible conveyance loss. To avoid confusion in this section, diversions from natural
surface-water bodies and aquifers are considered withdrawals and diversions from canals are
considered deliveries to the field.
Surface water irrigation systems rely on water diverted or pumped from a river, stream, lake,
or reservoir. Water diverted or pumped from a surface or groundwater sources can be
measured using a variety of techniques. For example this can be done by measuring flow in
the diversion (the point where water is withdrawn from the stream) or by measuring the flow
upstream and downstream from the diversion when the diversion is a significant part of the
flow. Similarly, the return flow can be determined by measuring flow at the point of
discharge into the stream or measuring the flow upstream and downstream of the discharge.
The difference between withdrawals and return flow is consumptive use, which consists of
evaporation, deep percolation, and evapotranspiration by and incorporation into the plant.
However, to measure consumptive use accurately using this method requires large quantities
of data to be collected together and a calibration process to be undertaken.
4.6.1 Measurement of conveyance losses
Conveyance losses (including evaporation and seepage) can be measured after the return flow
of one user and before the withdrawal of the next user. Conveyance losses can also be
measured by determining the loss attributable to canal seepage and adding an estimate of
evaporation. Several methods commonly used to measure canal seepage include:
• Ponding tests;
• Inflow-outflow studies;
• Seepage-meter studies.
4.6.2 Ponding tests
Ponding tests give the most reliable results. To conduct a ponding test, a section of canal is
blocked off with dams at each end and filled with water to, or slightly higher than, the level at
which it usually flows during the irrigation season. As the water level in the canal section
declines, the time is recorded and a seepage rate determined. Corrections should be made for
temperature and evaporation and the seepage loss-rate computed. Ponding tests are usually
conducted during the non-irrigation season, and are applied in a non-flow situation in which
actual flow conditions are not being met. This technique is discussed in more detail in
Reference 4.18.
4.6.3 Inflow-outflow studies
Inflow-outflow studies are conducted using long reaches of the canal and require the least
extrapolation of the three methods. However, the inaccuracy of an inflow-outflow
measurement is proportionate to the total flow in the canal, and can be a much larger value
than the amount of seepage that occurs in a reach that has little seepage. Inflow-outflow
studies using discharge measurements are described in detail by Rantz et al in 1982
(Reference 4.17). One of the major advantages of using the inflow-outflow method is that it
can be applied during the irrigation season.
4.7 Irrigation water use productivity

Over the past 30 years the use of conventional irrigation has grown significantly both
throughout the world and in the SADC region. Irrigated agriculture has been an extremely
important source of food in the SADC region in recent decades. Figure 4.10 shows that the
yields that can be obtained from irrigated agriculture can be more than double the highest
yields obtained from rainfed agriculture. Even low-input irrigation is more productive than
high-input rainfed agriculture.
8000 Irrigated crops

high yielding varieties
high inputs
7000
Cereal production (kg/ha)
6000
5000 Irrigated crops

low inputs
4000
Rainfed crops
3000 optimal inputs
2000 Rainfed crops

low inputs
1000
0
1000 2000 3000 4000 5000 6000 7000
Water used (m3/ha)
Figure 4.10 Yields and water requirements of irrigated and rain fed
agriculture
However, conventional irrigation cannot continue to grow as quickly as it has over the past 30
years. This is because the environmental costs of conventional irrigation schemes are also
high (and are often not reflected in food prices). High-intensity irrigation can lead to
waterlogging and/or salinization. About 30% of irrigated land is now severely or moderately

affected by salinization. The salinization of irrigated areas is reducing the existing area under
irrigation by 1% to 2% a year. In many countries the real cost of irrigated food production is
far from clear since, irrigation is often very heavily subsidised. In spite of these reservations
it is likely that not only will irrigation continue to be used but the area under irrigation will
also expand. However, what needs to be improved is irrigation water use efficiency or
productivity i.e. increasing the crop yield per drop of water applied.
There are basically five types of irrigation:
• Surface irrigation, in which the entire or most of the crop area is flooded;
• Sprinkler irrigation, which imitates rainfall;
• Drip irrigation, in which water is dripped onto the soil above the root zone only;
• Underground irrigation of the root zone by means of porous pots or pipes placed in the
soil;
• Sub-irrigation, in which the groundwater level is raised sufficiently to dampen the root
zone.
The first two of these, surface and sprinkler irrigation, are together known as conventional
irrigation. Surface irrigation is currently by far the most common technique, and is used
particularly by small farmers since it does not involve operation and maintenance of
sophisticated hydraulic equipment. For the same reason, surface irrigation is still likely to be
dominant in 2030, even though it is wasteful of water and is a major cause of waterlogging
and salinization (Reference 4.27).
Drip irrigation and underground irrigation are examples of localised irrigation, an

increasingly popular form of irrigation in which water efficiency is maximised because water
is applied only to the places where it is needed and little is wasted. However, technology is
not the only way in which irrigation water productivity can be increased. Small-scale
irrigation and the use of urban wastewater promise to increase water productivity as much as
changes in irrigation technology.
Drip irrigation has been applied only on a small part of the area for which it is suited. It
depends on a pressurised system to force water through perforated pipes running above
ground, at rates of 1 litre to 10 litres per hour per emitter. Although the technology is simple,
it does require both investment and careful maintenance because emitters can easily become
clogged. However, results from many countries show that farmers who switch from furrow
systems or sprinkler irrigation to drip systems can cut their water use by 30% to 60%
(Reference 4.27). Crop yields often increase at the same time because plants are effectively
‘spoon-fed’ the optimal amount of water (and often fertiliser) when they need it. Drip
irrigation has been used in sub-Saharan Africa to significantly increase water use productivity
(i.e. the amount of crop per drop) on large irrigation schemes. A case study based in
Swaziland is described in Section 4.10.
To produce an optimal yield crops must have adequate water at the right times to increase the
productivity of agricultural water, that is, produce more crop per drop of water used. The
maximum yield of crops is related to the actual yield by the equation below.
Ya = 1 – ky(1 – ETa)
Ym ETm
Where: Ya is the actual harvested yield

Ym is the maximum harvested yield
ky is the yield response factor for each stage of the crop development
ETa is the actual evapotranspiration which is related to the quantity of water
the crop receives

ETm is the maximum evapotransiration
Further information on the above equation is given in the FAO’s Irrigation and drainage paper
no. 33 entitled “Yield response to water”. Many crops’ yields are most susceptible to water
shortages during the flowering stage of their development (Reference 4.30). Water shortages
that occur in the ripening stage of their development often have significantly less effect on
yields. For example water stress at the flowering stage of maize, for example, will reduce
yields by 60%, even if the water is adequate for the rest of the year. When and how water is
applied to a crop thus has a major effect on the water efficiency of the water use in terms of
the yield per unit of water applied. An example of this is given below.
Example water use efficiency for wheat in Zimbabwe
A large agricultural estate in southern Zimbabwe cultivates wheat during the winter season.
The wheat crop entirely depends on sprinkler irrigation, as it rarely rains during winter. During
five consecutive growing seasons (1995 to 1999), the total amount of water entering five
wheat plots was measured, as well as the yields obtained. These data were plotted in a
graph Figure 4.11. The research also found that the field application efficiency of sprinkler
irrigation on the estate was 80%.
5000
4500
4000
3500
3000
Yield (kg/ha)
2500
2000
1500
1000
500
0
2000 3000 4000 5000 6000 7000 8000 9000 10000 11000 12000
3
Water use (m /ha)
Figure 4.11 Water use versus yield relationship for irrigated wheat in southern
Zimbabwe for 1995 to 1999

From Figure 4.11 it can be seen that the maximum yield of wheat (Ym) is around 4500 kg/ha.
3
In order to obtain this yield the irrigation requirement is around 8000 m of water has to be
applied per hectare of the crop. For the maximum yield of the crop the water use efficiency is:
4500 3
Water use efficiency = = 0.56 kg/m
8000
However, if the scenario is considered where there is insufficient water to obtain the
3
maximum yield. To obtain a yield (Ya) of 3250 kg/ha an irrigation requirement of 5400 m of
water has to be applied per hectare of the crop. The water use efficiency is:
3250 3
Water use efficiency = = 0.60 kg/m
5400
Although the maximum yield of the crop has not been obtained the water use efficiency is
10% higher than that which occurs at the maximum yield.
An example of the improved water use efficiency that can be gained using drip irrigation in
the Cape Verde islands is given below.
Example of the improved water use efficiency using drip irrigation in Cape Verde
In the early 1990s an FAO funded project sought to develop horticulture in Cape Verde. The
project was a success but its extension was limited by the availability of water - average
precipitation on the islands is about 230 mm/year, providing little more than 700
3
m /person/year. Drip irrigation was then introduced, first in experimental plots and then in
farmers' fields. The new system increased production and saved water, allowing for an
expansion of the irrigated land and cropping intensity. Convinced by the experiment, many
farmers spontaneously adopted drip irrigation on their land. In 1999, six years after the first
experiment, 22% of the irrigated area of the country had been converted to drip irrigation, and
many farmers had converted their crops from water-consuming sugar cane plantations to
high-return horticultural crops such as potatoes, onions, peppers and tomatoes. Total
horticultural production increased from 5700 tonnes in 1991 to 17,000 tonnes in 1999. It is
estimated that a plot of 0.2 hectares provides farmers with a monthly revenue of US$1000.
The relationship between water use and yield can be used to assist in assessing the marginal
value of supplementary irrigation. The example below from Zimbabwe illustrates how this
value can be estimated.
Example to illustrate the value of water for maize in Zimbabwe
For selected plots in Nyanyadzi irrigation scheme in Zimbabwe research was undertaken that
3
found that one additional m of water (irrigation water plus rainfall) supplied to the maize crop
-3
(rainfed with supplementary irrigation) gave an added yield of 1.5 kg of maize per m of water
2
(Correlation coefficient r = 0.81). Assuming a maize price of US$0.10 per kg, it follows that
3
the marginal value of water (rainfall plus irrigation) is US$0.15 per m .
Yields were also correlated with net total irrigation water (Inet in mm). The following
mathematical relationship was found:
2
Y = 1450 + 18.67Inet (Correlation coefficient r = 0.71)
The constant of 1,450 kg/ha indicates the yields obtainable for a rainfed crop without

irrigation. The marginal productivity of net summer supplementary irrigation water was
/ 3 3
around 19 kg/ha mm, or 1.9 kg/m . This means that 1 m of supplementary irrigation water
will produce an additional 1.9 kg of maize, which is valued at US$ 0.19. The marginal value
3
of supplementary irrigation for maize in Nyanyadzi is therefore 0.19 US$ per m . Figures 4.12
and 4.13 illustrate the discussed relationships.
7000
6000
5000
Yield Y (kg/ha)
4000
3000
Y =-4500 +15.39TNWU
2
r = 0.81
2000
1000
0
400 450 500 550 600 650 700
Total water use TNWU (mm)
Figure 4.12 Relationship between net water use and yield for maize for
Nyanyadzi in Zimbabwe
6000
5000
4000
Yield Y (kg/ha)
3000
Y = 1444 + 18.67Inet
2
r = 0.71
2000
1000
0
0 50 100 150 200 250
Net irrigation Inet (mm)
Figure 4.13 Relationship between net irrigation water and yield for maize for
Nyanyadzi in Zimbabwe

4.8 Improving rain fed agricultural production
Increasing the productivity of rainfed agriculture, which still supplies some 60% of the
world's food, would make a significant impact on global food production. However, the
potential to improve yields depends strongly on rainfall patterns. In arid areas, rainwater
harvesting can both reduce risk and increase yields (Reference 4.27). Figure 4.14 shows,
there are various forms of rainwater harvesting. These include:
• Using microstructures in the field to direct water at specific plants or plant rows (in situ
water conservation);
• Capturing and directing external water from the catchment area to the field in which crops
are grown (flood irrigation);
• Collecting external water from the catchment area and storing it in reservoirs, ponds and
other structures for use during dry periods (storage for supplementary irrigation).
Characteristics of arid zones

Drought
Erratic rains
Short rainy season
High evaporation
High runoff Periodic crop Dry spells at

water deficits critical times
In situ Flood Storage for

water irrigation supplementary
conservation irrigation
Eyebrow terraces, Spate irrigation,

contour strips, stream flow
furrows, diversion, Subsurface
contour bunds, groundwater dams, surface
trash lines, recharge, runoff dams, spring
terraces, pits, spreading development,
ridges, systems, tanks
conservation, road runoff
tillage
Low risk Medium risk High risk

reduction reduction reduction
Low investment Low investment High investment
Figure 4.14 Methods of improving rainfed agricultural production

4.8.1 In-situ water conservation
In dry areas, poor land management can greatly reduce crop yields, even to below 1 tonne per
hectare. One reason is that land degradation often affects the soil surface, leading to crust
formation and other phenomena that prevent infiltration by rainwater. Most rainfall then
simply runs off the land surface, collects in silt-laden torrents and produces severe gully
erosion. Crops benefit little.
A major cause is turning the soil, by hand, with animal traction or with a tractor, too often.
This leaves the soil exposed and prone to both wind and water erosion. While ploughing
techniques developed in temperate regions, with their gentle rains and light winds, are
harmless enough, they are often poorly suited to tropical climates and soils (Reference 4.27).
Alternative forms of tillage - such as turning the soil only along plant lines, deep ploughing to
break up soil crusts, building raised ridges that follow the contour, growing crops in pits, and
building eyebrow terraces round trees and shrubs - can improve crop yields and reduce
erosion. They lead to a much more efficient use of limited rainfall. Trials in Tanzania have
shown, for example, that breaking up the plough-pan increased maize yields from 1.8 tonnes
to 4.8 tonnes per hectare in a year with good rains and if manure was applied as fertiliser.
The effects of these trials are shown in Figure 4.15. In Damergou in the Niger, 310 hectares
were equipped with micro-catchments and contour furrows in less than one month using
special ploughs. Costs were only US$90 per hectare. Average yields were 2 tonnes per
hectare of sorghum with an annual rainfall of only 360 mm (Reference 4.27).
4
Yield (tonnes/hectare)
3
sub-soil
control
2
0
1
with fertiliser without2 fertiliser with 3fertiliser 4
without fertiliser
drought good rains
Figure 4.15 Effects of sub-soiling in Tanzania

4.8.2 Supplementary irrigation
Crop yields from rain fed agriculture can be significantly improved via the use of
supplementary irrigation. Runoff during spells of rain can be stored using tanks, ponds,
cisterns and earth dams. During dry spells this water can be used for supplementary
irrigation. Although structures such as tanks and ponds can be costly and require
considerable know-how on the part of the farmers who have to build them, they have the
advantage of greatly reducing the risk of small or non-existent harvests as a result of drought.
Small-scale farming can be productive in marginal rainfed areas if supplementary irrigation is
available to overcome short-term droughts which are critical to the crop and reduce yield
considerably. If there are cost-effective ways to store water before critical crop stages and
apply it when the rain fails in these critical stages, crop production can be considerably
increased (Reference 4.27). This results in considerable improvements in water use
efficiency.
4.9 Deficit irrigation

In the 2000 the FAO produced a report entitled “Deficit Irrigation Practices”. This was
published as Water Report No. 22. The studies presented in the report are the latest research
concepts and involve various practices for deficit irrigation. Both annual and perennial crops
were exposed to different levels of water stress, either during a particular growth phase,
throughout the whole growing season or in a combination of growth stages. The overall
finding, based on the synthesis of the different contributions, is that deficit or regulated-deficit
irrigation can be beneficial where appropriately applied. Substantial savings of water can be
achieved with little impact on the quality and quantity of the harvested yield. However, to be
successful, an intimate knowledge of crop behaviour is required, as crop response to water
stress varies considerably (Reference 4.31).
A study carried out on winter wheat in the North China Plain between 1992 and 2000 showed
possible water savings of 25% to 75% by applying deficit irrigation at various growth stages,
without significant loss of yield and profits. A dynamic model was used to calculate the net
profits of the irrigation treatments. Procedures were developed to schedule irrigation
applications according to the number of irrigations required. In deficit studies carried out in
India on irrigated groundnuts, it was possible to increase field water use efficiency and dry
matter by imposing transient soil moisture-deficit stress during the vegetative phase, i.e. 20 to
45 days after sowing. Water stress applied during vegetative growth may have had a
favourable effect on root growth, contributing to more effective water use from deeper layers.
While most studies were able to demonstrate the benefits of deficit irrigation, potatoes grown
under sprinkler irrigation in the semi-arid environment of eastern Oregon, United States of
America, did not show an economic benefit when exposed to stress. Growing four varieties
of potatoes under various deficit irrigation treatments resulted in gross revenues declining by
more than the production costs, and hence reduced profits. The results of this case study
suggest that deficit irrigation of potatoes would not be a viable management option for that
region under current economic conditions (Reference 4.31).
Fruit crops such as peach and pear trees and grapevines reacted favourably to deficit irrigation
practices, with important water savings and improved fruit quality. In south-eastern
Australia, regulated deficit irrigation (RDI) of peach and pear trees increased water use
efficiency by 60%, with no loss in yield or reduction in vegetative vigour. In Washington
State, United States of America, regulated deficit irrigation of grapevines prior to fruit set
(veraison) was effective in controlling shoot growth and pruning weights, with no significant
reduction in yield. RDI applied after veraison to vines with large canopies resulted in greater
water deficit stress. Wine quality improved with pre-veraison RDI applied as compared to

post-veraison RDI. RDI applied at anytime resulted in better early-season lignification of
canes and cold hardening of buds (Reference 4.31).
In addition to RDI, partial root zone drying (PRD) is also a promising practice for inducing
stress tolerance in fruit trees. PRD is a new irrigation technique that subjects one-half of the
root system to a dry or drying phase while the other half is irrigated. The wetted and dried
sides of the root system alternate on a 10 to 14 day cycle. Both RDI and PRD systems require
high management skills. Close monitoring of soil water content is recommended. Both
practices improve the water use efficiency of wine grape production. Micro-irrigation
facilitates the application of RDI and PRD. Practical guidelines for using RDI were developed
(Reference 4.31).
Subsurface drip irrigation (SDI) also improved the water use efficiency of crops and reduced
farming costs. An approach was developed for deficit SDI on cotton grown in arid east
Texas, United States of America, to enable farmers with a limited supply of water to decide
on the optimal area to plant and the best row width/pattern to apply. By applying deficit SDI,
it proved more economical to use the available water resources over the entire farm, rather
than to try to maximise water and yield on part of the farm. Moreover, with SDI, it proved
possible to apply a large part of the water required as pre-planting irrigation, thus effectively
advancing the timing of water application to the beginning of the season when more water is
available (Reference 4.31).
4.10 Case study of improved water use efficiency for an

irrigation scheme in Swaziland
Irrigation is vital to the production of sugarcane production at the Simunye sugar estate in
Swaziland. In an average year the sugarcane crop receives 440 mm of effective rainfall and
requires an additional 750 mm of net irrigation for optimum growth conditions to occur. The
drought that occurred in southern Africa in the early 1990s led the managers of the estate to
investigate methods of improving water use efficiency and sugarcane yield per unit of water
in order to release water to grow additional areas of sugarcane (Reference 4.21).
When the full commercial production of sugarcane commenced in 1982 the estate had two
main irrigation systems. Overhead sprinklers were used on 77% of the land and surface
furrow was used to irrigate the remainder. Although the sprinkler system initially installed
worked well and was simple to operate by the mid-1990s it was nearing the end of its
serviceable life. Figure 4.16 shows the growth in sugarcane area and the use of drip irrigation
between 1982 and 2001 (Reference 4.21).
There were several reasons for changing the irrigation system. These included:
• Improving water use efficiency;

• Making water available to increase the area of sugarcane production;
• Increasing the sucrose yield of the sugar cane per unit of water used;
• Improve soil water drainage.
A number of options were considered. On economic grounds the option of converting from
sprinkler to drip irrigation was ranked the highest. One of the primary benefits of the drip
irrigation option was its sucrose yield per hectare compared to sprinkler irrigation. A
performance evaluation carried out on a number of pilot plots showed an average increase of
1.6 tonnes per hectare up to 1997 when the redevelopment decision was made. This is shown
in Figure 4.17.

Once the new drip irrigation scheme was installed a post investment audit was carried out to
assess whether the benefits accruing from the cost of converting sprinkler to subsurface drip
provided an acceptable rate of return. The four major benefits from the redevelopment were:
• An increase in the yield. Sucrose yield was found to increase by 1.6 tonnes per hectare;
• Water saving. A water saving of 1.5 Ml/ha/year has been made. This has an opportunity
cost of some US$160/ha/year
• Operation and maintenance saving. Savings in maintenance, power and labour have
resulted in a saving of US$140/ha/year;
• Savings owing to power levelling The improvement in load factors on the power
supply has led to a saving in power of 4.6 kVA/ha/year.
12,000
10,000
Subsurface drip
8,000
Sugarcane area (ha)
Surface drip
Sprinkler
6,000 Furrow
4,000
2,000
0
1 2
1982 2001
Year
Figure 4.16 Growth in sugarcane area and the use of drip irrigation for the
Simunye sugar estate in Swaziland

22
21
20
Tonnes of sucrose per hectare
19
18
17
16
15
14
13
12
Severe drought with some
11 furrow fields put on a dryland
regime
10
1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997
Year
Furrow Sprinkler Drip
Figure 4.17 Sucrose productivity by irrigation type for the Simunye sugar
estate in Swaziland
Figure 4.18 shows the relative benefits as measured at post investment.
Water value
33%
Operation and
maintenance saving
45%
19%
3%
Increase in
Power levelling sucrose yield
saving
Figure 4.18 Relative value of the project benefits for the Simunye sugar
estate in Swaziland

Further monitoring and analysis was carried out on 46 fields converted to subsurface drip in
1998 and compared to 13 sprinkler fields of equivalent age and ratoon class distributed
throughout the scheme. The findings were that the average sucrose increment was 23% for
plant cane and 24% for the first ratoon and the water use efficiency was 29% better for plant
cane and 18% for the first ratoon (Reference 4.21). These results were superior to the
parameters used in the post-investment audit and confirm the gains in water use efficiency
and yield per unit of water that can be made by converting to more efficient irrigation
methods.
4.11 Livestock water use

The water requirements of livestock are influenced by several factors, including:
• Type of livestock,
• Pregnancy;
• Lactation;
• Type of diet;
• Feed intake;
• Temperature.
Estimating total livestock consumption is relatively simple. The assessment of livestock

consumption should be carried out as follows.
• Determine types of livestock;

• Use typical water consumption figures per head for each type of livestock;
• Determine number of each type of livestock in the area being assessed.
Typical water consumption figures for livestock are given in Table 4.2.
Table 4.2 Typical water consumption figures for livestock
Type of livestock Water consumption
Beef cattle 25 litres to 45 litres per head per day

Dairy cattle 40 litres to 60 litres per head per day
Horses 30 litres to 45 litres per head per day
Pigs 10 litres to 20 litres per head per day
Sheep and goats 4 litres to 10 litres per head per day
Chickens 30 litres to 40 litres per 100 birds per day
Turkeys 40 litres to 55 litres per 100 birds per day
Table 4.3 shows estimates of water requirements of beef cattle in different physiological
states and in different thermal requirements.

Table 4.3 Water consumption estimates for beef cattle in different thermal
environments
Thermal Water requirements

environment
>35°C 8 litres to 15 litres of water per kg of dry matter feed intake

25°C to 35°C 4 litres to 10 litres of water per kg of dry matter feed intake
15°C to 25°C 3 litres to 5 litres of water per kg of dry matter feed intake
-5°C to 15°C 2 litres to 4 litres of water per kg of dry matter feed intake
Note: Young and lactating animals require 10% to 50% more water.
Source: Reference 4.24 and 4.25
Dry matter intake per day can vary between 7 kg per day and 20 kg per day depending on the
type of livestock (Reference 4.24 and 4.25). It should be noted that estimates for livestock
water use are very variable. However, livestock water use is usually a minor user of water
when compared to irrigation water use.
4.12 References
General
4.1. Rothert, S. and Macy P. (1999) The Potential of Water Conservation and Demand
Management in Southern Africa: An Untapped River
http://www.irn.org/programs/lesotho/ws.report/ws4c.potl.agr.shtml
4.2 Jensen M.E. (1983) The design and operation of farm irrigation systems, American
Society of Agricultural Engineers
4.3 Hansen, V.E., Israelson, O.W. and Stringham, G.E. (1980) Irrigation principles and
practice, John Wiley and Sons
4.4 FAO (1995) Irrigation in Africa in figures, Food and Agriculture Organization,
Rome, Italy
Climatic data and stations
4.5 FAO (1976) Agro-meteorological field stations, Irrigation and drainage paper no. 27,
Food and Agriculture Organization, Rome, Italy
4.6 FAO (1984) Volume 1: Data for Africa, countries north of the equator Irrigation and
drainage paper no. 22, Food and Agriculture Organization, Rome, Italy
4.7 FAO (1984) Volume 2: Data for Africa, countries south of the equator Irrigation and
4.8 FAO (1993) CLIMWAT for CROPWAT, Food and Agriculture Organization, Rome,
Italy

Crop water requirements
4.9 Allen, R.G., Pereira, L.S., Raes, D, Smith, M. (1998) Crop evapotranspiration:
Guidelines for computing crop water requirements Irrigation and drainage paper no.
56, Food and Agriculture Organization, Rome, Italy
4.10 Doorenbos, J. and Pruitt, W.O., (1977) Crop water requirements, Irrigation and
4.11 FAO (1992) CROPWAT – A computer program for irrigation planning and
management Irrigation and drainage paper no. 46, Food and Agriculture
Organization, Rome, Italy
4.12 US Department of Agriculture Soil Conservation Service (1967) Irrigation water

requirements, TP21
4.13 Water Research Commission South Africa (2000) Irrigation requirements of selected
crops under small-scale production: Linking on-farm and on-station research WRC
Report No. 689/1/00
4.14 Water Research Commission South Africa (1996) SPAWAT 1.0 - A computer
program for estimating irrigation requirements in South Africa WRC Report No.
379/1/96
Effective rainfall
4.15 Dastane, N.G. (1974) Effective rainfall, Irrigation and drainage paper no. 25, Food
and Agriculture Organization, Rome, Italy
4.16 US Department of Agriculture Soil Conservation Service (1969) Engineering field

manual for conservation practices
Irrigation efficiencies
4.17 Bos, M.G. and Nugteren, J. (1974) On irrigation efficiencies IRLI publication
number 19
4.18 Rantz, S.E., et al (1982) Measurement and computation of streamflow U.S.

Geological Survey Water Supply Paper 2175, Volume I and II, p631
4.19 Rohwer, C. (1948) Seepage losses from irrigation channels Fort Collins, Colorado,
USA, A and M College, Technical Bulletin Volume 38, p98
4.20 Water Research Commission South Africa (1994) Maximising irrigation project
efficiency in different soil-climate-irrigation situations WRC Report No. 226/1/94
4.21 Merry, R.E. (2001) The challenges of an irrigation redevelopment project BSSCT
Autumn Technical meeting
4.22 Seckler, D. (1996) The new era of water resources management: From “dry” to
“wet” water savings http://www.cgiar.org/iwmi/pubs/pub001/REPORT01.PDF
International Irrigation Management Institute

Leaching requirements
4.23 Ayers, R.S. and Westcot, D.W. (1976) Water quality for agriculture, Irrigation and
Livestock water requirements
4.24 NRC (1981) Effect of Environment on Nutrient Requirements of Domestic Animals
4.25 Schulz, R.S. and Austin, T.A (1976) Estimating stock water use in rural water
systems, American Society of Civil Engineers, Journal of the Hydraulic Division Vol
102 no HY1 pp15-18
4.26 USGS (2000) National Handbook of recommended methods for water data
acquisition – Chapter 11 – Livestock use
Water use efficiency
4.27 FAO (2000) Crops and drops –Making the best use of water for agriculture ,
http://www.fao.org/DOCREP/005/Y3918E/y3918e00.htm#TopOfPage Food and
Agriculture Organization, Rome, Italy
4.28 Personal communication van der Zaag, P University of Zimbabwe (2002)
4.29 Pazvakawambwa, G., and van der Zaag, P (2000) The value of irrigation water in
Nyanyadzi smallholder irrigation scheme, Zimbabwe. Proceeding of the First
WARFSA/WaterNet Symposium 'Sustainable Use of Water Resources'. Maputo, 1 to
2 November 2000
4.30 FAO (1979) Yield response to water Irrigation and drainage paper no. 33, Food and
Agriculture Organization, Rome, Italy
4.31 FAO (2000) Deficit irrigation practices Water paper no. 22,
http://www.fao.org/docrep/004/Y3655E/Y3655E00.HTM Food and Agriculture
Organization, Rome, Italy

5. RURAL DOMESTIC WATER DEMAND AND USE
5.1 Background
5.1.1 Definition of water demand, consumption and use
Estimating domestic water demand and use at a catchment level for rural areas in southern
Africa is problematic owing to the lack of measured data available. Estimates of rural
domestic water demand and use are further complicated by the lack of definitions of the terms
used. Hence in the context of assessing rural water demand and use it is important that
various terms are clarified. Water demand is different from consumption. Consumption is
the actual volume of water consumed, whereas demand is how much water people would use
if they had the opportunity. Many of the surveys that have been carried out in rural areas in
southern Africa have assessed water consumption, not water demand. Definitions of water
consumption, water extracted, water use, water need and water demand are given below.
• Water consumption is the volume of water that is actually gathered from water points
such as wells, boreholes and communal taps.
• Water extracted is the measure of the water actually gathered at a water point and
associated wastage i.e. a measure of what is removed from the water point in total.
• Water use is the sum of water that is utilised but not always collected. An example of
water that is not collected from a water point is washing clothes in a river.
• Water need is the required water volume based upon a series of pre-set, often qualitative
assumptions and values regarding water use e.g. there are accepted values for the volume
of drinking water required by a person per day. An example of such a value is the World
Health Organisation’s recommended minimum figure of 50 litres per person per day;
• Water demand is the assessment of the perceived need of water by the recipient. It will
be affected by a variety of factors including the expectations, wealth, and education of the
individual. In many parts of rural Africa water demand can exceed water use.
For water use, consumption and extraction it is physically possible to measure the quantity of
water. However, water need and demand are subjective values that require a qualitative
assessment in order to be derived.
In some parts of southern Africa the concept of “primary water use” is enshrined in water
laws. Primary water use refers to the reasonable use of water for basic domestic needs. As
well as including human requirements they may also include requirements such as animal
consumption, brick-making, and dip tanks. Primary water uses, together with environmental
water requirements are often given priority with regards to water allocation under many
recent pieces of water legislation implemented in southern Africa.
5.1.2 Typical rural domestic water sources
In rural areas of southern Africa communities often rely upon a complex system of multiple
sources, which are generally used for various activities. These often include non-potable
sources of water where water quality is not the prime concern e.g. small dams and rivers are
often used for washing of clothes. Typical sources of water in rural areas of southern Africa
include:
• Community tap. A tap facility with communal access, usually gravity fed by a filtered
steam or groundwater supply.
• Homestead tap (yard). This is a private water supply from a tap located in the yard of
the homestead.
• Homestead tap (inside). This is a private water supply by tap located within one of the
homestead buildings.
• Community borehole. A borehole with a hand driven pump for community access.
• Private borehole. This is similar to a community borehole but with the access limited to
a select few.
• Unprotected open well. A well that is generally open to the environment and generally
not very deep (e.g. 2 m to 10 m). The water is prone to contamination by pathogens.
Water is collected using a bucket and rope.
• Protected well. A well that has been constructed with a cover, windlass/winch and
bucket that protects against pathogens entering the water.
• Protected/unprotected spring. Similar to a protected/unprotected well, except the water
wells up from the surrounding rock.
• River or stream. Water is accessed by direct collection from a watercourse.
• Permanent dam. Water is collected from the reservoir formed by the dam.
• Rain water harvesting. A method of collecting rainwater that runs off the roof of a
building.
The seasonality of the regional rainfall in rural areas of southern African countries has a
significant impact on the following:
• The type of water source used e.g. in the rainy season rain water harvesting may be the
main source of water, whereas in the dry season a deep borehole may be the only reliable
source of water;
• The security and reliability of the supply;
• The quality of the supply;
• Access to water.
These sources of water and the seasonality of the rainfall affect how much water is used in
rural areas of southern Africa.
5.1.3 Background to rural domestic water supply and sanitation

schemes
Rural domestic water supply and sanitation schemes are usually typified by the following:
• Lack of physical infrastructure (i.e. a low level of service). For example:

- There is often a lack of a piped water system or a water borne sewerage system;
- Water sources are commonly rivers, hand dug open wells or boreholes;
- Travel times to water sources can be high;
• Operation and maintenance of rural water supply schemes in sub-Saharan Africa often takes
place at a community level;
• The outreach of rural government agencies to rural water and sanitation schemes is often
limited.
Figures for rural domestic water coverage in South Africa are shown in Table 5.1.

Table 5.1 Rural domestic water coverage in South Africa
Level of Description Percentage

service coverage (%)
Minimal No infrastructure in place 40

Upgradeable Upgrading required to be classified as basic 25
Basic 25 litre per person per day within 200 m of every resident 20
Intermediate Households have access to yard taps 10
High Households have access to in-house connections 5
Table 5.1 indicates that even for South Africa, which is a relatively developed country by
African standards, the level of service for rural domestic water supply is generally low.
Domestic water demand and use in rural areas of the developing world may be defined as the
water required to fulfil basic water supply and sanitation needs. These are defined as follows:
• Basic water supply is normally taken to mean all water used for drinking, food preparation,
bathing, laundry, dishwashing, and cleaning, but can also include the water needed for
watering animals and small gardens;
• Basic sanitation is the water used for waste disposal.
Numerous rural water supply studies have shown that the daily level of domestic water use
varies considerably in rural areas of Africa. Generally domestic rural water demand will be
between the following two figures:
• In Angola rural domestic water use is estimated to be approximately 15 litres per person per
day;
• In Madagascar rural domestic water use is estimated to be approximately 270 litres per
person per day.
In the majority of rural areas of Africa domestic water use was found to vary from some 20
litres per person per day to 40 litres per person per day. However, the World Bank quotes a
figure of 50 litres per person per day. The World Health Organization quotes a figure of 150
litres per household per day in order to provide adequate health and sanitation. It is important
to note that both rural domestic water demand and use are affected by a number of complex
factors. It is rarely possible to use one simple per capita figure on a catchment basis to assess
the demand and use. These factors are discussed in Section 5.3. As a consequence any rural
domestic water demand and use assessments should be based on the various factors that
influence water demand and use. Section 5.2 gives typical figures for rural domestic water use
for a variety of countries in southern Africa.
5.2 Typical rural domestic water use figures for southern

Africa
In general there are few measured rural water use figures available for southern Africa.
However, details of the figures that are available are given in the Sections below.

5.2.1 Angola
There is very little data available on rural water demand and use for Angola. A feasibility
study carried out in 1995 for the Epupa hydropower scheme used a figure of 50 l/person/day
for rural water use. However, this figure is about twice what had been previously reported
(Reference 5.13).
5.2.2 Botswana
The available data for rural areas in Botswana pertains to “urban villages”. These are
settlements in rural areas that have populations of several thousand and are supplied by a
piped water supply scheme. Table 5.2 gives typical water production for urban villages in
Botswana. It should be noted that the figures refer to water production and not consumption.
The quantity of water actually consumed will be lower than these figures owing to losses in
the water supply networks.
Table 5.2 Water production for urban villages in Botswana in 1995

(l/person/day)
Location Production (l/person/day)
Kanye 62
Ramotswa 51
Maun 103
Moshopa 48
Molepolole 26
Serowe 96
Palapye 91
Thamaga 54
Mochudi 81
Tlokweng 64
Lethakane 85
Tsabong 113
Mahalapye 99
5.2.3 South Africa
In May 2000 a survey of primary water use was carried out in the Mkonmazi River catchment
in the province of KwaZulu-Natal South Africa. The primary water demand and use is the
water that is required to maintain life and livelihoods. The survey was carried out of some
1,000 settlements. The average water use for the Mkonmazi River catchment was found to be
30.8 l/person/day. The minimum figure supplied was 5 l/person/day. The maximum reported
was 122.4 l/person/day (Reference 5.3). These figures incorporated the following uses:
• Drinking;
• Cooking;
• Washing people;
• Washing clothes;
• Drinking water for animals (primarily dogs and chickens);
• Gardening although water may be re-used for gardens after its primary use.

At the time of the 1994 democratic elections, rural South Africa was very poorly served with
water supplies. A severe drought in 1991/1992 highlighted the weaknesses in the operation
and maintenance of such systems that had been installed. Since then, the Department of
Water Affairs and Forestry has embarked on a programme to provide a water supply service
to everyone, at a minimum level of 25 litres per person per day within 200 metres walking
distance.
5.2.4 Swaziland
In 1999 a survey was carried out in the Mbuluzi catchment in Swaziland. Some 950
homesteads over an area of 200 km2 were questioned as to their water use. These homesteads
house approximately 7,600 people, which is around 10% of the total population of the
catchment. The average water use was found to be 15.7 l/person/day (Reference 5.4).
Water use figures used by the Swaziland Government in the planning of rural water supply
schemes are given in Table 5.3.
Table 5.3 Estimated rural water demand for Swaziland
Description Water demand (litres per

person/head of livestock per day)
Domestic demand 40
Cattle 35
Sheep and goats 5
5.2.5 Tanzania, Kenya and Uganda
Water use data for the countries of Tanzania, Kenya and Uganda have been collected as part
of the Drawers of Water II (DOW II) study. The DOW II study was carried out to assess
water use in east Africa. The study collected data during 1997 to 2000 and revisited the sites
surveyed in the Drawers of Water I (DOW I) study conducted between 1967 to 1968. The
data reported in DOW I were obtained by interviews and observations at 34 study sites in
Kenya, Tanzania and Uganda. These sites are shown in Figure 5.1. Research for DOW II
sought to carry out a comprehensive, repeat, cross-sectional analysis by replicating the
original study closely and adding additional lines of inquiry, where appropriate. As in the
original study, information was collected through a household survey in which interviewers
spent an entire day with each household, observing water use patterns and gathering other
socio-economic and environmental health data. Additional data were collected separately
about each site through interviews with key informants, field observations and review of
secondary literature. DOW II achieved a considerably higher sample size of 1015 households
compared with 713 in DOW I (Reference 5.5).
At a regional level, average daily per capita water use has declined by 30 percent over the last
thirty years, from 55 to 39 litres. This is a reflection of the almost universal decline in water
use by piped households. While water use by unpiped households has almost doubled (rising
from 11 to 19.5 litres), use by piped households has decreased by approximately 50 percent.
Despite this decline, piped households still use over three times the amount of water
consumed by unpiped households (Reference 5.5). The results of the study are illustrated in
Figure 5.2.
The average per capita water use was found to be 38 litres per day. However, there were
major differences in the quantity of water used by piped and unpiped households. By
definition, ‘piped’ households have piped water supplied to their home, while ‘unpiped’
households do not, and therefore obtain water from sources outside the home (Reference 5.5).
Piped households used on average almost three times more water per capita than unpiped
households. Similarly, urban sites had significantly higher levels of per capita water use than
rural sites.
Lake Terkane
UGANDA
Alemi
Lake Kyoga Lake Bisina
KENYA
Lake Albert
Kamuti
Moey’s Bridge
Tororo
Kasangati Iganga Mutwat
Muluyu
Lake Edward Manyata
Kiambaa
Mwisi Karuri
Nairobi
Lake Victoria Nasii
Makaa
Mkuu
Lake Kayasi Moshi
Lake Manyara
TANZANIA
Lake Tanganyika
Kipanga Dodoma
Dar Es Salaam
Water carried from sources 0 150 km
Piped connections to households
Figure 5.1 Settlements surveyed for the Drawers of Water II study

Region 6 45.4
22.4
Urban 5 65.3
26.2
Rural 4 58.5
8.3
Uganda 3 44.1
18.7
Tanzania 2 60
23.7
Kenya 56.5
1
20.5
Unpiped Piped
0 10 20 30 40 50 60 70
Water use (l/person/day)
Figure 5.2 Water use figures for piped and unpiped sources from the
Drawers of Water II study
A recent paper on rural water use in the Mwanza region of Tanzania provides figures for rural
domestic water use. These figures are shown in Table 5.4.
Table 5.4 Rural domestic water use for Mwanza in Tanzania
Location Cooking and drinking Hygiene

(l/person/day) (l/person/day)
Minimum for healthy living 2 8

Mwanza rural average 5 14
Typical figure for a developed country 8 192
5.2.6 Zimbabwe
In 1999 a survey was carried out of domestic water use in the Mhondoro communal lands in
the Mupfure catchment of Zimbabwe. The estimation of per capita water demand was based
on the water brought to the homestead only. As a result water used for purposes other than
those undertaken within the homestead are not accounted. Approximately 1,470 households
were surveyed with a combined population of some 7,300 people. The amount of water used
was found to be 16 l/person/day (Reference 5.8). However, as previously stated it should be
noted that this is only the quantity of water used for functions undertaken in the homestead
e.g. cooking, washing dishes, drinking, bathing. Hence water used for other purposes such as
washing clothes, bathing or gardening away from the homestead is not accounted for in this
figure. If water used for other purposes undertaken elsewhere, away from the homestead,
then the average per capita water use would be much higher than 16 litres.
Tables 5.5 and 5.6 provide water use figures used in a national master plan for rural water
supply and sanitation carried out in Zimbabwe in 1985.
Table 5.5 Rural domestic water demand figures for Zimbabwe 1985
Description Water demand

(l/person/day)
Individual connection in a rural area 60

Communal taps within 300 m of homestead 40
Communal taps greater than 300 m from the homestead 25
Boreholes with handpumps less than 300 m from the homestead 30
Wells less than 300 m from the homestead 30
Table 5.6 Domestic water demand figures for institutions in rural areas
Description Water demand
Rural clinics 10 l/outpatient/day

60 l/inpatient/day
Rural hospitals 200 l/patient/day
Rural shops 200 l/shop/day
5.3 Indirect and direct methods of estimating rural water

demand and use
There are two main methods of assessing rural domestic demand and use. These are:
• Indirect methods, where the quantity of water consumed is calculated from population
levels and estimated demand levels in terms of per capita consumption;
• Direct methods where socio-economic surveys and participatory techniques involving the
relevant stakeholders are used to estimate the current and future water demand and use.
In general estimation of rural water demand and use is difficult because:
• The majority of rural domestic water supply systems are unmetered;

• Data concerning domestic rural water demand and use is often expensive and time
consuming to collect;
• The level of service provided by the water supply system is often unknown.
For catchment management purposes indirect methods are the most appropriate for
establishing rural domestic water demand and use. However, in some cases the results of
socio-economic surveys and participatory techniques can be extrapolated to a sub-catchment
and catchment level. The main methods for estimating rural domestic water demand and use
are discussed below. It should be noted that the accuracy of a demand assessment is a trade
off between the budget needed for an accurate demand assessment and the predicted
usefulness of the results.

5.4 Factors affecting rural domestic water demand and use
Many water supply schemes in rural areas are fairly basic, however, the patterns of use and
demand are more complicated and tend to be dependent on many factors. Such factors
include:
• Population;
• Household occupancy rate;
• Level of service of the water supply for each household;
• Tariff levels;
• Willingness and ability to pay;
• Local knowledge and indigenous practices;
• Cultural values, traditions and religious beliefs;
• Climate;
• Water quality.
The more important of these factors are discussed briefly below.
5.4.1 Population
Water demand and use is directly related to the population. However, in rural areas it is often
difficult to estimate the population levels accurately. This due to the following:
• Lack of accurate, up-to-date census data;

• Lack of up-to-date aerial photography or remote sensing data from which to estimate the
numbers of settlements in an area;
• Migratory labour with the male members of households often working in urban areas for
long periods of time.
5.4.2 Household occupancy rates
Household occupancy rates affect water use. Studies have shown that low occupancy
households generally use more water per head than higher occupancy ones. This trend is shown
in Figure 5.1.

Water use per person in household per day
Number of people in household
Figure 5.3 Water use per household versus household occupancy
5.4.3 Level of service
The level of service can be defined in the following terms:
• Technical terms as follows:

- For water supply in terms of the quantity and quality of the water available within a
given distance;
- For sanitation in terms of whether there is a pit latrine, pour-flush latrine or piped
sewerage;
• Performance terms e.g. a stipulated measure of reliability.
Typical definitions of levels of service are shown in Table 5.7.

Table 5.7 Typical levels of service providing access to safe water supply and
sanitation
Level of Water supply Drainage water Sanitation

service disposal
Deficient Water source unsafe or None Open defecation or

inadequate or return dirty communal
time travel to source latrine
more than 30 minutes
Minimum Communal point source Soakaway or other Simple pit latrine on
with safe and adequate drainage at public householder’s plot
water and appropriate water point. Some
drainage, return time disposal of wastewater
less than 30 minutes at household level on
plot or onto field, or in
urban areas, gutter or
opened or covered
drainage channel
Intermediate Point source (e.g. yard Soakaway on plot or Ventilated improved
tap) on householder’s open or covered drain pit latrine or pour-
plot with safe and from plot to safe flush toilet on
adequate water supply disposal; connecting householder’s plot
and appropriate channels within plot
drainage (made by
householder)
High Piped connection into Open drain to safe Flush toilet with
house with safe and disposal or pipe to septic tank or
adequate water under septic tank or sewerage (if water
continuous pressure sewerage supply is sufficient)
Table 5.8 gives water use for various levels of service. The figures indicate that there is a
considerable variation between various studies that have been carried out.

Table 5.8 Estimated domestic water use for various levels of service
Level of Water use (litres per person per day)

service
Cairncross Hofkes et Jinja PDG 1996 Van Webster
and al 1981 Uganda Schelkwyk 1999
Feachem
1993
Communal
standpipes
Greater than <16 - - 15 25 -

200 m walking
distance
Less than 200 16 30 35.5 30 35 25
m walking
distance
Yard >16 40 50 70 80 80
connection
House
connection
Single tap 50 155 120 130 130
Multiple tap 150 250
Table 5.9 indicates the effect of different sanitation systems on water use.
Table 5.9 Effect of the type of sanitation system on water use
Type of sanitation system Water required for operation (litres per

person per day)
Simple pit latrine Nil

Pour-flush latrine 5 to 25
Ventilated improved pit latrine Nil
On-site septic tank 5 to 40
Sewered interception tank 5 to 40
Conventional sewerage >60
Many studies have been undertaken to try and quantify rural water use and demand in relation
to distance from the drinking water source. Most studies have indicated that an inverse
relationship exists between travel distance and consumption. As the distance to the drinking
water decreases domestic water use increases markedly. However, there is a threshold
distance at which the basic water use will not decrease. One study in southern Africa
concluded that, although per capita water use drops rapidly once there is not a water
connection on the property, water use varies little between households around 100 m from a
standpipe and the ones where the water source is several kilometres away. This concept is
shown in Figure 5.2.

Water use per person per day
Threshold distance after which water use

does not decrease significantly with distance
from the source
Distance of household from water source
Figure 5.4 Water use versus distance from source
5.4.4 Tariff levels
Water is often viewed as primarily a social good. Many countries’ constitutions enshrine the
basic right of every citizen to a safe supply of water. However, water can also be viewed as
an economic good. In many circumstances rural domestic water users have to pay a tariff
towards the cost of their water supply. In some cases consumers are able to choose the level
of service for which they are willing to pay. Demand functions and curves can be produced
that are related to the following:
• Quantity of water demanded;

• Price or tariff level of the water;
• Price of other related goods or services;
• Household income;
• Other socio-economic factors.
The above variables are often known as determinants of demand. Many researchers agree
that different levels of supply and different levels of service will display different functions
regarding demand. However, as a rule, water use will generally increase as the level of
service increases and decrease as tariff levels increase. This is shown diagrammatically in
Figure 5.5.

Tariff
level
Ho
us
Ya ec
rd on
Co
co ne
cti
mm
nn on
ec
un
t io
al
n Increasing level
sta
of service
nd
pip
es
Q1 Q2 Q3 Consumption
Figure 5.5 Effect of tariff levels and levels of service on water use
The determinants of demand influence the “elasticity of demand” i.e. the impact a variable
will have on demand. Major factors that influence water supply demand elasticity are:
• Price i.e. tariff levels;

• Income and affordability of the supply;
• Metering of the supply;
• Other factors e.g. distance from the source, level of service.
The demand for drinking water (and water for other basic needs e.g. basic hygiene) is
generally considered to be “price inelastic”, whereas demand for non-drinking water is “price
elastic”. This concept is shown in Figure 5.6.

Tariff
level
Non-drinking water
(elastic)
Drinking water
(inelastic)
Consumption
Figure 5.6 Elasticity of water demand for different uses
5.5 Indirect methods of estimating rural domestic water

demand and use
Indirect methods for estimating rural demand and use are relatively straight forward to use
and are the most practical methods for the estimation of water demand and use on a sub-
catchment and catchment basis. The following information is required:
• Population data;
• Per capita water demand;
• Unaccounted for water levels i.e. the difference between the total quantity of water
abstracted and the quantity of water consumed.
Figure 5.7 shows diagrammatically the process for calculating rural water demand and use.

Population surveys
Population data Census data
Settlement count
Tariff levels
Per capita water

demand
Level of service
Total water demand

Level of leakage
Unaccounted for Unauthorised

water abstractions
Other
losses
Total water use
Figure 5.7 Indirect method of calculating total water demand and use
Indirect methods are the simplest, and often the most appropriate method to use when
estimating water demand and use on a catchment and sub-catchment basis.

Example of applying an indirect method
Rural domestic water demand and use is to be estimated for the River Kadzi sub-catchment in northern
Zimbabwe. The estimate could be carried out as follows:
(i) The number of individual settlements has been calculated from up-to-date aerial photography.
There are estimated to be 16400 individual settlements within the Kadzi sub-catchment.
(ii) Local census data, health surveys, economic and engineering studies indicate the following:
• 35% of settlements have a well or borehole more than 200 m from a water source and use
pit latrines for sanitation. These settlements have been estimated to use 20 litres per
person per day;
• 45% of settlements are less than 200 m from a well or borehole and use pit latrines for
sanitation. These settlements have been estimated to use 30 litres per person per day;
• 15% of settlements have a yard tap and use pour-flush latrines. These settlements have
been estimated to use 80 litres per person per day and have unaccounted for water of 25%
owing to leakage;
• 5% of settlements have multiple taps in houses and use flush toilets connected to septic
tanks. These settlements have been estimated to use 150 litres per person per day and
have unaccounted for water owing to leakage of 30%.
(iii) The occupancy rate of settlements in the Kadzi sub-catchment has been found to be
approximately 7.3 people per household.
The calculation would be performed as follows:
Water demand = number of settlements x occupancy rate x water demand per capita
Water use = number of settlements x occupancy rate x water use per capita
Water demand for settlements > 200 m from a water source = 0.35x16400x7.3x20 = 0.84 Ml/day
Water demand for settlements < 200 m from a water source = 0.45x16400x7.3x30 = 1.62 Ml/day
Water demand for settlements with yard taps = 0.15x16400x7.3x80 = 1.44 Ml/day
Water demand for settlements with taps in the house = 0.05x16400x7.3x150 = 0.90 Ml/day
Total rural domestic water demand for the Kadzi sub-catchment is 3.92 Megalitres per day
The total rural domestic water demand for settlements supplied by wells and boreholes has been
assumed to be equal to water use (i.e. unaccounted for water is negligible).
Water use for settlements > 200 m from a water source = 0.35x16400x7.3x20 = 0.84 Ml/day
Water use for settlements < 200 m from a water source = 0.45x16400x7.3x30 = 1.62 Ml/day
Water use for settlements with yard taps = (0.15x16400x7.3x80)/(1 – 0.25)
= 1.92 Ml/day
Water use for settlements with taps in the house = (0.05x16400x7.3x150)/(1 – 0.3)
= 1.28 Ml/day
Total rural domestic water use for the Kadzi sub-catchment is 5.66 Megalitres per day
5.6 Direct methods of estimating rural domestic water

demand and use
Various studies in sub-Saharan Africa have shown that rural domestic water use is related to a
“level of living index”. The level of living index is a function of the following:
• Population;
• Income;
• Education;
• Agricultural activity;

• Type of dwelling;
• Household size.
Direct methods of estimating rural domestic water demand encompass responsive approaches.
These approaches to domestic water supply development have gained considerable support in
the recent years in South Africa and internationally. This approach is being promoted because
supply-driven projects, where water is simply delivered to communities with little or no
involvement of community members, have largely failed around the world.
It should be noted that direct methods are primarily designed for detailed planning (e.g.
feasibility studies and design) of rural water supply schemes. For catchment managers and
water resources planners operating at a catchment and sub-catchment level it is recommended
that indirect methods are used to estimate water demands and use. Various direct methods are
outlined briefly below. Details of demand assessment techniques are also given in Table
5.10.
5.6.1 Direct interviews with individuals
Direct interviews with stakeholders can reveal a lot of information that cannot be obtained
from elsewhere. Often personal interviews may reveal more information than group
discussions, especially on matters relating to individual households. Sample questions
include:
• How many individuals are in the household ?

• What does your household use water for (e.g. cooking, bathing) ?
• How much water does your household use per day (in summer, in winter)?
• Does all the domestic water used by your household come from the same source or from
various sources?
• What are the sources (e.g. irrigation canal, well) and how far from your settlement are
they?
• Do you think the water supply is adequate?
• If not, how much do you think your household really needs a day for domestic use?
• Who collects the water for domestic use?
• How long in terms of minutes/hours does it take to walk to the nearest water supply
point?
• How much water can they carry at one time?
• How many times do they collect water a day?
• Which do you perceive as most important – water for domestic or agricultural use?
These observations and surveys can be used to estimate the level of service and per capita
water demand. It should be noted that such direct interview techniques are time consuming to
carry out, especially for catchment management purposes. It is also important that a
standardised structure is used for any interviews that are undertaken and that training of the
interviewers takes place to ensure consistent results.
5.6.2 Community discussions and focus groups
On more general issues, group discussion may provide more information as many different
opinions will be put forward and discussed. Sample questions could be:
• Are you satisfied with the current water supply and sanitation services?
• Does everyone have equal access to water for domestic use or are some people
disadvantaged? Why?

• Have you been given any incentives to conserve water?
• Would you be willing to pay for water? How much and what services would you expect
for your money?
• Do you currently buy water from vendors? If so, how much do you pay?
• Have you received any training on the water and sanitation services available to you?
• Have you got any local traditions and beliefs that affect the take up of water and
sanitation services?
5.6.3 Seasonal calendars and diaries
Seasonal calendars can be used like diaries as a record of water used per household over a
year for different domestic tasks. It should be noted that incentives may be necessary to
encourage people to complete the calendars accurately. Alternatively diaries can be used to
record amounts of water used for different uses and from which source.
5.6.4 Transect walks and direct observation
Direct observations can be used to note the following:
• What services already exist;

• What indigenous technologies and appropriate technologies are in use;
• How far people are walking to obtain water;
• How much water is being carried per person on any one journey;
• At what time of day most water is collected;
• What other water sources are available (e.g. irrigation canals, rivers, dams) and whether
these are being used.

Table 5.10 Demand assessment techniques: Water supply and sanitation
Elicit relative demand Participatory rapid Participatory rapid Revealed preference Contingent valuation ‘Real’ detailed options
between different appraisal (PRA): appraisal (PRA): surveys (RPS) method (CVM)* considered by community or
services* Internally facilitated* Externally facilitated* ballot
Description of Improvements to a wide Community volunteers are A variety of PRA RPSs estimate time and A questionnaire survey to Detailed options and their
technique variety of different encouraged and trained to techniques are used by financial costs of current determine the maximum implications are considered by
services such as water, undertake a participatory trained researchers or household behaviour (e.g. willingness to pay of communities using PRA or
drainage, roads are survey in their own facilitators to confirm the payments to water vendors) individuals for various ballot
considered by the community. Preferences preferences of different and time saved in collecting options for level of service
communities who express and commitments are then community groups who water. (including improved
their relative demand for agreed in meetings. are also involved in the reliability), payment
these services. Total analyses. arrangements, within the
funds available for each context of the current or
community area should be specified institutional regime
reasonably fixed.
Potential Simple easy and Very good community Good community sense Can provide reasonably Provides good data for More precise cost estimates
benefits understood sense of ownership of ownership accurate estimates of Project Appraisal lead to less confusion
Express ‘real’ demand if Enhances empowerment Extension staff can current time and cost Good data on willingness to Institutional charging of O&M
only in real terms Useful if the demand assess appropriate time expenditure and hence pay and potential revenues implications can be thoroughly
Preferences can be assessment involves to elicit the demand possible willingness to pay for different service levels assessed
refined during micro- ongoing negotiation Can enhance for service improvements assuming a thorough survey Can be used in a changing
planning empowerment Data and analysis is undertaken institutional environment
Inexpensive Can be used in a requirements are modest Can guide tariff subsidy and
Compatible with PRA changing institutional Good baseline data for cost-recovery policy
work environment impact assessment Similarity to public opinion
Compatible with PRA polls means the
conceptually easy for non-
specialists and politicians to
understand
Potential risks Possible group or Possible group bias Possible group bias Cannot estimate household Risks inhibiting community Risk of key decisions being
and strategic bias Liable to lack Process can be response to price increases decision making and based on misleading results
constraints Willingness to pay for technical/financial rigour manipulated by (including for new levels of ownership for instance by from an unrepresentative group
different service levels Reliant on skills being in extension workers who service options) raising expectations about unless care is taken to avoid
not readily known the community may not use sufficient Poverty may constrain particular options group bias
Process can be Requires substantial technical/financial rigour ability of poor people to Relatively high cost and Requires detailed cost
manipulated by extension flexibility by external if not adequately convert time saving results requires specialist information so earlier demand
workers who do use funding agencies and supervised from service into cash consultants for reliable assessment may need to use
sufficient technical or local support institutions Extension workers with payments for themselves results other methods
financial rigour good facilitation skills are Rarely used for sanitation Inaccuracies may occur in a Detailed work on some options
required projects changing institutional can be redundant
environment
Typical usage Suitable for village or More suitable where low- Suitable in most Suitable where substantial Suitable for informing Suitable where difficult choices
slum general tech, low-cost solutions situations, possibly water supply problems strategic decision on levels to be made
improvement projects. are definitely viable e.g. complimented by other exist. To be used in of service in large
NGOs often use this handpumps and latrines methods conjunction with PRA investment programmes.
technique
* Estimated costs of technically viable options are needed for these techniques

5.7 References
5.1 Webster M (1999) Effective demand for water supply in South Africa – Technical
and financial implications of design to meet demand, Water, Engineering and
Development Centre, Loughborough University, UK
5.2 Water and Engineering Development Centre Loughborough University (1999)

Guidance manual on water supply and sanitation programmes, Department For
International Development
5.3 Geodata Institute, University of Southampton, UK and School of Applied

Environmental Science, University of Natal, South Africa (2000) Mkomazi catchment
rural water use survey: Final report (unpublished)
http://www.iwrms.uni-jena.de/download/other-
reports/mkomazi_community_survey.pdf
5.4 Geodata Institute, University of Southampton, UK and Department of Geography,

University of Swaziland (2000) Mbulizi catchment rural water use survey: Final
report (unpublished)
5.5 International Institute for Environment and Development (2000) Drawers of water II
http://www.iied.org/agri/dowrv-general.html
5.6 Higgins, C. (2000) Rural water tenure in east Africa: A comparative study of legal
regimes and community responses to changing tenure patterns in Tanzania and
Kenya
5.7 Interconsult Consulting Engineers Norway (1985) National master plan for rural
water and sanitation in Zimbabwe (unpublished)
5.8 Madulu, N. F. (2000) Changing the water needs and uses in rural communities of the
Mwanza region, Tanzania, First WaterNet symposium: Sustainable use of water
resources
5.9 Tevera, D. and Kundishora, P. (2000) Water sources and demand patterns in the
Mupfure Catchment: The case of Ward VIII in Mhondoro Communal Area
(unpublished)
5.10 Calvo, C. M. (1994) SSATP Working Paper No. 11: Case Study on the Role of
Women in Rural Transport: Access of Women to Domestic Facilities (World Bank)
http://www.worldbank.org/gender/transport/Case_Studies/calvo11.htm
5.11 Mendiguren, J. and Mabelane, M. (2001) Economics of productive uses of domestic

water in rural areas: A case study from Bushbuckridge, South Africa Association for
Water and Rural Development
http://www.nri.org/WSS-IWRM/Reports/productive%20water%20use.pdf
5.12 Alcock, P (1986) An examination of household water consumption in the Inadi ward,
Kwazulu, The Civil Engineer in South Africa
5.13 Burmeister and Partners Namibia (1995) Epupa hydropower scheme project
formulation report: Part II technical/economic site assessment (unpublished)

5.14 Kgathi, D.L. (1998) Water demand, population and sustainability in Botswana:
Implications for development policy (unpublished)
http://www.iiasa.ac.at/Research/POP/pde/docs/bw/kgathi.htnl
5.15 Turton, A., Mucheleng'anga, C., Mbao, E., Musonda, W., Mukubesa, L. and Ng'oma
M (2001) Water Demand Management, Natural Resource Reconstruction and Social
Adaptive Capacity: A Case Study from Kolomo, Zambia Paper presented at the
WARFSA/Waternet Symposium

6. INDUSTRIAL WATER DEMAND AND USE
6.1 Background
Industrial water use includes water used for the following:
• Industrial processes such as fabrication, processing, washing and cooling;

• Mining;
• Hydropower generation;
• Thermal electric power generation.
This chapter provides specific water consumption for a wide variety of industries and
examples of how to carry out water audits for industrial plants.
6.2 Industrial processes water use

Urban industries tend to be metered, usually according to the volume of water used, through
an average charge or a variable marginal charge based on block rates. Where larger
commercial and industrial users tend to be metered, determining demands can be facilitated
through analysing meter records. Alternatively, they may have their own source, especially if
they are outside urban centres. Monitoring records may exist or there may be some indication
as to the energy consumption of pumps or other references to demand and use.
The factors affecting water demand vary widely between the different industrial operations.
The major factors common to the industrial and commercial sectors in determining water
demand are:
• Economic activity at a local, national and international level;

• Population;
• Industrial composition;
• Price of water;
• Access to supply and alternative supplies;
• Access to technology;
• Working practices.
Although information on industrial water use is often available from meter records, additional
information may be required to establish industrial water use including:
• Type of industries;
• Expected rates of growth or decline of industrial sector;
• Percentage of industries connected to the public supply;
• Tariff levels for water;
• Extent of the use of water saving technologies;
• Extent of water reuse;
• Effect of conservation measures on demand;
• Government incentives.
Another method for estimating industrial water use is look up tables. Appendix A provides
look up tables that detail typical water consumption figures for a variety of industries.
Industrial water use activities include water withdrawals from ground and surface water;
deliveries from public water suppliers; consumptive use through evaporation and product

incorporation (as in a bottling plant); water and wastewater treatment, recycling, releases to
wastewater collection systems, and return flow to ground and surface water. Large industrial
water users are more likely to obtain water directly from wells, rivers, lakes, and estuaries,
and may supplement this with water purchased from public water suppliers. Small industries,
especially in cities, are more likely to obtain water from public water suppliers. Even if water
is purchased from a public water supplier, the water may be treated by the industry before
use, especially if pure water is required.
Industrial consumptive use occurs either through evaporation during cooling and open-air
washing, or through product incorporation, especially in food processing, such as bottling or
canning. In recent years, industries have tended to decrease water withdrawals as they began
to recycle water within their plants to a greater extent than previously, often due to the high
cost of wastewater treatment required to meet the provisions of statutory instruments such as
clean water acts. There is an increased consumptive use rate associated with recycled water.
More recently, industries have decreased water withdrawals in response to decreasing supply
and an emphasis on decreasing use and production of hazardous-waste materials. After use,
wastewater may be treated onsite, released to wastewater collection systems, returned directly
to surface water or to septic systems, or a combination. Figure 6.1 shows the way in which
water supply, use and treatment are related for a typical industrial plant.
Withdrawal Public water

Groundwater or supply
surface water
Water treatment
Industrial water
use and
recycled use
Reclaimed
Manufacturing processes
In-plant conveyance
wastewater
Recycled
Cooling
use
Washwater
Remediation
Recycled
use
Wastewater treatment
Wastewater
collection and
return flow
Figure 6.1 Water supply, use and treatment for a typical industrial plant

6.3 Industrial water consumption statistics
The aim of the look up tables included in Appendix A is to allow staff responsible for water
management to assess whether the licensed volumes for an existing or proposed abstraction
are reasonable for a given industrial process. There is also some information on the water use
steps for each industrial process and potential water saving initiatives that could be employed
to reduce consumption levels.
Water consumption figures for industry are often expressed in terms of cubic metres of water
used per unit of product produced. For example for steel production water use is measured as
m3/tonne of steel produced, for beer it is measured in m3/m3 of beer produced. The method of
manufacture used by a particular industry affects its water use. Some industries are relatively
consistent in their water use because they use the same processes, the same equipment and
produce similar projects. It is these industries that are generally covered by the look up
tables. For some industries, e.g. the chemical industry, use different processes, different
equipment and produce very different products. For these types of industry the look up table
approach is not applicable and each process and site should be assessed on its own merits, not
a ‘typical consumption basis’.
The quality of the data in the look up tables varies. There are some industrial sectors where
data was available from a wide range of sources, for other industries the data available was
limited. In some cases water use figures are based on limited data provided directly by one
company. The data in the tables should be treated with care. However, the water
consumption figures generally provide at the very least an idea of the order of magnitude of
water consumption for various industries.
For industries that consume significant quantities of water, water consumption varies in an
approximately linear manner with an increase in production. Hence the planned production
capacity of an industrial plant is important in establishing the water consumption. The data in
the lookup tables tend to concentrate on industries that use relatively high quantities of water.
There is a clear distinction between the terms industrial water use and industrial water
consumption. Industrial water use is defined as the quantity of water that is abstracted from a
source (e.g. a river or borehole) for use by an industrial plant. In many industrial processes
significant quantities of the water that is abstracted can be re-cycled and re-used. Industrial
water consumption is the quantity of water that is “lost” (e.g. by being incorporated in a
product such as soft drinks, or through evaporation from cooling towers) during the
manufacture of a particular product. The look up tables in Appendix A provide details of
specific water consumption in terms of m3 of water consumed per unit of product produced.
6.4 Participatory methods for obtaining industrial water use

consumption data
Water use surveys can be carried out through the use of questionnaires. When such exercises
are undertaken it should be stressed to the various participants that the information they
provide will remain confidential. One method by which a high level of participation may be
achieved is to offer the chance of obtaining a free water minimisation audit. An example of a
questionnaire is shown below in Table 6.1.

Table 6.1 Example of an industrial water use questionnaire
Company name: Company address:
Contact name:
Job title:
Telephone number:
Please indicate business sector in which the company operates from the list below
Breweries Machinery and electronics Retail

Chemicals Metal processing Rubber
Construction Mining Textiles and leather
Dairies Mineral products Other (please specify)
Food and drink Paper and printing
Give details of the main company activities/products:
Please provide details of all water sources, annual consumption and purpose of use
Water source Annual consumption Units (e.g. m3) Purpose
Public supply
River
Borehole
Other
If water is re-cycled on site, please give details of total annual water use for each source
Water source Annual use Water source Annual use
Public supply Borehole

River Other
To assess individual water efficiencies it is necessary to have information on annual

production, output and the number of employees.
Annual production Units (e.g. tonnes)

Number of employees <10 10 to 50 50 to 100 100 to 250 >250
Have any water minimisation studies been carried out before ? Yes/No
Are any water saving devices or water efficient processes installed ? Yes/No
6.5 Water audits for industry

Carrying out a water audit out of an industrial establishment is one of the key activities that
assists in improving the specific water consumption of a site. An example of a simplified
water audit for a generic site is shown in Figure 6.2.

Water in
Water for
Evaporative gardens
losses
Parts
cleaning
Chiller and
Cooling rinsing
towers bath
Chiller
Office
Manufacturing area
General washing and
sanitation
Boiler
Wastewater Process mixing
treatment and
vat cleaning
Canteen
Washrooms
Air compressor Domestic effluent
cooling once- Process effluent
out out
through
Water balance summary
Source of water use Quantity used Percentage of

per year (litres) total (%)
Cooling: Cooling towers, boiler 7,966,000 38.3

Process use: Parts and mixing vat cleaning 3,848,000 18.5
Domestic: Toilets, taps, showers 3,536,000 17.0
Once-though cooling: air compressors and pumps 2,388,000 11.0
Gardening 832,000 4.0
General washing, sanitation and maintenance 561,600 2.7
Leaks (detected) 416,000 2.0
Food preparation: Canteen 312,000 1.5
Sub-total 19,859,000 95.5
Total water delivered to site 20,800,000 100.0
Unaccounted for water 941,000 4.5
Figure 6.2 Simplified water balance for a manufacturing facility
Details of how a water audit should be prepared for and undertaken are detailed in Section
6.5.1.
6.5.1 Preparation for a water audit
Before a water audit of an industrial establishment is carried out the following should be
undertaken:
• Establish the size of the facilities including the number of buildings and floor space;

• Obtain plumbing drawings if possible;
• Establish the operating schedule of the facility including the number of employees per
shift;
• Estimate or measurement of the total water consumption and specific water consumption
per unit of the product produced on a monthly basis;
• Produce an inventory of all water using equipment;
• Collation of all the records and measurements of water coming into the plant including:
water meters, tankers, private water sources (e.g. borehole owned by the plant);
• Location and size of all water meters;
• Calibration data for any meters.
The methods that can be employed to measure the in-plant industrial water use are detailed
below.
6.5.2 Measuring in-plant industrial water use
There are three main ways in which in-plant industrial water use can be measured. These are:
• Metering;
• Bucket and stopwatch method;
• Micro-weirs.
Metering
For large industrial establishments metering the specific processing equipment used for
different operations enables the water consumption for each process to be established. It is
possible to obtain temporary strap-on meters that can be used to determine approximate
flows. Temporary meters can be used to establish whether it is cost effective to install
permanent meters.
Bucket and stopwatch
The use of a bucket and stopwatch is a simple, cost effective and accurate tool for measuring
water use. This method involves collecting a specified water in a set amount of time (e.g. if a
five litre bucket was taking 10 seconds to fill, the flow rate would be 0.5 l/s).
Micro-weirs
Micro-weirs are hand held weirs that be used to measure low flows in confined spaces.
6.5.3 Conducting a water audit
At the commencement of a water audit of an industrial facility a walk-through survey should

be carried out with personnel from the facility who are knowledgeable of how water is used in
each part of the site. The following steps should be taken:
• Record the hours of operation of each piece of equipment. Identify water piping layout
and areas of older equipment. Note equipment that has multiple uses of water (e.g. a
water-cooled ice machine);
• Establish the discharge and quality of the flow for each use. This allows it to be
determined if water from one use can be re-used as a potential supply for another
application. Parameters that should be measured include:
- Temperature;
- pH;
- Total dissolved solids;
- Conductivity
- Biological and chemical oxygen demands;
- Metals;
- Oil;
• Measure the actual quantity of water being used. These methods are outlined in Section
6.5.2;
• Check the water quantity and quality specified with the industrial equipment’s operating
manuals. Equipment is sometimes operated at higher flows than actually required;
• Establish a regular programme of water meter reading. Meters monitoring processes that
use large quantities of water should be monitored more frequently than those monitoring
processes that do not consume large volume of water;
• Identity the quantity and quality of the wastewater from each process;
• Estimate any water that is generated by the processes. Establish whether it is possible to
use this water in another part of the site e.g. as cooling water;
• Prepare a water balance diagram as shown in Figure 6.2;
• If the unaccounted for water is greater than 10% revisit the major areas of water use, talk
further with plant operators or take additional measurements.
6.5.4 Measuring leakage
When carrying out a water audit it is important that an estimate of the leakage is made. Leaks
often occur at joints and seals. A visual or audio inspection is the best method by which to
assess leaks. Employees should be requested to report any leaks that occur. Under floor
leaks can be determined using specialised sonic leak detection equipment.
The volume of the leaks has should be established. Small drips can be measured using a
bucket. An approximate method for converting the number of drops per second to litres per
minute is given in Table 6.2.
Table 6.2 Conversion of drops per second to litres per second
Number of drops per second Litres per minute
1 0.023
2 0.045
3 0.068
4 0.091
5 0.114
Note: If a pipe is dripping at more than 5 drops per second the leak will be a steady stream of water
Rates of water loss for an approximately circular hole can be determined using Greeley’s
equation (Reference 6.2). This is of the form:
Q = 0.215A P
Where: A is the cross-sectional area of the leak in cm2;

P is the pressure in the pipe in Pascals;
Q is the leakage rate in litres per minute.
Leaks in joints or cracks can be estimated by an equation of the form.

Q = 0.0161A P
Where: A is the cross-sectional area of the leak in cm2;

P is the pressure in the pipe in Pascals;
Q is the leakage rate in litres per minute.
For example in a pipe with a 0.1 cm wide crack that is 3 cm long with a line pressure of
500 kPa, the leakage rate would be about 34 litres per minute.
6.6 The effect of improving water use efficiency on effluent

water quality
Often the advantages of improving water efficiency for an industrial plant requires a holistic
examination of the facilities operations in terms of not only water use by various components
used in the process but also in terms of effluent water quality. A good example of this is the
Abakor Abattoir in Johannesburg, South Africa. This abattoir uses single cell protein to
manufacture stock feed. Improvements in the treatment of the effluent (through single cell
protein recovery) reduced chemical oxygen demand (COD) values from 9,000 mg/l to
200 mg/l, resulting in industrial treatment charges reducing from R5 million to R300,000 per
annum. Owing to the improved quality of the effluent, further tertiary treatment was possible
to produce water of a high enough quality to be reused in the abattoir. These system
improvements reduced the water consumption and effluent discharge by 50%. At full
capacity production the cost of reclaimed water is 42% of the cost of the potable supply
(Reference 6.3).
6.7 Case studies of industrial water demand management

for three industries in Zimbabwe
6.7.1 Background to case studies
The city of Bulawayo in Zimbabwe has a population of one million and is faced with
extensive deterioration of its sewerage infrastructure. To finance the rehabilitation of the
public sewer system there will have to be substantial increases in water and effluent tariffs.
Recent studies have indicated that industries are responsible for 25% of Bulawayo’s total
water consumption (30,000 m3/day). A number of heavy industries in Bulawayo are a major
source of water pollution. Most “wet” industries discharge their effluents untreated to the
municipal sewers (Reference 6.4).
Detailed studies of water use and effluent water quality were carried out for three industries in
Bulawayo by the University of Zimbabwe. The industrial plants studied included a soft drink
manufacturing company, a sugar refining plant and a wire galvanising factory. For each of
the three industries an industrial water use survey was carried out. Water quality sampling
points were also identified within each process train.
6.7.2 Results from the galvanised wire manufacture
This enterprise about 2.104 x 10-4 m3/m2 “galvanised wire surface” municipal water and 0.058
m3/m2 “galvanised wire surface” borehole water for the production process. The calculated
specific water consumption for the industry is 0.059 m3/m2 of wire surface area treated.
According to research carried out in South Africa the target specific water consumption for
the industry is 0.100 m3/m2 for industries for operations treating in excess of
10,000 m2/month and 0.200 m3/m2 and for factories treating less than 10,000 m2/month
(Reference 6.3). A simplified flow diagram for the plant in Bulawayo is shown in Figure 6.3.
Raw wire
Descaling
Water Acid pickling Effluent

[5.260] (40% hydrochloric acid solution)
[5.260]
Water Rinsing Effluent
[5.260] (water) [5.260]
Water
Lubrication Effluent
[0.701] (Borax 90oC) [0.701]
Water
Wire drawing Effluent
[2.104]
[2.104]
Annealing
(lead bath 732oC)
Charcoal wipe
Fine wire drawing
Water Effluent
Quenching 1
[102.400] (Water 82oC) [92.200] Lubrication
(Borax 90oC)
Water Acid pickling Effluent
[3.507] (40% hydrochloric acid solution)
[3.507]
Water
Hot water rinsing Effluent
[470.600]
[470.600]
Flux application
(Ammonium chloride 82oC)
Hot dip galvanising

(Zinc 460oC)
Water Quenching 2 Effluent

[2.104] (Water 82oC) [2.104]
Coiling
-4 3 2
Note: Figures in 10 m of water used per m of galvanised wire surface area
Figure 6.3 Simplified flow diagram for the wire galvanising process plant
Although the plant in Bulawayo operates well below the target specific water consumption
values there is still the opportunity to save water.
During the process of drawing and galvanising the wire, after the annealing step, the wire is
quenched in a water bath. This step is necessary to prevent overheating of the acid bath
which is the next step in the process. When the temperature of the water in the bath ahs risen
to a level that renders the quenching capacity of the water ineffective, water is discharged to
the municipal sewer.
From the water balance carried out for the wire manufacturing industry in Bulawayo the
quenching process consumes approximately 0.01 m3 of water per m2 of galvanised wire
surface produced. This represents 17% of the total water consumption. Instead of
discharging the hot quench water as wastewater it can be recycled through a cooling tower
and then used as quench water again. The industry would realise about 0.009 m3 of water per
m2 of galvanised wire surface. This would improve the specific water consumption to
0.050 m3/m2 (Reference 6.4).
6.7.3 Results from the soft drink manufacturer
The manufacturing of soft drinks requires large volumes of water. In the case of the plant
studied the water is pre-treated on the site to meet the product quality requirements before
being used in the manufacturing process (Reference 6.4). This process is shown in Figure
6.4.
From research carried out in South Africa the target specific water consumption for a soft
drink manufacturing plant is 2.3 m3 of water for each m3 of soft drink produced. The soft
drink manufacturing process is operated at 3.5 m3 of water consumed per m3 of soft drink
produced. If a system of recycling the filter backwash water was introduced the specific
water consumption could be lowered to 3.3 m3/m3. However, the reduction in specific water
consumption is not significant because the backwash process only consumes about 5% of the
total water used in the plant. Alternatively a large percentage of filter backwash water can be
cascaded for use as service water. Once the initial solid content dirty water has gone to drain,
the remaining water used in the backwashing process can be reclaimed into a recovery
holding tank and then used for services requiring lower quality water e.g. floor washing. An
over-capacity water treatment plant can often result in large water wastage due to the
backwashing of unnecessary sand and carbon filters. The industry therefore has to optimise
on the amount of water for treatment and the backwashing process (Reference 6.4).
Water usage should be included as part of the selection criteria when purchasing major
equipment such as bottle washers, sprays and bottling machines. Of particular importance is
the water usage efficiency in the bottle-washers as they are responsible for a large percentage
of water intake (about 13,300 m3/month or 54% of the total water used). Older bottle washers
should be modified to ensure that bottle spraying is discontinued once the machine is shut off.
Automatic shut-off valves and high pressure, low-volume jets for hose pipes could also prove
to be effective in helping reduce water intake. Attention should also be paid to future
developments such as varying heat transfer systems, e.g. oil as a substitute for steam. At
present steam is used for heating up the caustic soda solution in the bottle washing machines.
Steam is generated from boilers that use about 0.354m3/100t “raw sugar” or 10% of total
water intake (Reference 6.4).

Municipal water
[1.148] Sugar
Filter backwash Water treatment

[0.158]
[0.990] Simple syrup
Final syrup Flavour/acid/colours
Effluent from
Carbonating/Blending/Filling cleaning in place
[2.104]
Crowning
Packaging
and storing
Consumer outlet
Product
3 2
Note: Figures in m of water used per m of soft drink manufactured
Figure 6.4 Simplified flow diagram for the soft drink manufacturing plant
6.7.4 Results from the sugar refinery
The process of refining sugar consists of four basic steps:
• Washing the raw sugar crystals;

• Adding water to crystals to form a solution;
• Clarifying and decolourising the solution;
• Re-crystallising and finishing the sucrose.
Cooling water from cooling towers water is used to condense the water vapours that boil out
of the sugar solution during the re-crystallisation step. This process is shown in Figure 6.5.
The condensed vapours, the condensate, that mixes with cooling water is contaminated with
sugar (in an amount the refiner desires to minimise for business reasons), which contributes
substantially to the chemical oxygen demand (COD) of the wastewater. Owing to this
contamination, water from the cooling circuit has to be periodically discharged (cooling tower
blow down) as wastewater, otherwise the sucrose content and the COD of the cooling water
would become excessive. There are a number of ways that water consumption can be
reduced. At present, the refinery is utilising approximately 166 m3 of water from the
municipal supply per 100 tonnes “raw sugar” and 169 m3 of water from boreholes per 100
tonnes of “raw sugar”. About 7400 tonnes of raw sugar is refined per month.

Condensate
[56 m3/100t]
Steam Vacuum pan
Condenser Liquid sugar

Evaporation
[23 m3/100t]
Hot well Cooling tower
Blow down Borehole water

[35 m3/100t] [58 m3/100t]
Figure 6.5 Existing cooling water and condensate system for the sugar
refinery
There are a number of ways in which water can be conserved at the sugar refinery. Cooling
water can be separated from the vapour by introducing heat exchangers to separate the
condensate from the cooling circuit. The condensate will be contaminated with sugar carried
over with the vapour from the vacuum pans, but can be utilised within the process as “sweet
water” for dissolving sugar or de-sweetening muds. The cooling water to the heat exchanger
would be that from the current cooling water circuit. From the heat exchanger, the now hot
cooling water would pass to the cooling towers, after which it would return to the cold water
inlet of the heat exchanger. In the condensers, cooling water currently enters at 29oC and
after contact with hot vapour, exits at 49oC. The design characteristic of the heat exchanger is
to provide a similar temperature for the cooling of the condensate. As occurs in the cooling
water system, the volume of condensate would increase at the rate at which the water vapour
currently condenses and is entrained (137 m3 of water per day or 56 m3 of water per 100
tonnes of “raw sugar”). This system could also help reduce amount of cooling tower blow
down because cooling water contamination would be minimised. Figure 6.6 shows the
existing and the proposed non-mixing system incorporating a heat exchanger, respectively
(Reference 6.4).

Condensate
[56 m3/100t]
Steam Vacuum pan
Condenser
Liquid sugar
Enclosed Build up
condensate [56 m3/100t]
tank
Evaporation
Heat exchanger [23 m3/100t]
To other processes Cooling tower

e.g. melting muds, de-sweetening
Blow down
[3.5 m3/100t] Borehole water
[26 m3/100t]
Figure 6.6 Proposed cooling water and condensate system for the sugar
refinery
6.7.5 Conclusions
For the three industries in Bulawayo that were studied, there is no doubt that the identified
water demand management measures show potential for savings in water. There is a lot which
manufacturers can achieve by firstly carrying out audits to identify areas of improvement
within the manufacturing process.
Regular, simple water monitoring surveys of the different water-using areas could be devised
to assist in the monitoring of water consumption, as well as to supply information as to the
state of equipment e.g. taps, pipes and valves. Initial selection of water using equipment is
also crucial. Modifications can also be done although usually costly in the short-term to
incorporate recycling and reuse (closed-loop systems). Above all, good house keeping and
awareness training for personnel could further help industries practice water demand
management and cleaner production.
The initiatives of water demand management can also help reduce the strength and volume of
industrial effluents and probably eliminate the use of toxic substances through substitution.
The benefits do not accrue only to the industry but to the local authority through reduced
costs of conveying and treating industrial effluents. Subsequently, the quality of effluent
discharged from municipal sewage treatment works to the environment would improve,
thereby limiting environmental damage and social costs leading to sustainable development.

6.8 References
6.1 Hadley, H.K. (1995) Water use guidelines: Industrial water use, US Geological
Survey
6.2 North Carolina Department of Environment and Natural Resources (1998) Water
efficiency manual for commercial, industrial and institutional facilities
6.3 Personal communication (1998) with Muller, J. General Manager of technology

development Abakor, Pretoria
6.4 Gumbo, B. Mlilo, S., Broome J., and Lumbroso D. (2002) Industrial water demand
management and cleaner production: A case study of three industries in Bulawayo,
Zimbabwe Third WaterNet Symposium: Integrating Water Supply and Water Demand
for Sustainable Use of Water Resources, Tanzania October 2002

7. URBAN WATER DEMAND AND USE
7.1 Background
Urban water demand and use can be divided into the following categories:
• Domestic water use and demands including:

- In-house use e.g. drinking, cooking, ablution, sanitation, house cleaning, laundry;
- Out-of-house use (garden watering, swimming pools, livestock, car washing);
• Commercial water use and demands for:
- Shops;
- Offices;
- Restaurants;
- Hotels;
- Railway stations;
- Airports;
• Institutional water use and demands for:
- Hospitals;
- Schools;
- Universities;
- Government offices;
- Military establishments;
• Unaccounted for water which takes into account the following:
- Authorised unmetered water use such as street cleaning, public parks, sewer
flushing and fire-fighting;
- Unauthorised unmetered water use such as leakage, consumer wastage
distribution losses, illegal connections and metering errors.
In many African cities urban water demands are often non-homogeneous owing to a range of
levels of service occurring within the same urban area. Levels of service can vary from
household connections to standpipes or to no service at all. Figure 7.1 shows the percentage
of the population in African cities served by various water sources. This figure was produced
from data collected from 43 African cities including the following urban areas in southern
Africa: Dar Es Salaam, Gaborone, Harare, Luanda, Lusaka, Maseru, Maputo, Port Louis and
Windhoek (Reference 7.1).
There are numerous ways in which the level of service can be defined. The South African
Government has defined four levels of service for the formally developed areas of the city.
These are shown in Table 7.1.

Borehole or pump
3% Other
2%
Public tap Unserved
21% 31%
Household connection
or yard tap
43%
Source: References 7.1 and 7.2
Figure 7.1 Water supply in large African cities: Source of water
Table 7.1 Service level categories used in South Africa for domestic water
supply
Level of service Definition
Inadequate No access to basic water supply as defined below

Basic (i) A minimum quantity of 25 litres of potable water per
person per day. This should be provided at a minimum
flow rate of not less than 10 litres per minute. The source
should be within 200 metres of a household. There
should not be more than seven days interruption of the
supply to any consumer per year i.e. water should be
available 98% of the time.
(ii) The provision of appropriate education in respect of
effective water use should be provided.
Intermediate Yard tap or yard tank
Full House connection
Section 7.2 provides details of typical urban water demand figures for a number of urban
areas in southern Africa. Sections 7.3 and 7.4 discuss the various methods available to
establish urban water demand and use.
7.2 Typical urban water demand figures for southern Africa

An overview of urban supply for southern Africa is given in Table 7.2. Urban water demand
is often quoted in terms of litres per capita per day (l/c/day) or litres per person per day
(l/person/day). It should be noted that the figures in Table 7.2 represent the water production
figures. These are the quantity of water produced at a water treatment plant and not the
quantity of water used at the point of consumption. In urban water supply schemes there is
usual a considerable difference between the water production figures and the water used at the
point of consumption. The difference is known as unaccounted for water which is caused by
a variety of factors (e.g. leakage, illegal connections, errors in metering readings).
Table 7.2 Overview of urban water supply for southern Africa
Country Total Percentage Largest city Population Water

urban of urban of largest production for
population population city the largest city
(million) covered (%) (million) (l/person/day)
Angola 4.40 34 Luanda 4.0 30

Botswana 0.82 100 Gabarone 0.13 286
Democratic Republic 15.64 89 Kinshasa 5.7 86
of the Congo
Lesotho 0.60 98 Maseru 0.27 81
Malawi 2.72 95 Blantyre 0.80 -
Mauritius 0.48 100 Port Louis 0.15 200
Mozambique 7.92 86 Maputo 0.97 133
Namibia 0.53 100 Windhoek 0.27 -
Seychelles 0.26 100 Greater Victoria 0.12 140
South Africa 20.33 92 - - -
Swaziland 0.27 - Mbabane 0.94 100
Tanzania 11.02 80 Dar Es Salaam 3.00 150
Zambia 3.63 88 Lusaka 1.21 225
Zimbabwe 4.12 100 Harare 2.38 156
Notes: The urban population is defined as the population living in urban centres according to national
criteria.
The percentage of the population covered by the water supply infrastructure is defined in terms
of access to water, based upon the type of technology employed, distance from the house and
quantity available. Access includes (1) household water connections that can have taps within
the house or within a private plot of land, or (2) public water points, including public standpipes,
boreholes with handpumps, protected dug wells, protected springs, rainwater collection, or
other locally defined technologies. Reasonable access to a public water point is broadly
defined as the availability of 20 litres per person per day of safe water from a public water point
located within one kilometre of the users’ dwelling.
Over the past five years performance indicators for African water utilities have been collected
by the Water Utilities Partnership. One of the parameters that was collected was per capita
domestic consumption. This indicator represents the average daily water consumption per
person per day.
For utilities where most domestic properties are metered, the total domestic consumption can
be estimated quite accurately. However, for utilities where the majority of the domestic
customers are not metered it can be difficult to determine the spilt between the actual
customer consumption and the unaccounted for water. In the UK the per capita domestic
water consumption figures range from around 130 l/person/day to 170 l/person/day. The
average domestic per capita consumption in the UK is approximately 150 l/person/day. In
southern Africa the average urban domestic per capita water consumption ranges from
35 l/person/day to over 370 l/person/day (Reference 7.4).
It should be noted that the accuracy of the various estimates of urban demand will vary
significantly and is dependent on the accuracy with which they are recorded. The figures
given in the Tables below provide an indication of the range of values that can be expected
for domestic urban water consumption in southern Africa.

7.2.1 Angola
In the year 2000 the population of Angola was estimated to be 13.1 million with 34.2% of the
population living in urban areas. The majority of Angola’s urban population is located in the
capital Luanda (Reference 7.2). The existing water supply infrastructure has degraded over
the past 20 years owing primarily to the civil war and the lack of capacity and policies in
place. There is very little data concerning urban water demand for urban centres in the
country. In 1997 it was reported that tankers, low-level tanks and standpipes supplied 70% of
Luanda’s population with only 30% of the population being supplied by the reticulation
network (Reference 7.5).
7.2.2 Botswana
Tables 7.3 and 7.4 provide details on urban water demand estimates for a number of cities in
Botswana in 1996. The domestic or residential per capita consumption ranges from about
64 l/person/day to 108 litres/person/day. The total urban water demand that includes
institutional and commercial users ranges from 131 l/person/day to 270 l/person/day
.
Table 7.3 Estimates of water demand for Botswanan cities in 1996
(million litres per year)
City Population Domestic or Institutional Commercial Total use

residential or industrial
Gaborone 170,000 6,243 5,287 2,722 14,252

Francistown 89,000 1,664 1,374 1,227 4,265
Selebi-Pilkwe 44,000 1,745 561 2,032 4,338
Lobatse 26,000 607 439 1,061 2,107
Table 7.4 Estimates of water demand for Botswanan cities in 1996

(litres per capita per day)
City Domestic or Total use

residential
Gaborone 100.6 229.7

Francistown 51.2 131.3
Selebi-Pilkwe 108.7 270.1
Lobatse 64.0 222.0
Average 89.8 212.0
It has been reported that in some urban areas, such as Selebi Phikwe, that are centred around
the mining industry, urban domestic water consumption is significantly higher than in other
urban areas. A water use and affordability study carried out in Botswana in the early 1990s
indicated that in some of the mining towns the mining company pays for the water bills of
their employees. This can lead to a culture of wasteful use and relative indifference to water
leakage resulting in high overall consumption figures.

7.2.3 Democratic Republic of the Congo
Owing to the continuing civil war in the Democratic Republic of the Congo there is very little
reliable information on urban domestic water demand. The figures in Table 7.1 indicate that
in 2000 water production for the town of Kinshasa was 86 l/person/day. However,
unaccounted for water for Kinshasa was reported to be 47%. Hence at the point of supply
only some 40 l/person/day of water is available. This is a low figure for an urban area and
indicates that either the reliability and/or the coverage of the water supply system in Kinshasa
is low.
7.2.4 Lesotho
The population of Lesotho was estimated at 2.0 million in 2000. About 28% of the
population reside in urban areas. In 1995 the urban water demand was estimated to be
109 l/person/day. This figure is relatively low for urban areas. It has been hypothesised that
that this is because some urban dwellers are not connected to the reticulation system or that
those that are connected use them infrequently. This is because there are water vendors who
sell water from community stand pipes at a lower rate than is charged by the water utilities
(Reference 7.5). In 2000 the Water and Sewerage Authority in Lesotho reported that their
urban per capita domestic water demand was 67 l/person/day (Reference 7.4).
7.2.5 Malawi
There are five water boards that are responsible for urban and peri-urban water supply in
Malawi. These are the Lilongwe, Blantyre, Southern, Central and Northern Water Boards. In
1995 the total urban water demand for Malawi was estimated to be 730 million m3 per year
equivalent to a per capita daily water demand of 180 litres.
Table 7.5 provides details of the urban domestic water consumption figures for three water
utilities in Malawi. It should be noted that the figures in Table 7.5 only refer to the proportion
of the flow that is used by domestic consumers.
Table 7.5 Urban domestic water consumption for water utilities in Malawi in 2000
Water utility Per capita demand (litres/capita/day)
Central Region Water Board 53.3

Northern Region Water Board 35.4
Lilongwe Water Board 95.0
The per capita urban domestic water demand for the Central and Northern Water Boards in
Malawi appear to be low indicating that the water supply to domestic consumers is prone to
interruptions.
7.2.6 Mauritius
In 2000 the population of Mauritius was some 1.2 million in the year 2000 of which 41.3%
reside in urban areas. The urban water consumption for the island has been estimated to be
approximately 135 litres/person/day (Reference 7.2).

7.2.7 Mozambique
In recent years there appears to be little readily available data for urban water demand for
Mozambican cities. The figures detailed in Table 7.6 are based on data from a study carried
out in 1992.
Table 7.6 Estimates of water demand for Mozambican cities in 1992

(million litres per year)
City Population Domestic Non- Losses Total Per capita

domestic domestic
demand
(l/person/day)
Maputo 590,324 15,527 10,500 16,920 42,950 199

Beira 83,594 2,850 1,330 3,136 7,317 240
Quelimane 19,986 342 293 254 889 122
Nampula 54,234 895 1,139 1,220 3,255 164
Lichinga 5,811 104 86 48 238 112
Chimoio 11,364 292 267 25 584 141
7.2.8 Namibia
Namibia is the most arid country in southern Africa. In recent years the Namibian
Government and Namibian water utilities have implemented many water demand
management and water conservation measures. As a result domestic water consumption for
cities in Namibia is relatively low compared with other urban areas in southern Africa. Table
7.7 provides per capita daily domestic water demand figures for the cities of Walvis Bay and
Windhoek.
Table 7.7 Estimates of domestic water consumption for Namibian water utilities in
2000
Water utility Per capita demand (l/c/day)
Municipality of Walvis Bay 108.2

City of Windhoek 120.3
7.2.9 South Africa
The population of South Africa was 42.8 million in 2000. Some 55% reside in urban areas.
Table 7.8 provides estimates of the domestic water demand for a number of metropolitan
areas.

Table 7.8 Estimates of domestic water demand for South African water utilities
and cities in 2000
Bloem Water (Bloemfontein) 98.6

Cape Town 219.2
Port Elizabeth 163.9
Durban Metro Water Services 158.0
City of Tygerberg 100.7
Manguang Local Municipality 171.1
Drakenstein Municipality 154.0
7.2.10 Swaziland
The main water utility in Swaziland is the Swaziland Water Services Corporation. The 1995
annual urban water demand was estimated to be 25 million m3 with a corresponding daily
urban domestic water demand of 57 litres/person/day (Reference 7.5).
7.2.11 Tanzania
The population of Tanzania was 33.7 million with 27.8% living in urban areas. The 1995
annual urban water demand for the country was estimated to be 1,690 million m3, equivalent
to a demand of 584 litres/person/day. Table 7.9 provides an estimate of domestic water
consumption for a variety of water utilities in Tanzania.
Table 7.9 Estimates of domestic water demand for Tanzanian water utilities
and cities in 2000
Mwanza Water and Sewerage Company 40.4

Tabora Urban Water and Sewerage Authority 88.5
Dar Es Salaam Water and Sewerage Authority 29.3
Urban Water Supply and Sewerage Authority 105.4
Dodoma Urban Water Supply and Sewerage Authority 194.5
Shinyanga Urban Water and Sewerage Authority 37.4
Tanga Urban Water Supply and Sewerage Authority 59.4
Arusha Urban Water Supply and Sewerage Authority 78.3
Table 7.9 indicates that several urban areas in Tanzania have very low per capita domestic
demand figures (e.g. the domestic water demand for Dar Es Salaam was stated to be only
29.3 l/person/day according to the local water utility). However, some of the figures are
unlikely to reflect the true level of water demand. For example study carried out in 1991 in
Dar Es Salaam revealed that the registered house and yard connections in the city accounted
for only 30% of the domestic urban water consumption (Reference 7.9). In Dar Es Salaam
many domestic water consumers rely on water vendors and other initiatives to fulfil their
water demands. In this way private, commercial and community initiatives compensate for
the shortcomings of the public water distribution system.

7.2.12 Zambia
The population of Zambia in 2000 was 10.1 million. The percentage of people residing in
urban areas is approximately 36%. Table 7.10 provides estimates of domestic water
consumption for a variety of water utilities in Zambia.
Table 7.10 Estimates of domestic water consumption for Zambian water utilities
in 2000
Chipata Water and Sewerage Company 115.0

Lusaka Water and Sewerage Company 254.4
Kafubu Water and Sewerage Company 371.2
7.2.13 Zimbabwe
The population of Zimbabwe in 2000 was 12.6 million of which 35.3% lived in urban areas.
In 1995 the annual urban water demand for the country was approximately 697 million m3.
This is equivalent to a per capita water demand of 169 litres. Estimates of urban domestic
water consumption for the cities of Mutare, Masvingo and Gweru are given in Table 7.11.
Table 7.11 Estimates of domestic water consumption for Zimbabwean water

utilities in 2000
City of Masvingo 171.0

City of Mutare 103.2
City of Gweru 137.4
7.3 Estimation of water use where records are available

There are number of methods by which urban water demand and use can be estimated.
Where demands are metered or records are available the following methods can be used:
• Estimation of water use using meters;

• Estimation of water use from pumping records;
• Estimation of water use using test metering.
Where water demands are not metered there are various other methods that can be employed.
These are detailed in Section 7.4.
7.3.1 Estimation of water use using meters
Where the water supply infrastructure is comprehensively metered water use can be
calculated from information that is often available through the public water supply utility.
Urban water use can be established from the following information:
• Records of total water production from water treatment works;

• Records from individual pumping stations (i.e. hours of pumping, duty flows of pumps,
source output meter records);
• Other water meter records (e.g. domestic and commercial properties);
• Estimates of unmetered consumption (alternative sources, government connections,
illegal connections).
It is important that water meters are cleaned and serviced regularly. The accuracy of meters
can be checked using a temporary weir. The number of broken or unreadable meters should
also be estimated.
7.3.2 Estimation of water use from pumping records
Urban water use may also be estimated from pumping records. This can be carried out by
collecting the following information:
• The pumping regime should be determined;

• It should be established whether water is being pumped to storage (e.g. elevated tanks,
service reservoirs) or directly into distribution (e.g. house connections, community tanks,
standpipes). It should be noted that water systems that pump to storage would have a
constant flow. This is because they are pumping against an approximately constant head.
Where water is pumped directly into the distribution system the flow will be determined
by the water use that may vary considerably;
• The number of hours per day on average that the pumps are operated, together with the
seasonal operating regime;
• The duty point of the pump should be established. The duty point usually coincides with
the peak efficiency of the pump. The duty point is shown diagrammatically in Figure 7.2.
Establishing the average number of hours of pumping per day and the normal flow of the
pumps (the duty point) will provide an estimate of the average flow per day in the water
system, and this can be used as an estimate of total consumption. It should be noted that in
old water supply schemes the pump might no longer be operating at its duty point resulting in
a different flow to the duty point marked on the pump or in the records.
The duty point can change for the following reasons:
• Water level in the sump may change;

• The roughness of the water distribution system pipeline will change with age;
• Wear and tear on pump impellers;
• Partial clogging of the suction pipe may occur.
An example of how the duty point of a pump can change is shown in Figure 7.3. If the water
level in a borehole or well falls by an amount ∆S, the curve for the water distribution system
will change and as a consequence the quantity of water pumped into the water distribution
system will decrease from Q1 to Q2.

Pump characteristics
Head H
Duty point
System
characteristics
Efficiency
curve
Operational
range
Discharge Q
Figure 7.2 Duty point of a pump
Pump characteristics
Head H
System
characteristics
∆S
Efficiency
curve
Hs
Q2 Q1 Discharge Q
Figure 7.3 Effect in the change in static head on discharge

7.3.3 Estimation of water use for unmetered conumers using test
metering
Where an urban water supply system is not metered water use can be determined using test
metering. This is carried out as follows:
• Classify domestic housing into five or six classes. Examples of domestic housing
classifications are given in Table 7.12;
• Select a relatively small sample (i.e. approximately 30 households) from each housing
class;
• Check that there is no leakage from these households;
• Carry out test metering (ascertaining the number of persons in each household over the
period of the tests). The test period should be between two to four weeks avoiding
holiday periods and extreme weather conditions e.g. periods of hot weather when water
use is likely to be higher than average. In parts of southern Africa where there is
pronounced difference in rainfall throughout the year, test metering should be carried out
both in the rainy and dry seasons;
• All standpipes should be test metered and the number of people using them in the test
period should be estimated;
• Calculate the mean per capita water use for each class of housing that has been defined;
• Test meter the largest water consuming establishments (e.g. institutions, government
buildings).
Table 7.12 Example of domestic housing classifications
Class Type of housing
1 High and upper middle income groups e.g. villas, detached houses, large flats
2 Average and lower middle income groups e.g. small houses and flats with one
or two toilets, one kitchen and one bath/shower
3 Low income groups e.g. tenement blocks with high density occupation (one
shower, one pour-flush latrine, one or two taps)
4 Lowest income groups e.g. low grade tenement blocks, one tap dwellings with
shared or no latrine, low pressure
5 Standpipe supplies
An alternative to a full test metering programme is to meter the flow through a single main
supplying a group of households. However, this masks exceptionally high or low
consumption rates and may include undiscovered leakage downstream of the metering point.
The total water use can be calculated as follows from test metered data:
• The total domestic consumption is estimated by marking the supply districts on a map of
the water distribution system, and plotting and measuring the amount of each class of
housing (including the population densities);
• The aggregate domestic consumption for each district can then be calculated using the
mean per capita consumption for each housing class;
• The commercial and institutional water use are apportioned per district from an analysis
of the test-metering;
• An allowance for unavoidable distribution leakage should be made for each district,
dependent on the age of the system and pressure. It should be noted that the minimum
night flow can be test-metered to evaluate system losses during near zero demand periods;

• Deductions should be made for any areas where consumption is below average (e.g.
where demand is not satisfied due to low pressure or excessive leakage).
The total urban water use is calculated as follows:
Total urban water use = Domestic water use + Institutional and commercial water use +
Unaccounted for water – Unsatisfied demand
Using an approach that utilises metering and information from pumps can provide a
reasonably accurate estimate of urban water use. However, in most urban situations, there
will be a substantial number of households (especially low-income ones) that do not have
access to a piped water supply, and obtain their water from other sources. In these cases an
approach based on demand estimation is required.
7.4 Estimates of unmetered urban water demand

Where water use is not metered there are several approaches to estimating urban water
demand. The conventional method uses current population and per capita demand for the
existing level of service to determine water demand.
There are several sources of data and methods that can be used to estimate the urban
population. These are briefly detailed below:
• Recent census figures. These are often the most reliable figures;
• Estimates made by the district or municipal administration;
• Settlement counts. These require access to aerial photographs or remote sensing
information. The population can be estimated by counting the number of houses and
multiplying this number by the mean household size;
• Self-survey, rapid assessment procedures can be used to estimate population levels.
Rapid assessment procedures include:
- walkabout observations and surveys;
- mini focus group discussions;
- informal talks with change agents/key informants living in the relevant areas;
• Population estimates can also often be obtained from community leaders.
7.4.1 Estimates of unmetered urban domestic demand
Urban domestic water demand is directly influenced by the following:
• Level of service;
• Type of housing and its location.
Table 7.13 gives some typical values for different types of housing and levels of service. It
should be noted that seasonal variations in demand should also be considered. Table 7.14
provides details of the figures used by the Lusaka Water and Sewerage Company for planning
water supply schemes.

Table 7.13 Typical design allowances for in-house water demand
Class Type of housing Demand

(litres per
person per
day)
1 High and upper middle income groups: villas, detached houses, large flats 200 to 250
2 Average and lower middle income groups: small houses and flats with one 130 to 180
or two toilets, one kitchen and one bath/shower
3 Low income groups: tenement blocks with high density occupation (one 70 to 130
shower, one pour-flush latrine, one or two taps)
4 Lowest income groups: low grade tenement blocks, one tap dwellings with 50 to 110
shared or no latrine, low pressure
5 Standpipe supplies (including washing) 25 to 70
Standpipe supplies (drinking and cooking only) 8 to 10
Table 7.14 Water consumption figures used by the Lusaka Water and Sewerage
company for planning
Class of area Domestic Garden Commercial/ Leakage Wastage

water use (l/c/day) Institutional rate rate
(l/c/day) water use
(l/ha/day)
Low density 250 80 NA 10% 10%

Medium density 200 40 NA 10% 20%
High density 100 20 NA 10% 30%
Compound 35 0 NA 10% 25%
Commercial NA NA 30,000 10% 10%
Institutional NA NA 30,000 10% 30%
Industrial NA NA 30,000 10% 10%
If levels of service are not known there are several methods that can be used to obtain them.
These include:
• Conducting a comprehensive survey to collect information on level of service in each

household (e.g. house connection or standpipe, number of water-using fittings) and the
size of household. It should be noted that this type of survey is expensive, time-
consuming and may not give reliable results;
• Rapid assessment procedures can be used to estimate the percentage of each level of
service and housing type, and population densities in each zone. However, the confidence
limits on the data collected is heavily dependent on the quality and experience of field
staff collecting it;
• Conduct participatory assessment of levels of service. Participatory methodologies are
not recommended purely for the estimation of water demand and use at a sub-catchment
and catchment level. However, if these methods are used as part of a feasibility study for
expansion of a water and sanitation system the results can be useful for water resources
planning purposes. Common methodologies include:
- Semi-structured interviewing;
- Focus group discussions;
- Preference ranking;

- Wealth ranking;
- Mapping;
- Participatory evaluation and planning.
The urban domestic water demand can then be estimated as follows:
• The supply districts should be marked on a map of the water supply and sanitation system
together with the class of housing and the level of service;
• The domestic demand for each district should be aggregated using the average demand
for each housing class in each location
Out-of-house demands (e.g. garden watering, filling and livestock watering) are dependent
largely on climate and metering. In hot, dry climates where water usage is not metered
garden watering can increase the water demand of higher income groups by some 30% to
50%. Similarly, water demand from standpipes can increase substantially where animals are
watered at or near the standpipe.
7.4.2 Estimate of the urban domestic demand of the population without

access to piped water supply
In many urban areas in Africa there is often a significant minority (or in some cases a
majority e.g. in Luanda) of the population who do not have access to a piped water supply.
Determining the demand of this proportion of the urban population is difficult. However, in
some urban areas it can represent a substantial proportion of total demand.
The following guidelines should be used for determining the demand of the population in
urban areas without access to a public water supply:
• Plot the existing supply areas on a map of the urban population;

• Estimate the size of the population without access to public water supply;
• Visit the areas which are not supplied with water and conduct studies (e.g. surveys, rapid
assessment procedures and participatory methods) to map the areas and confirm the
population (including the number of households, size of households, population density),
location of existing water sources, existing water consumption and expenditure on water,
existing sanitation, and any unusual local water use practices;
• The water demand can be estimated for the unsupplied areas based on observed or
surveyed water consumption. This figure should be compared with a minimum demand
of 25 litres per person per day.
7.4.3 Estimating unmetered urban institutional and commercial water

use and demand
Water use and demand information for commercial and institutional establishments may be
available from metered records. If commercial and institutional estimates of water use and
demand are not available they must be calculated. One estimation method is to apply a
demand allowance on a per capita basis for various institutions and commercial buildings.
Table 7.15 gives typical allowances for commercial and institutional establishments. These
allowances assume piped water connections and waterborne sanitation, and should be
adjusted down where the establishments have a lower level of service (e.g. standpipes,
handpumps or pit latrines at schools).

Table 7.15 Typical demand figures for commercial and institutional establishments
in urban areas
Usage Demand allowance
Small businesses, shops and offices Up to 25 litres per capita per day (applied as per
capita allowance to the whole urban population)
Offices 65 litres per day per employee*
Department stores 100 litres to 135 litres per day per employee*
Hospitals 350 to 500 litres per day per bed
Hotels 250 litres per day per bed
Schools 25 to 75 litres per day per pupil*
*Note: These figures should only be applied when the above are operating or open
Another method for calculating the institutional and commercial water use and demand is to
subtract the distribution system loss and domestic use and other uses (e.g. industrial and
public use) from the public supply figures.
7.4.4 Estimating unmetered public water demand and use
Water for parks, fountains, ornamental ponds, public gardens, and government buildings is
often supplied free of charge. Some studies carried out suggest that a figure as high as
45 litres per person per day may be appropriate for watering ‘green areas of town’ in hot dry
climates (provided there is sufficient water to meet the demand). Firefighting, hydrant-testing
and sewer flushing generally consumes an insignificant quantity, when expressed as an
average daily demand.
7.4.5 The effect of the number of occupants in a household on water

use
The number of occupants in a household can have a significant effect on water use. In 1987
the Water Research Commission in South Africa carried out a study to assess the difference in
per capita water consumption between houses and flats for middle income groups in the late
1980s. Table 7.16 indicates that the per capita indoor water consumption for houses is lower
than for flats. However, the occupancy rates of houses is higher than that of flats. This is
because there are certain common uses of water that are spread amongst people in the same
dwelling (e.g. laundry, kitchen use, floor washing). This leads to a drop in per capita water
consumption the more occupants there are per dwelling. Figure 7.4 illustrates this
phenomenon.

Table 7.16 Comparison of water use and occupancy rates for middle-income flats
and house in South Africa
City Indoor water use Occupants per dwelling

(litres per person per day)
House Flat Houses Flat
Port Elizabeth 148 176 4.4 1.9

Pretoria 154 224 4.0 2.2
Bloemfontein 145 270 4.1 1.4
Kiowa (Bulk metered) - 170 - 2.4
Kiowa (Individual meters) - 130 - 2.4
300
31 flats in Bloemfontein
Mean for 12 blocks of flats in Pretoria

250
Mean consumption (l/c/day)
Mean for 19 houses in

Pretoria with borehole
200
150
Mean for 109 flats in Port Elizabeth
Mean for houses in Port Elizabeth

100
50
0
0 1 2 3 4 5
Occupants per dwelling
Figure 7.4 Average internal water consumption for middle income group
dwellings in South Africa in 1987
There is no other data available for other countries in southern Africa relating occupancy rates
of households to water consumption. However, water consumption data for urban areas in
other parts of the world indicates that per capita water consumption generally decreases with
an increase in the number of occupants of a household.

7.5 Unaccounted for water
7.5.1 Background
Unaccounted for water can be defined as the difference between the volume of water put into
the supply and the authorised volume used by the consumers. In some respects unaccounted
for water is a misleading term. This is because unaccounted for water includes unmeasured
water put to beneficial use (e.g. firefighting) as well as water losses form the system.
Unaccounted for water is made up of the following:
• Authorised unmetered uses include fire fighting, main flushing, process water at for water
treatment plants and landscaping of public areas;
• Unauthorised water losses from the water supply system e.g. caused by leakage or illegal
connections;
• Meter reading and billing errors;
• Accounting procedure errors;
• Malfunctioning distribution system controls e.g. meter error at various levels in the
supply system.
Table 7.17 and 7.18 give unaccounted for water figures for a number of southern African
water utilities and countries. Unaccounted for water figures vary from 16% for Walvis Bay in
Namibia up to 65% for Mwanza in Tanzania. As a comparison, for many countries in Europe
an unaccounted for water level of 10% is the recommended value. According to data
collected by the World Bank the mean level of unaccounted for water for water utilities in
developed countries is 16%.
Table 7.17 Unaccounted for water for large southern African cities
City Percentage of connections Unaccounted for water

that are metered (%) (%)
Luanda, Angola 40 60
Gaborone, Botswana 100 20
Kinshasa, Democratic 76 47
Republic of the Congo
Maseru, Lesotho 97 32
Port Louis, Mauritius 100 45
Maputo, Mozambique 100 34
Windhoek, Namibia 100 11
Greater Victoria, Seychelles 100 26
Mbabane, Swaziland 100 32
Dar Es Salaam, Tanzania 10 60
Lusaka, Zambia 44 56
Harare, Zimbabwe 85 30
It should be noted that the figures for unaccounted for water in Tables 7.17 are crude. Water
utilities often overstate their performance indicators or manipulate data so that they appear
favourable. In some cases unaccounted for water figures may not be credible when water
production and consumption are not well metered. As Table 7.17 shows there are some cities
in southern Africa were the percentage of connections that are metered is extremely low.
Even taking into account that the figures for unaccounted for water in Table 7.18 may be an
underestimate it is still the case that many water utilities in southern Africa are failing to meet
standards achieved by their peer utilities.
The Global Water Supply and Sanitation Assessment 2000 Report also stated that:
• Many water utilities have difficulty in collecting and maintaining records related to the
performance indicators;
• The quality of the data need to be cross-checked, as many of the water utilities are not
happy with the quality of their own data;
• There is a need to address concerns among utilities that they are providing confidential
information that may be made public;
• There is a slow response to questionnaires and evidence of lack of commitment on the
part of some utilities.
Table 7.18 Unaccounted for water figures for water utilities in southern Africa
Water utility Country Year of Unaccounted

data for water (%)
collection
Water Utilities Corporation Botswana 2000 20.8

Water and Sewerage Authority Lesotho 2000 22.5
Central Water Authority Malawi 2000 43.2
Blantyre Water Board Malawi 2000 37.8
Northern Region Water Board Malawi 2000 28.6
Lilongwe Water Board Malawi 1999 34.3
Central Water Authority Mauritius 2000 43.2
Walvis Bay Municipality Namibia 2001 16.0
City of Windhoek Namibia 2000 19.8
Bloem Water (Bloemfontein) South Africa 2000 50.0
Durban Metro Water Services South Africa 2001 20.0
City of Tygerberg South Africa 2001 18.2
Manguang Local Municipality South Africa 2001 26.6
Drakenstein Municipality South Africa 2000 25.9
Arusha Urban Water Supply and Sewerage Authority Tanzania 2001 38.9
Mwanza Water and Sewerage Authority Tanzania 1999 65.2
Dar Es Salaam Water and Sewerage Authority Tanzania 2000 34.8
Iringa Urban Water Supply and Sewerage Authority Tanzania 2000 45.0
Shinyanga Urban Water and Sewerage Authority Tanzania 2000 44.0
Dodoma Urban Water and Sewerage Authority Tanzania 2000 42.2
Urban Water Supply and Sewerage Authority Tanzania 2000 30.0
Tabora Urban Water and Sewerage Authority Tanzania 2000 36.9
Department of Water Development Tanzania 2000 30.0
Tanga Urban Water Supply and Sewerage Authority Tanzania 2000 34.3
Chipata Water and Sewerage Company Zambia 1999 40.1
Mulonga Water and Sewerage Company Zambia 2000 37.0
Lusaka Water and Sewerage Company Zambia 2000 50.0
AHC Mining Municipal Services Limited Zambia 2001 59.5
Kafubu Water and Sewerage Company Zambia 2001 50.0
City of Mutare Zimbabwe 2000 57.6
City of Gweru Zimbabwe 1999 35.0

7.5.2 Methods for reducing unaccounted for water
In order to reduce the level of unaccounted for water it is important that the factors causing
the unaccounted for water are well understood. Estimation and detection of the loss in a
water supply system can be applied only if the following basic requirements are in place:
• A thorough understanding of the structure and functioning of the system;

• Availability of adequate measuring devices for flow volumes/flow rates;
• Availability of engineering drawings, instruction and operation manuals.
Metering of water volumes supplied is an essential part of a demand orientated management

strategy and any programme to reduce the unaccounted for water and increase the financial
revenue should be based on it. Some publications on demand management strategies stress
the need for financial and institutional changes as water pricing, institutional reforms and
policy changes. However, it should be well understood, that good intentions and policy
changes towards reduction of unaccounted for water could not succeed if the above-
mentioned requirements are not in place. Methods by which unaccounted for water can be
reduced are shown in Table 7.19.
Table 7.19 Methods by which unaccounted for water can be reduced
Area Issues Actions
Metering Unmetered connections Meter installation

Faulty meters Meter replacement/repair
Under registration of meters Bulk metering
Lack of confidence on billings Bulk metering
Lack of confidence in number of Bulk metering
customers
Leakage Leakage in reservoirs and mains Systematic maintenance, detection,
Poor quality pipe material and monitoring and maintenance of old
installation pipes
Lack of information on pipe network Information programmes to public
and others
Standardisation of installation,
material and control
Lack of maintenance Pipe database
Replacement connection policy
Adequate pressure regulation
Operational control Deficient operational control Monitoring indicators
Water distribution system automation
Designing operations control units
Commercial systems Inefficient billing system Database of users
Poor connection and/or Design/implementation of better
disconnection procedures commercial systems
High level of accounts receivable
Low income consumers not billed Improved users/demand data
Illegal/unregistered connections Disconnect policies
Water pricing policies Control of high volume users
The major sources of unaccounted for water in the majority of African cities are leakage and
illegal abstractions. The factors that affect the proportion of unaccounted for water that
occurs through leakage are:

• Pressure. This has the following effects:
- The higher the pressures the higher the rate of leakage through broken pipes or
faulty fittings. An example of this is given below.
- The higher the pressure the greater the frequency of bursts in pipes.
- Pressure surges can cause pipes to fracture and fittings to move out of place.
- Cycling the pressure between high and low value within the design pressure
causes fatigue.
• Movement of the soil. This can cause pipes to break, joints to move, or may result in
local stress concentrations within pipe or fitting which eventually leads to its failure.
• Soil characteristics. Soil characteristics can affect the running time of individual leaks.
For example, in some soils water from underground leaks may show on the surface fairly
quickly whilst in others leaks can run indefinitely without showing.
• Deterioration of water mains and pipes. The common forms of failure of pipes owing
to deterioration of their quality are hole formation and transverse or longitudinal fracture
of the pipe. Concrete or asbestos cement pipes can also corrode owing to the presence of
high levels of sulphates in the water or soil. Soft water also leaches calcite from asbestos-
cement pipes weakening them and introducing internal stresses that lead to failure
through longitudinal splitting. Poor quality pipes, fittings and workmanship can promote
leakage. Age also comes into play as older pipelines tend to deteriorate more than
pipelines that have just been laid.
7.5.3 Estimation of water losses in an urban water supply system
The estimation and detection of the loss in a water supply system can be applied only if the
following basic requirements are in place:
• A thorough understanding of the structure and functioning of the system;

• Availability of adequate measuring devices for flow volumes/ flow rates;
• Availability of engineering drawings, instruction and operation manuals and/or a
Geographical Information System (GIS).
Metering of water volumes (rates) supplied is an essential part of a demand orientated

management strategy and any programme to reduce the unaccounted for water and increase
the financial revenue should be based on it. The configuration of a water supply system may
vary in wide limits, depending on water source, topography and the engineering solutions
adopted. The discussion on this topic is based on a generalised scheme, shown on Figure 7.5
that assumes that water is abstracted from a surface impoundment.

Water source
M1
Water treatment works
M2
Service reservoirs
M3
Reticulation system
M4A M4B M4N

District DA District DB District DN
Sub Sub Sub Sub Sub Sub Sub Sub Sub

-district DA1 -district DA2 -district DAn -district DB1 -district DB2 -district DBn -district DN1 -district DN2 -district DNn
Consumers M5
Note: M1 to M5 represent points at which water flow or volume measurement should be metered
Figure 7.5 General scheme of a water supply system
7.5.4 Location of water losses in an urban water supply scheme
The main water losses in a water reticulation scheme are:
• Water losses along trunk mains;

• Water losses within the treatment plant;
• Water losses within the distribution system;
• Water losses within the consumers’ premises.
Water losses along trunk mains. These losses are due to leakage from trunk mains,
including the main from the source to the treatment plant, from the plant to the reservoir(s)
and from the reservoir to the junction within the reticulation system. Owing to the bulk
amounts transported, and in most cases high pressures, the volumes of water losses might be
considerable, but they are relatively easy to detect by visual inspection and lower in numbers.
Trunk mains are usually designed and constructed at high standards of reliability, because of
their importance, and leakage along them is an indication of ageing and the need of
replacement. In cases where they are very long and also complex hydraulic structures are
involved, leaks can reach a significant value, amounting to 60% of that from trunk
distribution mains, expressed as m3/km/day.
Water losses within the treatment plant. These are losses due to the treatment process
itself, their amount can be decreased by improving the operation of the plant or introducing
treatment methods for reclamation of the sludge generated. To estimate these losses and to
include them into the water balance of the system water flow measurements are imperative.
The quantity of water entering the treatment plant (point M1 on Figure 7.5) and leaving the
treatment plant (M2 on Figure 7.5) should be measured.
Water losses within the distribution system. These losses are usually smaller in volumes
but numerous in numbers. Their identification and location is difficult, because of the large
area covered by the reticulation system. Measures for leakage control and detection are
usually dedicated to the reduction of this type of water losses. This need is also supported by
the fact, that water supply projects executed in conditions of financial restrictions are
resulting in less reliable solutions, correspondingly the expected number of breakdowns is
higher, especially in high pressure zones.
Water losses within consumer’s premises. These might be relatively small in volume but
their large numbers can lead to a considerable percentage of the total water loss of the system.
If a measurement device (M5) is installed within each household, the amount of water
consumed, including losses will be recorded. Therefore, the losses within consumer
premises, usually associated with leaks from plumbing and fixtures, and illegal irrigation of
loans and home gardens, are difficult to measure and estimate. It is interesting to note that in
the case of Johannesburg that has unaccounted for water of 42% it has been estimated that the
majority of this results from leakage on private property (Reference 7.26).
Tariff regulations are aiming at the reduction of this type of loss through financial
punishment. However, it should be noted that all other types of leaks, described above lead to
financial loss of the municipality (managing institution), but the water lost in consumer
premises leads to a financial revenue, provided that it is correctly metered and billed
(Reference 7.13).
Provided that water consumed is paid on the basis of metered volumes, water losses within
the system, but excluding consumer premises, form part of the unaccounted for water. Under
this context, it is the responsibility of local authorities to reduce as much as possible of the
water loss within the system, excluding consumer’s premises, while consumers are
responsible for the reduction of water loss within their premises.
7.5.5 The level of metering required within an urban water supply

reticulation scheme
A major objective of the management of water supply systems is to account for water
volumes within the system. Thus monitoring of the system’s water balance is very important
in order to estimate demand trends, expenditure, as well as, to account for water losses. This
can be done only on the basis of adequate number and adequate accuracy of water discharge
measurement devices, located at characteristic points in the system. Preferably, these points
should be identified during the design stage, but experience gained during the exploitation
period can indicate the need of additional measurement points. Regular data collection,
recording and leakage history is essential for good management of the scheme. The various
ways in which water can be metered effectively in an urban reticulation scheme are discussed
below.
District metering
In district metering separately defined areas, typically containing 2000 to 5000 properties, are
metered continuously, and the total quantity of water entering the district is recorded. The
meters are read regularly and if supply is inexplicably high, inspectors are sent into that
district to locate leaks. Metering at a district level is shown as M4 on Figure 7.5.

This method has the advantage that water utility personnel can concentrate their efforts in
those districts where the highest levels of leakage occur. It also has the added advantage that
information regarding flows and use of water within the network is obtained that can be
useful for the day-to-day running of the network and for the planning and design of future
extensions. However, the method is not sensitive to changes in leakage and does not
determine the position of leaks. Although district metering allows the detection and
estimation of water losses within the district it is highly dependent on the range of
measurement, accuracy and sensitivity of the water meter (Reference 7.13).
Waste metering
The urban water distribution system are often sub divided into sub-districts containing 200 to
3000 properties. These are shown as Da1 to Dan on Figure 7.5. These areas are isolated and
fed through a single meter, capable of measuring and recording the low rate of flow that occur
during the early hours of the morning, also known as minimum night flows. It is assumed that
the night flows closely represent the water loss. This is because little legitimate consumption
is expected to take place during the period of measurement. The flows are recorded at regular
intervals. This allows water utility personnel to establish the districts with higher flow
records, indicating leakage (Reference 7.13).
The flow meter, which is used, is one that is capable of measuring low rates of flow and is
normally, referred to as a waste meter. The waste meter may be permanently installed on a
by-pass or carried on a mobile trailer and connected temporarily into the system via hydrants.
There is a type of waste metering technique that is referred to as step testing or valve
inspection. This technique involves closing the valves within the district, so that the metered
area is successively reduced. The resultant reduction in flow rate following the closure of a
particular valve indicates the total leakage plus legitimate night consumption in that section of
the distribution system. There are a number of ways in which the step test can be undertaken.
These include:
• Isolation method. Starting furthest from the waste meter, valves are successively closed
so that the different subdistricts are isolated sequentially and less and less of the district is
supplied via the meter. The sequence of closing valves is followed right up to the last
valve whereupon the flow should drop to zero.
• Close and open method. This entails closing valves on each step, noting the resultant
drop on the flow meter and then re-opening again.
The success of both waste metering and step testing depends to a large extent upon the ability
to isolate completely the waste meter district from the rest of the system and this depends on
valves shutting down tight. The method is sensitive to small leaks and also establishes the
position of that leak between valves within the sub-district. However, this leads to time being
spent in monitoring districts where no leakage has occurred and hence no benefits would be
obtained.
Waste metering is more accurate than district metering and locates water losses within sub-
districts. However, it requires the application of waste meters, with higher level of accuracy,
in addition to the already installed district meters. These could be installed permanently in
identified sub-districts where the incidents of leaks occurring is high or could be installed
temporarily for specified periods to identify zones of high leaks.
Combined district and waste metering
This method consists of both district metering and waste metering. When increases in supply
are indicated on the district meter, the waste meters downstream of it are read in order to sub-
divide the district into more manageable units and therefore guide inspectors to the areas
containing most leaks. The need to use waste meters in addition to district meters might be
overcome if the range of measurement of the district meter is chosen so that it can detect the
night flows in the sub-districts. A combination of two meters, one of them installed on a
bypass, is also a possible solution
7.5.6 Methods for locating leaks
There are three main methods for locating leaks. These are:
• Visual inspection;
• Regular soundings (proactive leak detection);
• Pressure control;
• Visual inspection.
Visual inspection
In this method only those leaks that become self-evident are located and repaired. A leak may
be self-evident because water shows on the surface or may become so upon investigation
following consumer complaints such as poor pressure or noise in the plumbing system.
Leakage is identified by visual inspection or based on consumer complaints. This method is
widely applied and requires regular inspection by the managing authority. No special
professional skills are needed. It is a low cost measure and is effective in areas where
pipelines are laid at lower depth and soil conditions are such that leaks quickly come to the
surface (Reference 7.13).
Regular sounding (proactive leak detection)
This method involves teams of inspectors seeking to locate leaks by systematic direct
sounding on all stopcocks, hydrants and valves through the distribution system and listening
for the characteristic noise of leaking water.
As water under pressure exits a crack or a small hole, the pipe wall and the surrounding soil
emit sound waves in the audible range. Water impacting the soil and circulating in a cavity
creates lower frequency waves that have limited transmission through the ground. Through
the use of surface microphones, leaks can be located with greater precision. The leak noise
detected will depend upon the position at which a sounding is made.
Proactive leak detection requires special sounding equipment, specifically trained personnel
and considerable practical experience in order to be applied successfully. Regular proactive
leak detection could be recommended to areas where leaks could cause severe damage of
surrounding structures or along parts of the system of highest importance (Reference 7.13).
Pressure control
Pressure control does not directly involve leakage detection, but sudden drops in pressure
may indicate to a possible leak. In general, reduction in pressure leads to reduced rate of
escape through each leak and may also affect the number of leaks occurring. Pressure
reduction is relatively cheap and can be quickly effected, but lower pressure may also
increase the leak population by making them less detectable. Pressure reduction can be
achieved in a number of ways such as reducing pumping heads, installing break pressure
tanks and using pressure-reducing valves. The control of pressure surges and cycling is likely
to reduce the numbers of bursts and leaks that occur, especially in plastic pipes. Pressure
control is a necessary tool for the technical management of the system and combined with any
other method of water loss estimation could give very useful information in order to identify

the causes of water lost through leakage (Reference 7.13). Pressure control does not directly
involve leakage detection, but sudden drops in pressure may indicate to a possible leak. In
general, reduction in pressure leads to reduced rate of
7.6 Water demand management measures

The provision of adequate water supply and sanitation to rapidly growing urban populations is
increasingly becoming a problem for governments in the SADC region. The continuing
expansion of the numbers of people in cities who need water and sanitation services and who
cannot readily get these services by self provision, form a continuous pressure to either invest
in additional production capacity or to stretch the available supplies to serve more people. At
the same time, industrial activity also demands the expansion of urban water supply services.
The predominant approach towards meeting these increasing water demands has been towards
supply augmentation schemes. However, the cost of developing new sources or expanding
existing sources is getting higher and higher as the most accessible water resources have
already been tapped. The real costs of water per cubic metre in second and third generation
projects in some cities have doubled between a first and the second project and then doubled
again between the second and third. At the same time, governments are becoming reluctant to
pay the rising investment costs as long as utilities are unable to meet these cost from user
charges.
It has been demonstrated in many countries that saving water rather than the development of
new sources is often the best ‘next’ source of water, both from an economic and from an
environmental point of view. Water demand management therefore is seen as the preferred
alternative to meet increasing water demand and can be defined as a strategy to improve
efficiency and sustainable use of water resources taking into account economic, social and
environmental considerations.
The main objective of water demand management is to contribute to more efficient and
equitable provision of water and sanitation services and to reach this objective a number of
instruments have been developed. These instruments are interdependent and mutually
reinforcing and the most optimal way they are applied will depend on the prevailing local
conditions and are the topic of a number of presentations in this symposium. With regards to
the domestic consumer, water demand management measures can be divided into:
• Water conservation measures including:

- Leakage detection;
- Reduction of illegal connections;
- In-house retrofitting;
- Out-of-house water saving measures;
• Water pricing measures including:
- Water metering;
- Tariff structures;
• Information and educational measures including:
- Awareness raising;
- Public involvement;
- In-school education;
• Legal measures including:
- Rules and regulations that form the basis of water demand management policy;
- Regulations on resale of water.
This section focuses on the most main and most promising water demand management
measures including:

• Reduction of the pressure in a water distribution system;
• Physical measures to reduce water consumption such as reducing toilet cistern capacity;
• The use of tariff structures.
7.6.1 Pressure reduction as a demand management measure
Pressure reduction in a water distribution system can be one of the simplest methods to reduce
water demand. The control of pressures can save water in a number of ways. High pressures
increase losses of water through leaks, and increase use when the amount of water used is
based on time rather than the volume of water discharged. The leakage from water
distribution systems has been shown to be directly proportional to the square root of the
distribution system pressure as indicated by the relationship below.
Leakage α (distribution system pressure)0.5
The objective of any pressure control strategy should be to minimise excessive pressure as far
as possible, while ensuring that sufficient pressures are maintained throughout the network to
make sure that consumer demands are satisfied at all times. The idealised objective of such a
strategy would be to always maintain a head profile in the network such that the pressure at
each connection is just sufficient to provide the corresponding demand. This is referred to as
an optimal head profile or the target pressure level. However, owing to the head-flow
relationships in the network, target pressure levels can only be achieved by few points of the
system while in the others the operational pressure remains higher. As the complexity of a
distribution system grows, the task of achieving the target pressure level becomes more
difficult and the average overpressure tends to increase.
Historically many water utilities have not managed pressures to optimum limits to reduce
leakage for a number of reasons including:
• Lack of awareness or understanding of the leakage problem;

• Lack of awareness or understanding of the pressure leakage relationship;
• Concerns over fire fighting regulations;
• Concern of customer complaints;
• Concerns of storage not filling properly at night;
• Lack of financial incentive if water was cheap and plentiful.
However, pressure management is becoming recognised as an effective tool for reduction of

real losses (in most cases leakage) and also in the reduction of unwanted demand. The South
African Code of Practice SABS 0306 “Code of Practice for the Management of Potable Water
in Distribution Systems” recommends that for water distribution networks that the static
pressure should be limited to between 30 m and 60 m. An example of the effect of pressure
reduction on leakage for the city of Bulawayo in Zimbabwe is given below.

Example of the effect of pressure reduction on leakage for Bulawayo in Zimbabwe
Example for a low density suburb

Queens Park East is a low density (high income), residential area in the city of Bulawayo that was
included as part of a pilot water conservation study in 2001. The total population of the suburb was
calculated to be 3,101 and the number of connections was 460. The total length of the water
distribution system is 12,640 m and the estimated length of connection pipes 9,500 m.
3
Total water consumption during a 24 hour period was recorded as 614 m /day. The minimum night
flow was recorded at 4.4 l/s. Minimum night flow consists of legal night consumption and leakage.
The legal night consumption was estimated to be 0.8 l/s. The losses in the area at night is therefore
4.4 l/s - 0.8 l/s = 3.6 l/s. Adjusted for pressure changes during the daytime this corresponds to a per
connection loss of 580 l/day.
Pressure was reduced from an average of 4.3 bar to 3.7 bar, resulting in an average consumption
reduction of 0.7 l/s. The flow is reduced by a factor of 1.09. This clearly verifies pressure reduction
as an effective method for reduction of losses.
Example for a high density suburb
Emganwini is a high-density (low cost) housing development in the city of Bulawayo. In December
1999 a study revealed that the pressure in Emganwini’s water distribution system was up to 10.6
bar at night time with an average of 9.5 bar. This is an extremely high pressure for a water
distribution system. The minimum night flow was recorded at 7.9 l/s. The legal night consumption
was calculated to be 1.5 l/s. Adjusting for the legal night consumption the leakage at night was
3
estimated to be 6.4 l/s, this is equivalent to a daily leakage of 495 m /day or 215 l/day/connection.
Using 1.5 l/s/connection as an average waste/loss figure at private premises, adjusted for the
3
existing pressure the losses in the distribution system can be calculated to be 338 m /day.
In April 2000 a new pressure reducing valve was installed for Emganwini and the pressure was
3
reduced to 5.5 bar. A total consumption of 916 m /day and a minimum night flow of 1.7 l/s was
recorded. Adjusting for the legal night time consumption, the leakage in the area at night was only
0.5 l/s. The reduction in the leakage resulting from reducing the pressure was considerable. Part of
this may have resulted from repairs carried out to the distribution system between December 1999
and April 2000. As the pressure is constant the net night flow can be converted into a daily leakage
3
of 43 m /day, corresponding to a per connection loss of only 38 l/day.
Pressure was reduced further to an average of 4.3 bar. The corresponding minimum night flow was
measured at only 1.12 l/s. This is almost identical to the calculated legitimate night consumption.
This example clearly illustrates the effect that reductions in pressure can have on water saving.
7.6.2 Physical measures to reduce water demand
At a domestic level there are many physical measures that can be introduced to reduce water
demand. These measures include:
• Reducing the cistern capacity of toilets;

• Use of low pressure shower heads;
• Garden timer taps to limit the time spent irrigating domestic gardens;.
• Flow restrictors at kitchen tap.
It has been shown that retrofitting plumbing fittings, such as installing low volume water
closets and low volume shower roses, reduce overall water use by at least 25% of domestic
water consumption (Reference 7.27). One immediate and cheap measure that can be
implemented is to reduce the cistern capacity of toilets. It has been estimated that water used
for flushing constitutes about 30% of total domestic water use (Reference 7.28). Adjusting
floats in existing installations, or simply putting one or two standard bricks in the cistern
would reduce cistern capacity by 10% or more. This means that each household would
reduce its consumption by approximately 3%, without requiring any significant investment,
thus saving money through a reduced water bill, while not compromising the quality of the
service enjoyed.
In New York 1.33 million inefficient toilets were replaced by efficient ones during 1994 to
1997, reducing the city's consumption by 0.3 million m3/day. Other demand management
measures were also implemented. As a result, per capita water use dropped from 738 l/day in
1991 to 640 l/day in 1999 (Reference 7.27).
7.6.3 The effect of tariff levels on water demand
With ordinary economic goods there is a relation between price and demand following a
demand curve. The dimensionless slope of this demand curve is called the price elasticity of
demand. It is defined as the percentage of increase in demand resulting from a percentage of
increase in price. This elasticity is a negative number since demand is expected to decrease as
price increases, and normally ranges between -1 and 0. The problem is that the elasticity is
not a constant. It depends on the price, it depends on the water use and it varies over time. So
it is an equation with limited applicability.
Primary uses of water have a special characteristic in that the elasticity (E) becomes rigid
(inelastic; E close to zero) when we approach the more essential needs of the user. This is
illustrated in Figure 7.6. People need water, whatever the price and for the most essential use
of water (drinking) few alternative sources of water are available. For sectors such as
industry and agriculture demand for water is generally more elastic (i.e. E closer to -1) which
is more in agreement with the general economic theory. This is because alternatives for water
use exist in these sectors (e.g. introducing water saving production technologies, shifting to
less water demanding products/crops). For basic needs, however, demand is relatively
inelastic or rigid.
increasingly vital uses

drinking
0 elasticity -1
cooking
sanitation
laundry
other uses
Figure 7.6 Diagram of different uses of domestic water and their

elasticities of demand

In urban water supply, elasticities are therefore generally close to 0, unless additional (non-
financial) measures are taken. Poor consumers often only can afford to use small amounts of
water (the basics), and any increase in tariffs will have little effect because they cannot do
with less water. For large consumers (e.g. the ones that irrigate their gardens and own cars
that need to be washed) the ability to pay is such that the need to save money on water is
limited. In the latter case, awareness campaigns, regulation, policing, leak detection and
renewal of appliances are often more effective than the price mechanism per se.
Figure 7.7 shows the different patterns of water use during the year for high-density (poor)
and low-density (rich) consumers for Ruwa, a town in Zimbabwe. The figure shows that for
high-density consumers seasonal fluctuation of water demand is low, indicating that most
uses are confined to essential purposes. Water consumption of low-density consumers is
much higher during the dry season than during the rainy season, probably related to non-vital
uses of water, such as watering gardens and filling pools when it is dry. One could
hypothesise that part of the additional water use during the dry season for these consumers is
relatively elastic, and could be influenced by tariffs, whereas this may not be the case for the
other types of water uses. Similar findings to these were found for the city of Masvingo in
Zimbabwe. This example is given below.
90
Coefficient of variation = 39%
80
Mean water consumption (m 3/month/connection)
70
60
50
40
30
20
10
Coefficient of variation = 20%
0
Sep-96 Mar-97 Aug-97 Feb-98 Aug-98 Feb-99 Aug-99 Feb-00 Aug-00
Date
High density (38 households) Low density (17 households)
Figure 7.7 Water consumption for high and low density suburbs for the
town of Ruwa in Zimbabwe

Water use by affluent and less affluent households in Masvingo, Zimbabwe
Trends in domestic water use in the city of Masvingo, Zimbabwe, were studied by
considering affluent and non-affluent residential areas. The residential areas of Rhodene
and Clipsham, comprising 1,050 households were considered affluent. The residential
areas of Rujeko and Mucheke, represented by a sample size of 3,350 households, were
considered less affluent. The total sample size represents some 34% of all domestic users
of Masvingo. Monthly water consumption was derived from billing data of 1999 to 2001.
The results of average consumption patterns in different residential areas are presented in
the Figure 7.8.
100
Affluent suburbs Rhodene and Clipsham
3
Mean water consumption = 59.3 m /month/connection
90 Coefficient of variation = 31%
Mean water consumption (m3/month/connection)
80
70
60
50
40
30
20
10 Less affluent suburbs Rujeko and Mucheke

3
Mean water consumption = 19.9 m /month/connection
Coefficient of variation = 11.5%
0
Dec-98 Mar-99 Jun-99 Sep-99 Dec-99 Mar-00 Jun-00 Sep-00 Dec-00 Mar-01 Jun-01 Sep-01
Date
Figure 7.8 Monthly billed water consumption for affluent and less
affluent households in Masvingo
Figure 7.8 shows that there is a large difference in water consumption between affluent
3
households that consume 59 m /month on average and less affluent households that
3
consume on average 20 m /month. Water consumption fluctuates much more in affluent
households that have coefficient of variation CV of 31% than that of less affluent
households that have a CV of 11.5%. This fluctuation is related to rainfall, as water use
tends to be higher in the hot dry months, especially for non-essential purposes such as the
use of treated water for watering gardens. In the hot dry month of October, for instance,
3
affluent households may consume as much as 80 m /month or more, whereas the less
3
affluent counterparts consume at most 25 m /month i.e. less than a third. In the poorest
section of the city (500 households within Mucheke residential area) average household
3
consumption was only 12 m /month. This amount may therefore be considered the basic
minimum or "lifeline" quantity, and is, with an average household size of 8 people equivalent
to 50 l/person/day, the minimum quantity recommended by the World Health Organisation.
The explanation for the observed trend in Figure 7.8 is clear: poor households cannot afford
to use a lot of water because of their inability to pay. In addition, they have relatively small
2 2
plot sizes (200 m to 300 m ). This puts an upper limit to the use of water for gardening
even if they did have the ability to pay. As a result, the seasonal variation in their water use
is relatively small, since water is mainly used for the most essential purposes. For the
affluent household the opposite is true: their ability and willingness to pay is large, and
2
water use is seemingly restricted by the size of their gardens (4,000 m on average), the
presence of a swimming pool as well as the number of cars they wish to wash. A large part
of water is thus applied to uses that are considered non-essential. It may therefore be
hypothesised that the CV of monthly consumption data is an indicator for the elasticity of
demand. This hypothesis would suggest that the elasticity of demand for water in affluent
households in Masvingo may be 2.7 times higher than that for poor households
(31/11.5=2.7). Detailed economic research would be required to falsify or accept this
hypothesis. This preliminary finding suggests that water pricing influences water use by
affluent households to a much larger extent than that of non-affluent households.
With regards to the effect of tariff levels on water consumption the following can be
concluded:
• The elasticity of water consumption is generally low;

• The price elasticity is greater when the price is higher;
• In the household sector, the price elasticity varies between -0.15 and -0.70;
• With respect to drinking water the demand-price relation will never be elastic (E < -1);
• In the industrial sector, the majority of estimates are in the range of -0.45 to -1.37
Increasing block tariff system
It should further be noted that any pricing policy aimed at influencing demand should
consider the basic right of people to access of safe drinking water. Thus demand management
through economic means should consider financial (full cost recovery) and equity criteria.
The increasing block tariff pricing structure implies a cross-subsidy from rich to poor users.
It is a good example of a satisfactory compromise between both criteria and is becoming
increasingly adopted, especially in water scarce regions.
The increasing block tariff system charges the water for the most vital human needs such as
drinking at the lowest levels and the uses for less vital needs highest. This tariff system
therefore makes a mockery of conventional economic theory, which would price the most
valued uses highest.
In order to find a satisfying compromise between full cost recovery and equity, each block
should have a clearly defined purpose, from which block size and tariff can be derived.
Below is an example of how the functions of four blocks could be defined:
(1) The poorest households have access to a lifeline amount of water and do not spend
more than a certain percentage of their income on water.
(2) The ‘ideal’ per capita water consumption level is defined, which will ensure “well-
being”. This “well-being” amount is, for example twice the lifeline amount. All
water consumed over and above the lifeline amount, but less than the well-being
amount, is charged at the Full Cost of Water Supply (FCWS expressed in e.g.
US$/m3); meaning that the average price of water is still less than FCWS, so these
households still receive a subsidy.
(3) Those households that use water over and above the well-being amount, but less than
a certain upper limit (e.g. four times the lifeline amount) will pay the full cost of
water over their entire use. This means that the tariff of the third block should off-set
the implicit subsidy that these users receive in the first block;
(4) Water use over and above the amount specified in the third block will be charged at a
rate that will off-set the subsidy received by households falling within blocks 1 and 2.
The above functions of the tariff blocks would ensure full cost recovery and equity.

1.8
1.6
1.4
1.2
Tariff (US$/m3)
0.8
0.6
0.4
0.2
0
0 10 20 30 40 50 60 70 80
3
Monthly water consumption per connection (m /month)
Hermanus (South Africa) Windhoek (Namibia) Gaborone (Botswana)
Figure 7.9 Block tariff structures for Windhoek, Gaborone and Hermanus
Case study of Masvingo’s tariff structure
The city of Masvingo in Zimbabwe derives its largest income from the water account, which
contributed between 25% to 40% of the council’s revenue since 1995. The City Council
made, on average, about 80% profit on each cubic metre of water sold during 1999 to 2001.
The water account is able to meet the present operation and maintenance needs of the water
supply utility.
Since 1999 Masvingo has adopted an increasing block tariff, with the first block covering
consumption up to 18 m3/month per connection and the second anything in excess, in
conjunction with a fixed monthly charge. The fixed charge is differentiated between
households in affluent neighbourhoods and those in other areas. Owing to the unstable
economic climate, since 2001, the tariffs are reviewed every six months. Table 7.20 shows
that the tariffs were increased significantly during 2001 and 2002. Table 7.21 shows that low
water consumers pay a relatively high price per cubic metre of water. This is due to the
relatively high fixed charge, which, for a household consuming only 12 m3/month, contributes
43% to its water bill. The largest water users, the affluent, pay on average about the same
unit water price of water as the non-affluent users.

Table 7.20 Domestic water tariffs for Masvingo 2001 and 2002
Year Fixed charge Consumption
Poor Rich Less than or equal More than

3 3
to 18 m /month 18 m /month
3
(Zim$/connection/month) (Zim$/m )
January 2001 80 287.30 8.98 12.82

July 2001 100 359.13 11.23 16.03
January 2002 155 556.45 17.41 24.85
July 2002 186 667.74 20.89 29.80
Table 7.21 Average water price at different consumption levels in Masvingo

January 2002
Household Monthly Total water bill Average water Fixed

3
consumption (Zim$/month) price (Zim$/m ) charge (%)
3
(m /connection
/month)
Less affluent 12 364 30 43

Affluent 60 1,914 32 29
Affluent 120 3,405 28 16
Note: In January 2002, 1 US$ was equivalent to Zim$ 55 (official rate) and Zim$ 300 (parallel market
rate)
Income levels of the majority of residents in Masvingo are not high. An oral opinion survey
was carried out which found that Zim$3,500/month was the average cash income per
household for the low-income bracket (the minimum set by government is Z$ 8,900 per
month). Using tariffs applicable from January 2002, and 12 m3/househehold/month as the
average water consumption per household in the poorest neighbourhoods, a household would
pay Zim$364 per month, or 10% of their estimated cash income. World Bank studies have
recommended that not more than 5% of income should be spent on water for basic
requirements. The high-income consumers, however, spend less than 5% of their income on
water. Assuming an average income of an affluent household of Zim$60,000 per month, and
average water consumption of 60 m3/month, its monthly water bill would amount to
Zim$1,914, or only 3% of household income.
It therefore appears that Masvingo's water tariffs could be improved: the fixed charge must go
down since it compromises equity; whereas the tariff of the second block should be increased
relative to the first block, since that would give a clear signal to high water users to reduce
their consumption. Such a change in the tariff structure is likely to have little effect on water
consumption by poor households, since these have a relative inelastic demand for water.
However, the tariff change would influence consumption levels of the affluent households,
which have a much higher elasticity of demand.
In 2002 a study carried out by the University of Zimbabwe proposed an alternative tariff
structure. This structure abolishes the fixed charge and introduces a four-stepped rising block

tariff structure that is based on the same principles developed by for the city of Windhoek.
The first block considers the lifeline amount (12 m3/household/month), and its tariff
recognises the limited ability to pay by the poor, assuming that the low income bracket may
not spend more than 3.5 % of cash income on water. The tariff for the second block, the well
being quantity (24 m3/household/month), is set at the real cost of water supply. The third
block has a tariff that meets the full cost of water supply. The tariff of the highest block
compensates for the subsidies enjoyed by the lowest two blocks, and caters for the financial
requirements of future extensions of the water supply utility that is shown in Table 7.22.
Table 7.22 An alternative tarrif structure for domestic water users for Masvingo
January 2002
3 3
Block (m /connection/month) Tarrif (Zim$/m )
0 to 12 10
13 to 24 20
25 to 36 35
>37 55
The alternative tariff would result in the following:
• Yields the same revenue as present (including an estimated 50% of income over and
above operation and maintenance costs, meant to cover infrastructure development);
• Considers the poor’s limited ability to pay in that the water bills of the poorest households
would be reduced by over a third.
Such a block tariff structure can assist with water demand management in that as
consumption increases the average charge per unit of water also increases.
7.7 Urban water demand case study for the city of Windhoek
in Namibia
In 1995 in the city of Windhoek, water demand was 242 litre per day per person in 1995, with
unaccounted for water being only 11%. Windhoek adopted an integrated policy on water
demand management in 1994, which is financed by a 0.5% levy. Efforts that started in the
1950s have primarily focused on re-use of water. Currently Windhoek can re-use all its
wastewater for the watering of parks, sport fields and cemeteries through a two-pipe system
and the reclamation of wastewater to a potable standard. Of all domestic water use, 13% is
treated for reuse. About 60% of all water used in the more affluent households is for gardens.
The infiltration of this water into lawns and gardens makes it unavailable for reuse. Water for
gardening still represents a large sector for water savings (Reference 7.9).
An important part of the water demand management programme involves appropriate tariffs.
When tariffs are sufficiently high, they tend to keep exterior irrigation demands to within
reasonable levels. Water tariffs were recently raised by 30% and any water demand
exceeding 45 m3 per month per household or enterprise was billed at US$ 1.30 per m3.
Other water demand measures that have been implemented include:
• Public awareness and education;

• A ban on the irrigation of gardens between the hours of 10:00 and 16:00;
• The mandatory use of swimming pool covers;

• Use of low-flush toilets. This has been a mandatory requirement for all new buildings
since the beginning of 1997;
• The use of low flow showers (i.e. showers with a flow of less than 10 litres/minute);
• On-site reuse of grey water
• Metering of all connections;
• Reuse of purified effluent for irrigation and reclamation to potable standard;
• Water conservation guidelines for wet industries.
The combined effect of all these measures was a decrease in the per capita water
consumption. In 1996 per capita water use decreased from 242 litre per day per person to 196
litres per day. Whereas the residential population grew 5%, total residential water
consumption decreased from 10 million m3/year to 7.8 million m3/year.
The benefits from water conservation include:
• Long-term water demand reductions of up to 30% can be achieved using the measures
adopted by Windhoek. It should be noted that larger, short-term reductions of up to 60%
reduction in water demand can be achieved through the implementation of a drought
management programme;
• Less waste water has to be treated, and less energy is used;
• The environment will benefit from reduced alteration of flow patterns and from fewer
dams and other infrastructure having to be constructed;
• Financial savings from reduced capital as well as operating costs. Figure 7.10 illustrates
savings due to delay in construction of “the next dam”.
Normal demand
Water supply/demand (volume)
Increased supply
Original water supply
Water demand with

water conservation
measures implemented
Four year delay

in construction of
infrastructure owing
to water conservation
measures
-8 -4 0 4 8 12 16 20 24
Time (years)
Figure 7.10 The effect of water conservation measures on the construction

of infrastructure

7.8 References
7.1 City of Cape Town (2001) Water services development plan
7.2 Water demand management programme for southern Africa Proceedings of the water
demand management regional workshop
7.3 Global water supply and sanitation report 2000 Water Utility Partnership (2001)
Report of performance indicators African water supply and sanitation utilities 2001
7.4 Baba, A.F. (1996) Water resources, water supply and sanitation in managing water
resources for large cities and towns, Report from the Beijing water conference
conducted by the United Nations Centre for Human Settlements
7.5 Bailey, R.J., Jolly, P.K., Lacey, R.F. (1986) Domestic water use patterns
7.6 Hadley, H.K. (1995) Water use guidelines: Commercial water use, US Geological
Survey Department
7.7 Novak, P., Moffat, A.I.B., Nalluri, C. and Narayanan (1996) Hydraulic structures, E
& F.N. Spon publishers second edition
7.8 Twort, A (1993) Water supply, Arnold publishers fourth edition
7.9 Macy, P. (1999) Urban water demand management in southern Africa: The
conservation potential, Swedish International Development Co-operation Agency
7.10 IUCN (2000) Water demand management: Towards developing effective strategies
for southern Africa IUCN Wetlands and Water Resources Programme
7.11 Mwandosya M.J. and Meena H.E. (1998) Dar Es Salaam water demand: An end-use
perspective The Centre for Energy, Environment, Technology Tanzania
7.12 Norplan SA (2001) Bulawayo water conservation and sector services upgrading
project: Final report summary Unpublished
7.13 UN Centre for Human Settlements (1999) Managing water for African cities:
Developing a strategy for urban water demand management Unpublished
7.14 Hranova R. (2001) Estimation of domestic water loss through leakage in high-density
municipal areas Unpublished
7.15 The World Bank Group Data and Statistics World-wide web page
http://www.worldbank.org/data/countrydata/countrydata.html
7.16 Clarke G. P. Kashti A., McDonald A and Williamson P. (1997) Estimating small area
demand for water: A new methodology Journal of the Chartered Institution of Water
and Environmental Management pp186 to 192
7.17 Tynan N and Kingdom B (2001) A Scorecard for Water Utilities in Developing
Countries World Bank
http://rru.worldbank.org/HotTopics/Hot_Topics_tynan_kingdom.asp

7.18 WHO and UNICEF (2000) Global Water Supply and Sanitation 2000 Report
http://www.who.int/water_sanitation_health/Globassessment/GlobalTOC.htm
7.19 WHO and UNICEF (2000) Global Water Supply and Sanitation 2000 Report Part II
Country Profiles http://www.whoafr.org/wsh/countryprofiles/
7.20 Kjellen M. (2000) Complementary water systems in Dar Es Salaam Tanzania: The
case for water vending Water Resources Development Vol. 16 No 1 143-154
7.21 Japan International Co-operation Agency (1991) The study on the rehabilitation Dar
Es Salaam Water Supply in the United Republic of Tanzania Final Report Vol. 2 of
the main report (unpublished)
7.22 North Carolina Department of Environment and Natural Resources (1998) Water
http://www.p2pays.org/ref/01/00692.pdf
7.23 Dube E. and Van Der Zaag P. (2002) Analysing the water use patterns for water
demand management: The case of the city of Masvingo, Zimbabwe, Third Waternet
Symposium October 2002
7.24 Gumbo B. and van der Zaag P., (2001), Water losses and the political constraints to
demand management: The case of the City of Mutare, Zimbabwe, 2nd
WARFSA/Waternet Symposium, "Intergrated Water Resources Management;
Theory, Practice, Cases", 30-31October 2001, Cape Town, South Africa.
7.25 South African Code of Practice (SABS 0306) Code of Practice for the management
of potable water in distribution systems
7.26 Personal communication with Hannes Buckle, Rand Water (2002)
7.27 Martindale, D. and Gleick, P.H. (2001) How We Can Do It. Scientific American
February 2001
7.28 Gumbo, B. (1998) Dual water supply systems “Is it just another pipe dream”
Proceedings of the International Conference on Small and Medium size domestic
water conservation, wastewater treatment and use. Bethlehem, Palestine 21 to 24
February 1998.
7.29 Work carried out by Chibaya, S. (2001), University of Zimbabwe
7.30 Personal communication with van der Zaag, P. (2002), University of Zimbabwe

8. FORECASTING WATER DEMAND AND USE
8.1 Background
Predicting future water demand and use is increasingly important in sub-catchment and
catchment management. It is generally easier to forecast water demand and use for the short-
term rather than for the long-term. For short-term forecasts it is possible to make assumptions
that some factors will not have changed significantly from the present. However, for long-
term forecasts there is a greater degree of uncertainty as to how the water demand and use
will change. It is important that uncertainties resulting from a lack of information are explicitly
reflected in the demand forecasts. At a catchment level demand forecasting should be carried
out for each of the major demand sectors. These are:
• Environment;
• Urban and rural domestic;
• Industry;
• Agriculture.
There is no absolute level of accuracy that is appropriate in all demand forecasting

circumstances. As a consequence the level of accuracy of the forecast should be sensitive to
its purpose. The cost of improving the demand forecast should be balanced against the
additional benefits that are derived from the greater accuracy. The demand forecast method
that is used will be dependent upon the purpose of the forecast and the cost of assembling the
data. For example a more accurate and costly forecast would be required to justify the
construction of a major new piece of infrastructure (e.g. a dam or a large pumping station)
than for general internal monitoring purposes of a catchment management agency.
8.2 Influences on water demand and use

There are many influences that affect water demand and water use. Some of the most
commonly cited factors relate to:
• Population;
• Level of service;
• Tariff levels;
• Demand management measures and increased efficiency in water use;
• Climatic conditions.
To predict demand accurately all of the above factors should be taken into account. There is a
number of forecasting methodologies that can be used. These are discussed in Section 8.5.
8.3 Criteria for assessing forecasting methods

There are numerous methods available to forecast water demand. To assess alternative
forecasting methods it is necessary to establish criteria by which the methods can be judged.
Suggested criteria are as follows:
• Consistency and transparency of the method;

• Logical/theoretical appeal of the method;
• Method incorporates and explains historical trends;
• The way in which factors that have not been taken into account in the past are treated;
• Empirical validation of the method;
• The acceptance of the method by regulatory bodies and other relevant organisations;
• The cost and feasibility of the method.
These criteria are discussed below.
8.3.1 Consistency and transparency of method
These are basic requirements of any systematic forecasting method. If other organisations
cannot understand or reproduce the results there is no way of them assessing the reliability of
the method. If the method has not been documented systematically it will be difficult in the
future to determine the reason for forecasting errors. Demand forecasts carried out by
organisations should be consistent. This aids revision of forecasts that are made in future
years. Aspects that contribute to consistency include:
• Geographical consistency. This is required at a number of levels ranging from small

areas (e.g. sub-catchment or district, local authority) to a catchment or regional level;
• Population forecasts should be carried out in the same manner at all geographical levels;
• Water demand and consumption figures available should be consistent with previously
published data unless anomalies can be explained.
8.3.2 Logical/theoretical appeal
Even if a particular forecasting method has produced good results it is unlikely to inspire
confidence among practitioners unless it has a logical appeal to practitioners. Forecasting
method without a sound theoretical basis are also unlikely to be acceptable to regulatory
bodies.
8.3.3 Incorporates and explains historical trends
Simple extrapolation of historical trends are often likely to be unsatisfactory for a variety of
reasons. In some cases what has happened in the past with regards to water demand and use
is mostly irrelevant to future forecasts. However, there are lessons that can be learned from
historical trends and if a forecasting method is unsuccessful in explaining past trends this
casts doubts on its usefulness in predicting future demands.
8.3.4 The treatment of factors not taken into account in the past
There may be several factors that have not been taken into account in past water demand
forecasts (e.g. metering, price, regulatory changes, reductions in unaccounted for water,
climatic changes). It is important that these influences are incorporated explicitly into
demand forecasts.
8.3.5 Empirical validation
This addresses how well the forecasting method has performed in the past i.e. how well has
the method predicted actual measured water demand and use.

8.3.6 Acceptance by the regulatory body
This is an important criterion, if a forecasting methodology is not accepted or recognised by

the regulatory body it would be unwise to rely on a method that is not widely recognised or
accepted.
8.3.7 Cost and feasibility
Some methods may have a strong theoretical basis and have performed well in the past.
However, the demands of the method on resources may make it impractical to implement.
These demands may take the form of unrealistic demands on data (e.g. with respect to quality,
availability or financial cost) or excessive demands on the skills and time required to operate
and maintain the method.
8.4 Choice of forecasting method

The are no simple rules that can be used to establish which water demand forecasting method
should be used and the scale of the resources that should be invested in demand forecasting.
In deciding which method to use in predicting future water demands, a balance needs to be
made between:
• The level of accuracy of the forecast required;

• The cost of obtaining the required level of accuracy;
• The benefits accrued from having a higher level of accuracy.
The advantages of producing as accurate forecasts as possible are as follows:
• Reduces the mis-allocation of resources and allows investment decisions to be delayed;

• Enables the effects of water resources policies to be examined with confidence;
• Identifies areas and sectors to be targeted for conservation programmes.
There are some principles that can be applied. These are as follows:
• The accuracy of the forecast should be in relation to the cost of the errors that may result
from an inaccurate forecast e.g. such as construction of a large piece of infrastructure that
is not needed because demand has been overestimated;
• The forecast of a demand component that represents a small proportion of the total
demand can be relatively inaccurate without causing too much concern. For example if
unmetered households only made up 5% of the total urban domestic demand of a city a
forecast of this component that was in error by 50% would not cause a major inaccuracy
in the overall forecast of urban domestic water demand.
• The most important water demand sectors and sectors and components should receive the
most attention. This does not mean that the sector or component with the largest water
demand should have the most resources allocated to it.
The question often arises of whether to use a simple or sophisticated forecasting technique.
This is often viewed as a trade off between cost and accuracy.
The methods used for long-term water demand forecasts (i.e. forecasts with a horizon of more
than ten years) are often unsuitable for forecasts over shorter periods of time (i.e. for the next
two years). In the short term demand is likely to be influenced by the weather or the precise
timing of the economic cycle. Often data available for the explanatory variables required by a
long-term forecasting model are not available at the necessary level of frequency making the

model difficult to apply. Checklists for determining and applying forecasting methods are
given below.
8.4.1 Checklist for determining the forecasting method
Before deciding which forecasting method to implement, the following questions should be
considered:
• What was the previous forecast and are the results available?
• What are the previous forecasts errors and why did they occur?
• What is the objective of the new forecasts? (e.g. long term or short term planning, sub-
catchment or catchment level);
• What data are available and over what period and at what level of detail? (e.g. water
demand data, data on population growth, demand management measures to be
implemented);
• What are the possible forecasting methods available?
• Which of these are feasible given the various constraints on data, budget and skills
available to implement the method?
• Is the method used previously still satisfactory for the purpose? (e.g. is a sophisticated
method still preferable to a simpler method);
• Do additional data need to be collected in the future?
8.4.2 Checklist for applying the chosen forecasting method
Once the forecasting method to be used has been decided upon the following issues need to
be considered:
• What specification is to be used?

• How well would the forecasting method have performed in the recent past? (e.g. can any
large discrepancies be understood);
• What assumptions need to be made for the forecast? (e.g. implementation of water
demand management measures, changes in tariff structures);
• Is there any information available to help form the assumptions? (e.g. are details of future
tariff levels available, are demand management measures to be implemented);
• Is the last year a “normal” year to predict from? (e.g. were water demand figures affected
by water rationing brought about by a drought);
• Are the forecasts plausible? (i.e. how do the forecast changes in demand compare with the
historical changes in demand).
8.5 Incorporating uncertainty into forecasts
Uncertainty is an integral part of water demand and use forecasting. All forecasting methods
for water demand and use are subject to uncertainty. It would be desirable to be able to
present a forecast with statistically based confidence limits around it. However, in practice
most forecasting methods do not yield purely objective confidence limits, and in some cases
the data are inadequate to permit statistical analysis to yield precise estimates. Uncertainty is
therefore better considered through the construction of a various scenarios. Incorporating
scenario modelling into forecasts allows the sensitivity of long-term forecasts to various factors
can be established. For example water demand and use forecasts are often carried out based on
high, medium and low population growth rates or a variety of demand levels, or levels of
reduction of unaccounted for water. A typical sensitivity analysis is shown in Figure 8.1.

There can be expected to be different degrees of uncertainty surrounding different components
of the demand. For example when forecasting urban domestic demand it is likely that
uncertainty over unmeasured household demand would be high. This is due to the limited
information about the impact of charges and prices on the demand.
Water use
High growth population

growth scenario
Medium population
growth scenario
Low population
growth scenario
2000 2010 2020 2030

Year
Figure 8.1 Forecasting water use for three population growth scenarios
8.6 Details of forecasting methods

There are a number of forecasting methods that are commonly used to predict future water
demands and use. These include:
• Judgemental forecasts;
• Extrapolation of historical data;
• Forecasts based on population growth and per capita consumption;
• Trend analysis;
• Component analysis;
• Multiple linear regression analysis;
• Multiple non-linear regression analysis.
These are discussed below.

8.6.1 Judgmental forecasts
Judgmental forecasts are based on personal or group knowledge. They may be purely
subjective or merely an adjustment of a more formal forecast. Judgmental forecasts have a
number of disadvantages in that they are subject to a number of biases that may add to or
filter existing knowledge and thus effect the overall forecast. These biases can be classified
as follows:
• Professional biases – the profession of a person may influence their judgement e.g.
professionals who are used to supply oriented solutions may produce forecasts greatly in
excess of professionals who look at demand management issues. Different professions
may also use different methods and data;
• Spatial and project biases – The forecasts may be based on areas that have common
factors (e.g. accessibility) and these areas may not be representative;
• Person biases – The stakeholders interviewed and assessed to determine demand
forecasts may in some cases overestimate their needs to ensure supplies under extreme
conditions;
• Seasonal and climatic biases – Judgements may have been made during untypical dry or
wet periods.
Although judgmental methods may prove useful in conjunction with other methods it is not
recommended that such techniques are used solely for demand forecasting of future water
use.
8.6.2 Extrapolation of historical data and trend analysis
A wide range of methodologies is used for forecasting water demand and use at a national,
regional and local level. Extrapolative techniques have been used frequently in the past but
there is a need to improve forecasts in line with more comprehensive planning approaches
and changing priorities (such as demand management techniques). Extrapolative techniques
suffer from a number of major drawbacks including:
• Very different predictions can be gained from trends which fit past data equally well;
• The methods assume past trends will continue into the longer-term future;
• Any errors that occur in the forecast do not provide a sound basis for future learning.
This is because the errors occur owing to changes in trends;
• Extrapolation techniques tend to use aggregate demands rather than components of
demand.
Forecasts based on a trend analysis use time as the independent variable and utilise
mathematically fitted functions to historical water demand and use to forecast future demands
and use. The function used to fit the data depends on the user and can have a significant
effect on future predictions. The method has the following advantages:
• It is relatively cheap to implement and takes account of past uses;

• It is generally not very demanding in terms of the data that is required
• It can be related to separate water demand and use sectors if the information is available.
The method has the following disadvantages:
• It assumes the historical water use and demand is representative of future water demand
and use;
• It can produce inaccurate results because the forecast is dependent on the fitting function;
• The method cannot account for the effects of a reduction in unaccounted for water.
An example of a forecast using trend analysis is given below.
Example of forecasts using trend analysis
The Mupfure sub-catchment in Zimbabwe has 20 years of recorded and reliable water use figures
for the catchment. Three different curves have been fitted to the available historical water use data
to forecast the future water use in the Mupfure catchment in Zimbabwe. Three methods have been
used to forecast future demand. These are:
• Exponential fit;
• Second degree polynomial fit;
• Linear regression.
The results of the analysis are shown in Figure 8.2.
250
Second degree polynomial fit Linear regression
2 y = 2.0249x - 3961.3
y = 0.0731x - 288.72x + 285329
2 2
R = 0.995 R = 0.959
200
Exponential fit
-24 0.0293x
y = 3x10 e
2
R = 0.984
Water use (million m3)
150
100
50
Historical water use data
0
1975 1980 1985 1990 1995 2000 2005 2010 2015 2020 2025 2030
Year
Figure 8.2 Results of trend analyses using different fitting techniques
Figure 8.2 illustrates that even though each of the fitting techniques chosen to forecast future water
2
use has a high correlation coefficient (R ) the choice of technique can have a significant effect on
the estimate of the future water use in the year 2025.
Another example of some of the problems that can occur with the extrapolation of historical
data is shown in Figure 8.3. Figure 8.3 shows the total water produced to supply the city of
Masvingo in Zimbabwe. Between 1991 and 1992 there was a serious drought in Zimbabwe
that had a significant impact on water demand for most urban areas. If a forecast of future
water demand had been made in 1991 by fitting an exponential curve to the available data
between 1977 and 1991 the forecast water demand for the year 2001 would have been almost
10 million m3/year. However, in 1992 a series of demand management, economic and water
rationing measures lead to in the rate of growth of water demand decreasing significantly. In
2001 the actual quantity of water produced for the city of Masvingo was some
6.8 million m3/year. The figure forecast for 2001 using a simple exponential curve fitting
technique for the recorded data between 1977 and 2001 is over 47% higher than the figure
actually recorded. This clearly illustrates the dangers of using simple curve fitting techniques
to forecast future water demand.
12
y = 2.475x10-48e0.055932x
r = 0.939
10
Total water produced (million m /year)
3
0
1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001
Year
Recorded water produced pre-1991
Exponential forecast based on fitting to pre-1991 data
Recorded water produced post-1991
Figure 8.3 The dangers of using extrapolation techniques for forecasting

water demand for Masvingo in Zimbabwe
8.6.3 Forecasts based on population growth rate
The estimated demand is derived from projected population growth figures multiplied by an
estimated per capita water use or demand. The per capita figure may be derived either
directly where water use is divided by the population (normally for a year in which a census
has been undertaken) or from international estimates for average water use, minimum
standards or from selective surveys.
The method has the following advantages:
• It is relatively inexpensive to carry out;

• It requires a limited amount of data.
The method has the following weaknesses:
• The accuracy can be poor for long term forecasts;

• In the light of the AIDS/HIV pandemic in southern Africa long term predictions of future
population growth are hard to make accurately;
• The method does not take into account variation in water use between sectors or even
within sectors.
• Past trends of total water demand are projected into the future;

• Estimates of population and per capita consumption are made for future dates without
reference to past trends in water consumption;
This latter technique is the most commonly used because it does not require a large quantity
of data. There are three main disadvantages with this technique. These are:
• It assumes that the factors influencing demand in the past will remain the same in the
future;
• It makes no attempt to understand why water consumption fluctuates over time;
• It relies on data about past water demand that may be poor or non-existent and which may
well under-estimate the actual water requirements of the population.
Data to carry out the forecast can be acquired as follows:
• Primary data sources. Water usage is based on amounts delivered by public supply and
licensing authorities for private abstractions. Population data is taken from national or
regional census.
• Secondary data sources. Water use data is taken from records of municipal authorities
of water delivered, or from surveys or direct measurement. Population data is taken from
local estimates.
• Derived data. Data can be derived from restricted surveys to obtain per capita demand
or use. Population figures can be extrapolated using statistical methods. Water usage
may also be estimated from levels of service.
An example of a forecast based on population growth rates is given below.
Example of a forecast based on the population growth rate
The population of the Mazowe catchment in Zimbabwe was estimated to be 1,671,000 in

1999. The per capita water use was estimated to be 140 litres per head per day. The
population growth rate has been estimated to currently be 3.2% reducing to 2.5% from the
year 2015. An estimate of the annual water use per year up to the year 2025 is required.
3
In 1999 annual water use = 140 x 365 x 1,671,000 = 85.4 million m
The water use is calculations are tabulated below.
Year Water use

3
(million m )
1999 85.4
6
2005 85.4(1+0.032) = 103.2
11
2010 85.4(1+0.032) = 120.8
16
2015 85.4(1+0.032) = 141.4
5
2020 141.4(1+0.025) = 159.9
10
2025 141.4(1+0.025) = 181.0
A graph of the above forecast is shown in Figure 8.4

200
180
160
140
Water use (million m 3)
120
100
80
60
40
20
0
1995 2000 2005 2010 2015 2020 2025
Year
Figure 8.4 Water use forecast based on population growth
8.6.3 Forecasts based on component analysis
The component analysis technique allows estimates of water use to be based upon an
individual component, e.g. household appliances, industrial machinery, crop and the extent of
usage or production of that individual component.
The component method of water use forecasting has the following advantages:
• It can be useful in determining the effect of changing uses and production technologies on
overall water use;
• It can be used to identify key areas of use, predict areas of use which may increase or
decrease and how these may effect the total water use.
The component method of water use forecasting has the following disadvantages:
• It requires a large amount of data;

• The analysis required is detailed;
• Representative surveys to collect the relevant information needed to carry out the analysis
can be costly;
• The method may not produce reliable results.
Component methods attempt to relate the water demand and use to demand determinants.
Component methods are more complex and rely on greater data and information needs.
However, if improved demand forecasts are required, these techniques are likely to be of
greater value to the users. Water demands and use are disaggregated into major components
e.g. the water demand of an individual component, be it household appliance or crop, is
estimated or measured. Examples of the components of an urban water supply are shown in
Figures 8.5 and 8.6.

Point of chargeable
abstraction licence Raw water abstracted
Raw water
Raw water retained Raw water imported
exported
Raw water collected
Raw water collected Raw water Raw water Non-

losses operational potable
use supplies
Point of production
Treatment Treatment Potable Potable
works operational works water water
use losses produced imports
Distribution input
Distribution input
Water taken Distribution losses

Distribution supply
Water delivered Distribution losses
Point of delivery operational use
Water delivered
Point of delivery
Water delivered billed Water taken unbilled
Unmeasured Measured Unmeasured Measured
Legally Illegally
household household non-household non-household
Water delivered billed

Underground supply
pipe losses Consumption
Underground supply Total plumbing

Customer use
pipe losses losses
Underground supply Above ground Internal plumbing
Customer use
pipe losses supply losses losses
Internal plumbing Customer use
Supply pipe losses
losses
Figure 8.5 Components of an urban water supply system

Distribution mains
Point of Point of
delivery consumption
Communication pipe Supply pipe
Underground supply
pipe losses Consumption
Underground supply Total plumbing

Customer use
pipe losses losses
Underground supply Above ground Internal plumbing
Customer use
pipe losses supply losses losses
Internal plumbing Customer use
Supply pipe losses
losses
Figure 8.6 Household level components
Future changes in each component are predicted separately and then aggregated. For water
demand and use forecasting carried out at a sub-catchment and catchment level, the following
should be taken into account:
• Population levels and growth;

• Socio-economic factors;
• Changes in unaccounted for water;
• Policy decisions affecting water resources management.
The first two components in the list are relatively easy to incorporate as data is usually
available. The population component tends to be based on information about demographic
characteristics such as age, sex, fertility rates, death and birth rates and migration. The socio-
economic situation component includes information on personal incomes, employment and
industrial production. Components such as changes in unaccounted for water and policies
affecting water resources management such as water pricing, metering, subsidies and water
quality standards, are harder to forecast, although they will have a significant impact on water
demand.
This technique requires more input and therefore cannot be implemented unless the requisite
data is available. Where data is available on other uses of water that are likely to affect future
demand, then they too should be incorporated into the forecast.
8.6.4 Multiple linear and non-linear regression analysis
Multiple linear or non-linear regression analysis can be used to relate water use to the various
parameters such as Gross Domestic Product (GDP), tariff levels or population levels. The
main disadvantage of this method is that it is time consuming to set up multi-variable
relationships and it requires large data sets to produce a reliable relationship. It is not
recommended that such techniques are used at a catchment or sub-catchment level to forecast
water use.
8.7 Forecasting agricultural water demand and use

Agricultural water demands are primarily a function of the following:
• Meteorological conditions;
• Crop type;
• Cropped area for each type of crop;
• Type of irrigation method and irrigation efficiencies;
• Water charges.
In assessing future agricultural demands the changes in one or more of the above must be
predicted. Owing to the potential for changes in the future and the relatively few variables,
component analysis is the forecasting methodology likely to yield the best results.
Component analysis is a forecasting method based upon the usage of water by an individual
component. The expected trends in each of the components (e.g. changes in irrigation
technology, climatic changes, crop type and irrigation area) should be analysed separately and
the overall result assessed.
Meteorological conditions are changing due to the effects of global warming. For agriculture,
the main considerations are rainfall and evaporation rates. As rainfall decreases and/or
evaporation rates increase, the irrigation needs are increased. Changes in crop type can have
a significant impact on water demands, which may or may not be positive. High value crops
such as garden vegetables generally have higher water requirements than, for example, grain
crops. Changes in cropped area can be a change in total area, as new land is developed or
existing land taken out of production, or it can be a result of changes in crop type, as the area
of one type is changed to accommodate changes in another type. Losses and inefficiencies
usually account for a significant proportion of total irrigation requirements. Their impact can
be improved through various means and the analyst will have to determine the likelihood and
magnitude of improvement programmes.
Instituting water user associations can also affect water demand and use. The main purpose
behind water user associations is to improve water security and equity in access to water.
Hence, while the water user association is beneficial in many ways, its formation may
increase water demands. Water charges have a significant impact on water demands because
of the incentive to reduce waste in water application, though consideration must be given to
willingness to pay and ability to pay on the part of the water users.
Changes in irrigation technology can also affect water demand. Modern technologies that
deliver water to the plant more efficiently reduce overall demand through reducing field
losses and non-beneficial evapotranspiration. Improvements in field level water management
and irrigation system operation also reduce irrigation water use through improving efficiency
and reducing losses. Other means of managing agricultural water use is practising new
techniques in application of water through precision irrigation and deficit irrigation, though
the ability of farmers to adopt such techniques must be assessed. Figure 8.5 shows how
switching from sprinkler irrigation to drip irrigation can reduce losses (i.e. increase irrigation
efficiency) and have a significant effect on agricultural water demand forecasting.

Phasing in of more efficient
Forecast irrigation
irrigation methods e.g. change
water use should
from sprinkler to drip irrigation
“inefficient” methods
continue to be used
Irrigation water demand and use
Reduction
in losses
High
losses
Irrigation water use

Irrigation water demand
2000 2010 2020 2030
Year
Figure 8.7 The effect of a change in irrigation technique on water demand

and use
Change in population is critical in the local sense and in the national sense. Nationally, a
change in population affects the nation’s food needs. The impact of rural-to-urban migration,
a feature in most countries in the last few decades, also changes food needs and also changes
the labour force available for agriculture, which affects water use. Changing economic
conditions, especially increasing affluence also has an effect on what crops are grown and
consumed.
8.8 Demand forecasting for rural domestic water supply

Historical information for rural domestic water demand and use are often not available. This
means that it is often not possible to assess future rural domestic water demand and use
through trend analysis. The two most important factors that affect future rural domestic water
demand and use are:
• Population growth;
• Change in the level of service owing to upgrading of the water supply system.
Population growth can be estimated from national, regional or local trends. It should however
be noted that improvements to infrastructure, such as water supply systems, could increase the
population growth above the average. Upgrading of water supply schemes and the
consequential changes in the level of service are difficult to predict. It has been postulated
that the upgrading of rural water supply schemes is related to:

• Economic conditions such as the Gross Geographic Product (GGP) the regional
equivalent of the Gross National Product (GNP);
• Tariff levels.
For simplicity, catchment managers and water resources planners will have to make
assumptions concerning upgrading (e.g. from a communal borehole to a yard connection to a
house connection). The increase in water demand may be estimated from other areas where
similar upgrading of the water and sanitation infrastructure has occurred. There are
significant discrepancies in the literature on rural water supply schemes as to what water
demand level can be expected from different levels of service. An example of rural water
demand forecasting is given below.
Example of rural water demand forecasting
The village of Manoti in Zimbabwe has a total of 500 settlements. These have the following
levels of service:
• 350 settlements are supplied by community boreholes within 100 m of the properties.
These have a water demand of 30 litres per person per day;
• 100 settlements are supplied by yard connections and have a water demand of 70 litres
per person per day;
• 50 settlements are supplied by a household connection and have a water demand of
130 litres per person per day.
The household occupancy rate is estimated to be six people per settlement. The water
demand over the next ten years is required. Demographic studies have indicated that the
number of settlements reliant on community boreholes will decrease by 1% a year over the
next decade, with the number of settlements using yard connections and supplied by
household connections will increase by 15% and 11% respectively over the next decade.
The calculations and water demand projections are shown below. The forecast is shown in
Figure 8.8.
Year Growth rate for different settlements Water demand

Number supplied Number supplied Number supplied (Ml/day)
by boreholes by yard by household
(-1% growth p.a.) connections (+15% connection
growth p.a.) (+11% growth p.a.)
2000 350 100 50 24.00
2001 347 115 56 25.74
2002 343 132 62 27.59
2003 340 152 68 29.68
2004 336 175 76 32.21
2005 333 201 84 34.98
2006 330 231 94 38.29
2007 326 266 104 41.92
2008 323 306 115 46.06
2009 320 352 128 50.88
2010 317 405 142 56.32

60
50
40
Water demand (Ml/day)
30
20
10
0
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
Year
Figure 8.8 Example of a rural domestic demand forecast
8.9 Forecasting industrial water demand and use

The simplest approach to forecasting industrial water demand and use is to carry out a trend
analysis as follows:
• Establish the relationship between industrial production and water use for various sectors;
• Assess the trends in industrial production;
• Project trends in industrial production into the future to establish the future water use.
Where industries are metered the relationship between water use and productivity should be
relatively simple to establish. However, the above approach will be more difficult for non-
metered industries. For non-metered industries a well-planned programme of surveys and
spot measurements will provide enough data to achieve an assessment to a reasonable level of
accuracy.
Gross Domestic Product (GDP) has become a common variable used in demand forecasting
models. Past studies in a number of countries have shown that there is a reasonable
correlation between GDP and industrial water use. However, no account is taken of the
changing nature of industrial production and the effect on water demand and use. For
example, in many industrialised countries production has moved away from manufacturing
and towards the commercial service sector and this provides different patterns of water use.
Technological improvements in industrial production can also lead to significant water
savings. For example, in Israel, industrial production increased by 250% whilst water
consumption rose by only 30% owing to regulation and licensing encouraging the use of
water saving and re-use technologies.
A study carried out by the International Water Management Institute in 1998 produced a
relationship between per capita water withdrawals and GDP in the year 2025 for countries
with a per capita industrial water demand of 10 m3 or above in 1990. This relationship,
shown in Figure 8.9, was based on GDP estimates for various countries provided by the
International Food Policy Unit and forecast industrial water withdrawals in 2025. It should be
noted that the correlation coefficient of 0.56 is relatively low indicating that there was a
relatively high degree of scatter in the data.
7
Per capita industrial water withdrawals Ind (m 3)
5
Natural log transformation
ln(Ind) = -3.405 + 0.881ln(GDP)

4
R2 = 0.56
0
4 5 6 7 8 9 10 11
Per capita GDP (1991 US$)
Natural log transformation
Figure 8.9 Example of a relationship between GDP and industrial water

demand
Component analysis is also applicable to the industrial sector because each industry uses
water differently. A component analysis looks at the use of water by an individual
component e.g. type of industrial machinery or application of water saving technology, and
the extent of usage or production of that individual component. Additional usage outside of
the aggregated components (depending on the coverage of the components) can be assigned
to miscellaneous uses and system losses. The relationship between the component and water
use is established using such tools as linear regression analysis. However, to use such a
method to forecast industrial water demand and use on a catchment or even sub-catchment
basis would be costly and time consuming.
8.9.1 Checklist of data for forecasting industrial water use
Before a forecast of industrial water demand and use is made the following checklist should
be consulted:
• For which industries are data currently available?

• Are there groups of industries with similar intensities of water use?
• Which are the major water using industries?
• Are there industrial sectors where past water use is not expected to provide a good guide
to future levels of water use?
• What is the present and likely future quality and reliability of the water supply?
• What are the current and likely future costs of the disposal and treatment of wastewater?
• What water demand management measures are likely to be implemented in the future?

8.10 Forecasting urban water demand
There are a number of factors that affect urban water demand and that should be taken into
account when forecasting urban water demand and use. These include:
• Demographic factors (e.g. population growth, migration, urbanisation)

• Socio-economic factors (e.g. standard of housing and living , employment opportunities);
• Climate;
• Type of sanitation;
• Extent of distribution leakage which is dependent upon the age of and pressure of the
system;
• Consumer wastage;
• Ratio of institutional and business demand and use to domestic consumption
• Availability of non-piped supplies;
• Tariff levels.
The main factors in estimating water use will be urban population growth and the level of
unaccounted for water. In many urban areas in Africa unaccounted for water levels can be as
high as 50%. Demand management measures including leakage reduction programmes and
other water conservation measures can often significantly reduce the level of unaccounted for
water. In some European countries unaccounted for water levels are of the order of 10%. In
any demand forecasting that is carried out for urban areas it is important that estimates are
made concerning the reduction in unaccounted for water levels in the future. Figure 8.10
shows an example of urban water demand and use forecasting. The forecasting method has
taken into account that the level of unaccounted for water will drop significantly over the next
35 years.
There are several methods available to forecast unmeasured household consumption. These
include:
• Micro-component analysis;
• Use of a constant annual per capita demand;
• Changes in consumption owing to assumed changes in household size.
Micro-component analysis involves forecasting the future demand for water by assessing the
future trends in the underlying elements of water use. It requires forecasts of the following:
• Water appliance ownership (e.g. washing machines, showers, baths);

• Frequency of the use of the various water appliances;
• The volume of water that is used each time the appliance is used;
The accuracy of a micro-component analysis is dependent on the quality of the baseline data.
This method requires large quantities of data and is relatively expensive to implement even
for a small sample size.

Water use
Water demand and use
Unaccounted
for water
Water demand
2000 2010 2020 2030
Year
Figure 8.10 Forecasting urban water demand
8.11 The effect of the HIV/AIDS virus on future water demand

in southern Africa
The spread of the HIV/AIDS virus in southern Africa has increased the difficulty in
accurately projecting possible population numbers over a long time frame. Population
projections that extend beyond a five to seven year time frame contain increasingly larger
inaccuracies and hence should be treated with caution. This is largely due to uncertainties
around probable behavioural changes in response to the anticipated massive number of deaths
that can be expected as the AIDS pandemic proceeds. Population projections are also
affected by uncertainties around the numbers of new immigrants that arrive in urban areas
each year, either from rural areas or from neighbouring countries. Therefore, due caution
must be exercised when population projections are made for periods beyond the next five
years. In addition, it is essential to remember that the limited availability of accurate,
widespread surveillance in southern Africa suggests that any estimates made are likely to
underestimate the true prevalence of HIV/AIDS.
Predictions of mortality rates owing to HIV/AIDS are based almost exclusively on the
realisation that no effective cure for HIV has yet been discovered. Currently it is believed
that every person recorded as HIV-positive appears certain to die within a period of between
seven to ten years from the date of first infection, unless antiretroviral therapy is administered
to halt the progression of the disease. In children, the situation is far worse and life
expectancy can drop to as low as two years for babies. Informed medical opinion considers
that the prevalence of HIV/AIDS appears to reach a plateau at between 32% and 35%, when

approximately 1 in every 3 individuals is infected. Prevalence survey data demonstrate that
somewhat fewer men than women are infected with HIV/AIDS (Reference 8.2).
Given the range of uncertainty around predictions of the possible numbers of people infected
with HIV/AIDS and the resultant mortalities, it is clear that estimates of population numbers
will also likely be inaccurate. Together, these uncertainties will reduce the accuracy and
reliability of future water demand estimates for specific geographical areas and countries.
If water demand estimates do not take HIV/AIDS-related mortality into account, demands for
water could be over-estimated by between 10% and 30%. This would pose several possible
unanticipated consequences for the construction and operation of large-scale water supply
schemes. In particular, if anticipated number of deaths from HIV/AIDS do reach the very
high levels suggested above, this would delay the demand for water by between 10 and 20
years. In addition, if this scenario were to hold true, construction of large water supply
schemes within current planning time frames would result in unnecessary expenditure of
capital. The converse situation is also important: if mortality rates are over-estimated, the
growth in water demand profiles of an area or country will not be anticipated correctly.
Given the relatively long lead-in times for water supply projects, a population would face
undue hardship if adequate water supplies cannot be provided in time (Reference 8.2).
8.12 Case study for demand forecasting for city of Masvingo

in Zimbabwe
8.12.1 Background
In 2002 the University of Zimbabwe carried out a case study of water demand for the city of
Masvingo in Zimbabwe (Reference 8.3). This section details the findings of the study.
The city of Masvingo is located approximately 300 km south of Zimbabwe’s capital Harare
and has an estimated population of around 70,000. Between 1977 and 2001 water use in
Masvingo increased from 2.4 million m3 per year to 6.8 million m3 per year. Lake Mutiriki,
which has a storage capacity of some 1,400 million m3 acts as a reliable source of water for
Masvingo. However, the main purpose of Lake Mutiriki is to support several thousand of
hectares of irrigated sugar cane. The water treatment works in Masvingo has a capacity of
24,000 m3/day. In the hotter months of September and October this is insufficient to meet
demand that can reach 28,000 m3/day (Reference 8.3).
Figure 8.11 shows the treated water production for Masvingo between 1977 and 2001. There
has been a general increase in water production. However, the effects of droughts in 1983/84
and 1991/92 on water consumption can clearly be seen. During droughts demand for water is
suppressed through rationing measures, decreased economic activity and public awareness
campaigns. Studies carried out between 1999 and 2001 indicate that unaccounted for water is
around 15% (Reference 8.3).

8
7
Treated water produced (million m 3/year)
0
1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001
Year
Figure 8.11 Treated water production for the city of Masvingo 1977 to 2001
In order to model past and to forecast future water consumption for Masvingo a multiple
linear regression model was developed. This is discussed below.
8.12.2 Use of a multiple linear regression equation to forecast demand
A multiple linear regression analysis was carried out correlating water consumption for
Masvingo with the following factors:
• Population based on annual data;

• Rainfall based on annual rainfall data;
• Economic development based on figures for Gross Domestic Product;
• Rationing. This was represented using a dummy factor with a “memory” of five years.
This factor is 1.0 for a drought year and decreases by a factor of 0.2 for each year
subsequent to a drought.
The multiple linear regression equation yield an equation of the form:
Q = 1,496 + 90.2 N − 1.5P + 26.8GDP − 837 R
R 2 = 0.965
Where:
Q is the annual quantity of water treated in 1000 m3/year;

N is the population of Masvingo in thousands;

P is the annual precipitation in mm/year;
GDP is the annual change in Gross Domestic Product in percent;
R is the rationing factor with a “memory” of five years starting a value of 1.0 and decreasing
by 0.2 in each year subsequent to a drought.
The formula implies the following:
• The constant 1,496 represents water uses that are more or less fixed and independent of
population, rainfall, GDP and rationing in 1000 m3/year. These uses include water losses
(950 x 103 m3/year, see above), to some extent institutional water uses, and to a lesser
extent industrial and commercial uses.
• The constant 90.2 is in m3/person/annum and is equivalent to 247 l/person/day. This
represents the “crude” per capita water consumption, and includes some industrial and
commercial uses. Population alone explains 88% of total water supply.
• The constant 1.5 is in 1000 m3/mm means that if rainfall is 100 mm above the average of
600 mm/year, water consumption decreases by 150,000 m3/year, if rainfall is 100 mm
below average, consumption increases with the same amount. Including rainfall improves
the correlation with 5%.
• The constant 26.8 is in 1000 m3/year and implies that change in GDP has relatively little
effect on water consumption: a 1% increase in GDP leads to an increase in water
consumption of 27,000 m3/year. Including this factor increases the correlation by only
0.4%.
• The constant 837 is in 1000 m3/year. This figure indicates that rationing has a significant
impact on water consumption. In a drought consumption drops by 837,000 m3/year. If
this factor is included the correlation is improved by 3%, yielding a total correlation of
96.5%.
A comparison between the multiple regression equation and the actual observed data is shown
in Figure 8.12.

8
7
Treated water produced (million m 3/year)
0
1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001
Year
Actual water water treated Modelled water treated
Figure 8.12 Actual and modelled water use for Masvingo 1977 to 2001
The projections in future water demand were made assuming that the factors that influenced
past water consumption, as given in the multiple linear regression, will remain dominant in
the future. The following assumptions were also made:
• The gross domestic product will be positive;

• Normal rainfall will occur and there will be no water rationing as a result;
• The population growth of 3% per year over the past four years will continue.
Table 8.1 below provides details of the unrestricted demand forecasts. The predicted water
use is shown in Figure 8.13.

Table 8.1 Forecast of unrestricted water production for Masvingo using a
multiple linear regression
Year Population Annual Annual change in Forecast unrestricted

(thousands) rainfall (mm) GDP (%) water demand
3
(million m /year)
2001 71 676 -4 6.8

2002 74 356 -1 7.6
2003 76 600 1 7.5
2004 79 600 2 7.8
2005 81 600 3 8.0
2006 83 600 4 8.2
2007 86 600 5 8.5
2008 89 600 6 8.8
2009 91 600 6 9.0
2010 94 600 7 9.3
2011 97 600 7 9.5
2012 100 600 7 9.8
14,000
12,000
10,000
Water use (1000 m /year)
3
8,000
6,000
4,000
2,000
0
1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013 2015 2017 2019 2021
Year
Observed water use Conventional projection of water use Current capacity
Stressed capacity Capacity by 2005 with WDM Capacity by 2012 with WDM
Capacity by 2020 with WDM Projected water use with WDM

Figure 8.13 Predicted water use for Masvingo up to 2021
8.12.3 The effect of water demand management measures on future

water demand
The multiple linear regression equation was used to estimate unrestricted water demand.
However, per capita water use in Masvingo is high especially in the more affluent parts of the
city. After forecasting the unrestricted water demand a forecast was carried out assuming that
demand management measures were implemented. The demand management measures that
were considered were:
• Reduction of water losses;

• Changes in the tariff structure;
• Retrofitting of water saving devices.
(i) Reduction of water losses
Unaccounted water for the city of Masvingo is about 15% equivalent to 2,600 m3/day. Some
of the losses are caused by the high static pressure (in many places between 80 m and 90 m)
that leads to a high number of pipe bursts. Reducing the pressure to between 30 m and 60 m
could lead to a reduction in water losses of some 7% equivalent to 180 m3/day. There is also
leakage of some 100 m3/day on one of the major reservoirs on the system (Reference 8.3).
Currently there are no passive leakage surveys being carried out in Masvingo. The
introduction of leakage surveys and quick responses to leaks coupled with a public awareness
campaign to encourage members of the public to report leaks could lead to a saving of
600 m3/day.
(ii) Changes in the tariff structure
There are currently 1,200 households in Masvingo that can be classified as affluent. These
are households with relatively large disposable incomes, large gardens and some times
swimming pools. The average water use of these households is 78 m3/month compared with
non-affluent households that consume around 20 m3/month. A change in the tariff structure
could result in reduction in water use of 480 m3/day. The proposed tariff structure would
comprise a rising block tariff. The effect of the new tariff on water consumption would have
to be carefully monitored to quantify the saving derived from this measure (Reference 8.3).
(iii) Retrofitting of water saving devices
The retrofitting of devices such as low volume toilets and shower heads have been shown to
reduce overall domestic water use. The expected savings that can accrue from these
measures have been shown to be at least 25% of domestic consumption. For Masvingo if this
were to be effectively applied to cover 50% of all the households there would be a water
saving of 1,540 m3/day. An immediate and cheap measure that can be implemented is to
reduce the cistern capacity of toilets. The flushing of toilets has been estimated to constitute
30% of total domestic water use. Adjusting floats in existing installations, or simply putting a
standard brick in the cistern would reduce its capacity by at least 10%. This means that each
household would save 3% of its consumption equivalent to 370 m3/day for the whole of
Masvingo (Reference 8.3).
8.12.4 Potential and short-term reduction of water demand for Masvingo
Table 8.2 shows the potential short and long term reductions in water demand for Masvingo.

Table 8.2 Potential long and short term reductions in water demand for
Masvingo
Water demand management Long term reduction of Short term reduction of

3 3
measure water demand (m /day) water demand (m /day)
Reducing losses
Pressure reduction 180 180
Repair of reservoir 100 100
Passive leakage control 600 300
Reducing demand
Block tariff structure 480 240
Retrofitting 1,540 370
Total reduction 2,900 1,190
Current water use 18,600 18,600
Percentage reduction of demand 15.6% 6.4%
The decrease in water demand of 1,190 m3/day or 0.44 million m3/year can be achieved in the
short term. This reduction in demand is equivalent to the expected growth in water use during
a period of two years. This would provide a window to implement the next supply option. In
the longer term with sustained campaigns coupled with political and technical support, the
reduction in demand may reach 1.0 million m3/year or 15.6% of current water use. Table 8.3
provides details of the forecast water production with demand management measures in place.
Table 8.3 Forecast water production with demand management measures in place
Year Unrestricted forecast water Water demand Forecast water

3
production (million m /year) management measures production with
demand
Domestic Industrial Losses Domestic Reduction management
demand factor for measures in place
3
reduction losses (million m /year)
factor
2001 4.76 1.22 0.82 1.00 1.00 6.8

2002 5.34 1.37 0.91 0.95 0.95 7.3
2003 5.26 1.35 0.90 0.94 0.94 7.1
2004 5.43 1.40 0.93 0.93 0.93 7.31
2005 5.56 1.44 0.96 0.92 0.92 7.5
2006 5.77 1.48 0.99 0.91 0.91 7.6
2007 5.95 1.53 1.02 0.90 0.90 7.8
2008 6.13 1.58 1.05 0.89 0.89 7.9
2009 6.30 1.62 1.08 0.88 0.88 8.12
2010 6.49 1.67 1.11 0.87 0.87 8.3
2011 6.67 1.72 1.14 0.86 0.86 8.4
2012 6.86 1.76 1.18 0.85 0.85 8.59

8.13 References
8.1 Seckler, D., Amarasinghe, U., Molden, D., De Silva R., and Barker R. (1998) World
water demand and supply, 1990 to 2025: Scenarios and issues Research Report 19,
International Water Management Institute, Sri Lanka.
8.2 Ashton, P. and Ramasar, V. (2001) Water and HIV/AIDS: Some Strategic
considerations in southern Africa published by the Council for Scientific and
Industrial Research (CSIR), South Africa
8.3 Dude, E. and Van der Zaag, P. (2002) Analysing water use patterns for water demand
management: The case of the city of Masvingo, Zimbabwe, Third Waternet
Symposium October 2002
8.4 Hunt, L.C. and Witt, R. (1995) An analysis of the UK energy demand using multi-
variate co-integration, Surrey Energy Economics Discussion Paper
8.5 Djokic, A. and Maidment, D.R.. (1989) Aggregated water use planning in Southern
California, Water use planning and management proceedings of the sixteenth annual
conference published by the American Society of Civil Engineers
8.6 UK Water Industry Research (1995) Demand forecasting methodology Environment

Agency England and Wales Report Reference NRA R&D 427
8.7 Rand Water South Africa (2000) Literature study on water demand management and
water demand management forecasting models: Final report unpublished
8.8 Singh, G. (1999) Water supply and sanitary engineering Standard publishers
distributors, Nem Chand Jain
8.9 Gumbo, B. (1998) Dual water supply systems “Is this just another pipe dream”
Proceedings of the international conference on small and medium size domestic water
conservation, wastewater treatment and use.
8.10 UK Water Industry Research (1997) Forecasting water demand components: Best
practice manual Report reference No. 97/WR/07/01
8.11 Asian Development Bank (1999) Handbook for the economic analysis of water –
Chapter 3 Demand analysis and forecasting
supplhttp://www.adb.org/documents/handbooks/water_supply_projects/Chap3-
r6.PDFy
8.12 Faures, J, Hoogeveen, J and Bruinsma The FAO irrigated area forecast for 2030
http://lnweb18.worldbank.org/ESSD/essdext.nsf/18DocByUnid/18A78C3BF51F68D
585256BE9007535F2/$FILE/Faures_Hoogeveen_Bruinsma.pdf
8.13 Varis, O. (2000) Global changes and water resources

http://www.water.hut.fi/wr/research/glob/egloshow/egloshow2.pdf

9. RIVER TRANSMISSION LOSSES
9.1 Background
In many parts of southern Africa there is considerable temporal and spatial variation in the
rainfall patterns. In general river flows show even more variation than rainfall. To make the
optimum use of these limited and varying water resources river flows are often regulated,
often via reservoirs, not only between seasons but also often over periods of more than one
year.
There are numerous storage reservoirs in southern Africa that are used to store and control
water resources. Ideally these reservoirs are located close to the point of use. However, in
many cases this is not possible owing to physical and economic reasons. Water often has to
be conveyed over long distances. To avoid the costs of pumping and constructing pipelines
or canals, water is often transmitted via natural watercourses. Although this method of
conveyance avoids constructing expensive infrastructure it can result in significant losses
particularly in sandy rivers that do not have a perennial flow. This chapter details the nature
of river transmission losses and methods that can be used to estimate them. The information
in this chapter is based on work carried out in 1998 by D. Kammer of the Department for
Water Development in Zimbabwe (Reference 9.1).
9.2 Nature of river transmission losses

When allocating water at a catchment level and carrying out catchment management plans it
is important that these losses are taken into account. The losses in rivers can be categorised
as follows:
• Natural losses. These are:

- Evaporation and transpiration;
- Bed and bank storage seepage;
• Artificial losses e.g. abstractions for domestic, industrial and agricultural uses.
This section concentrates on natural losses that occur from artificial releases from dams.
Losses caused by evaporation from open water surfaces and by transpiration from riverine
vegetation are often relatively small. The majority of transmission losses consist of two
components:
• Initial losses These are losses that occur at the front end of the release wave and
that are used to fill up pools and saturate the river bed.
• Continuing losses These losses occur after the initial wetting of the bed has taken place
and a steady flow has been established. They are caused by
evaporation and bed and bank seepage.
The river losses are to some extent offset by influent “seepage gains”. These gains are the
results of water stored in the river banks seeping back into the stream after the release wave
has passed and the channel storage has been depleted. These losses are illustrated in Figure
9.1.

Release from dam
Discharge
Continuing
losses
Recorded hydrograph
100 km downstream
Initial
losses
Seepage gains
Time
Figure 9.1 Illustration of transmission losses
The main parameters that affect these losses are:
• The discharge rate of release;

• The physical nature of the river channel (e.g. gradient, bed material);
• The degree of saturation of the river bed and banks;
• The duration and total volume of the release.
The loss process is complex and it is difficult to quantify the effects that differences in the
physical and release conditions have on the loss process. This is because they can influence
each other in a positive or negative way depending on the conditions. It is therefore necessary
to assess the effects on a case by case basis.
9.3 Example of transmission losses from southern Africa

9.3.1 Zimbabwe
In 1998 research was carried out into transmission losses on 13 rivers in Zimbabwe which are
used to convey water released from dams. The rivers that were investigated were grouped
into three broad categories (Reference 9.1). These are:
Category 1 Relatively steep rivers with gradients varying in the range of 1 in 150 to 1 in
400. The bed of such rivers consists mainly of rock and varies in total width
from 30 m to 300 m. Rivers in this category often contain small pools
created by rock, sandbars or small weirs. The riverine vegetation and
individual stream channels within the main bed are limited.
Category 2 These are less steep with gradients in the range 1 in 500 to 1 in 800. They
often range in width between 100 m to 800 m and contain more sand beds

and vegetation than Category 1 rivers. Pools are common and are often
connected by relatively narrow stream channels.
Category 3 These are usually very flat rivers with gradients shallower than 1 in 900.
They are often wide, (up to 1000 m) and have extensive, deep sand beds.
Large pools occur and vegetation can be abundant in wide braided rivers.
The other important factor that affects losses is the degree of saturation of the river bed and
the degree to which a natural flow is present. The degree of saturation is hard to quantify and
is subjective. In Zimbabwe a broad classification was devised to help categorise how wet or
dry a river is. This classification is as follows:
• “Very wet” The bed of the river was totally saturated, pools are full and there is a
substantial base flow (at least 5% of the release rate).
• “Partly wet or dry” There is relatively little base flow, the bed is only partly saturated
and the pools are partly dry.
• “Very dry” There is no base flow and the pools and bed are dry.
Tables 9.1 to 9.4 provide generic data gathered from 13 rivers in Zimbabwe on the initial
losses, continuing losses, the velocity of release waves, provisional guidance for seepage
gains. It should be noted that the rates in the tables often overlap. This is the result of that the
tables accommodate a wide range of results
Table 9.1 Guideline rates for “initial losses” from rivers in Zimbabwe
3 3
(x10 m /km or Megalitres/km)
3
Degree of Release rates in m /s
saturation
<1 1 to 3 3 to 5 5 to 7 7 to 10 10 to 15
Category 1
Very wet 0 0 to 1 1 to 3 3 to 6 5 to 9 8 to 13
Partly wet or dry 1 to 4 3 to 6 5 to 9 7 to 12 10 to 16 13 to 22
Dry 3 to 8 6 to 12 9 to 16 12 to 20 15 to 25 19 to 35
Category 2
Very wet 0 to 2 1 to 4 3 to 8 7 to 12 10 to 17 14 to 24
Category 3

Table 9.2 Guideline rates for “continuing losses” from rivers in Zimbabwe
3 3
(x10 m /km or Megalitres/km)
3
saturation
<1 1 to 3 3 to 5 5 to 7 7 to 10 10 to 15
Category 1
Partly wet or dry 3 to 7 6 to 11 9 to 14 12 to 17 15 to19 17 to 22
Category 2
Category 3
Table 9.3 Guideline rates for the velocity of release waves from dams
for rivers in Zimbabwe (km/day)
3
saturation
<1 to 3 3 to 7 7 to 12
Category 1
Very wet 10 to 30 20 to 38 30 to 45
Partly wet or dry 6 to 19 14 to 28 22 to 36
Dry 1 to 8 8 to 18 16 to 27
Category 2
Very wet 8 to 22 16 to 31 25 to 40
Dry 1 to 6 5 to 13 12 to 20
Category 3
Very wet 6 to 14 12 to 23 20 to 35
Dry >1 to 4 2 to 8 7 to13
There is little information available concerning the rates of seepage gain from rivers in
Zimbabwe. Provisional guidance on seepage gains for rivers in Zimbabwe are given in Table
9.4.

Table 9.4 Provisional guidance for “seepage gains” for
rivers in Zimbabwe
3 3 3
Release rate (m /s) Seepage gains (x 10 m /km or
Megalitres/km)
1 to 3 1 to 3
3 to 7 3 to 5
7 to 10 5 to 7
The initial and continuing loss rates vary considerable. In the case of initial losses they vary
from 0 m3/km to 100,000 m3/km for relatively large releases in very dry, flat and sandy rivers.
To determine which rate to use for a particular river it is important to know not only the
physical conditions but the also the conditions under which the release occurred.
An example of how a calculation could be used for estimating the losses from an artificial
release in Zimbabwe is given below.
Example of estimating transmission loses in Zimbabwe

3
Some 20 million m of water is to be released from a dam in Zimbabwe. This water has to be
conveyed 100 km via a natural watercourse to its point of use. The river has a gradient of
between 1 in 500 and 1 in 800 and a total width between 400 m and 500 m. The bed consists
party of rock but there are also significant sandy patches. Pools are fairly common. Prior to
the release the river bed is only partly saturated, there is little base flow and some pools are
3
partly dry. The release rate is to be kept at 4 m /s this means that the entire duration of the
release is 58 days
Interpolating from Tables 9.1 to 9.3 gives the following figures:

3
• Initial losses: 14,000 m /km
• Continuing losses: 17 litres per second per kilometre
3
• Seepage gains: 4,000 m /km
• Speed of release wave 14 km/day
3 3
The initial losses = 14,000 m /km x 100 km = 1.4 million m
6 3 3
The duration of the initial losses = 1.4 x 10 m /4 m /s = 4 days
The duration of the continuing losses = 58 days – 4 days = 54 days

3
The continuing losses = 17 l/s x 100 km x 54 days x 86400 seconds = 7.9 million m
3 3
The seepage gains = 4000 m /km x 100 km = 0.4 million m
3
The total losses = 1.4 + 7.9 – 0.4 = 8.9 million m
The losses equate to approximately 45% of the water that has been released from the dam.
It should be noted that the use of Tables 9.1 to 9.4 only provides guideline rates for
Zimbabwe.

9.3.2 Kenya
In Kenya the following equation has been developed for estimating transmission losses:
Q l = 57.73KA 0.5
Where: Q1 is the loss rate in l/s/km

K is a constant varying with the nature of the bed and the bank material e.g.:
K = 0.44 for sandy and gravelly channels;
K= 0.13 for sandy loam beds;
K = 0.08 for heavy clay
A is the cross sectional area in m2
Example of estimating transmission loses in Kenya

3
For a river in Kenya with a sandy bed (e.g. K = 0.3) that has a flow of 5 m /s at a mean
2
velocity of 0.5 m/s (i.e. a cross sectional area of 10 m ) the equation will give a continuing
loss of 55 l/s/km. If the bed were rocky (e.g. K = 0.1) the loss rate would be of the order of
18 l/s/km.
9.4 References
9.1 Kammer, D. (1998) River transmission losses in Zimbabwe, Department of Water
Development Zimbabwe (unpublished)

APPENDIX A
INDUSTRIAL WATER
CONSUMPTION
LOOK UP TABLES

A1.1 Background to look up tables
The aim of the look up tables is to allow staff responsible for water management to assess
whether the licensed volumes for an existing or proposed abstraction are reasonable for a
given industrial process. There is also some information on the water use steps for each
industrial process and potential water saving initiatives that could be employed to reduce
consumption levels.
The quality of the data in the look up tables varies. There are some industrial sectors where
data was available from a wide range of sources, for other industries the data available was
limited. In some cases water use figures are based on limited data provided directly by one
company. The data in the tables should be treated with care. However, the water
consumption figures generally provide at the very least an idea of the order of magnitude of
water consumption for various industries.
For industries that consume significant quantities of water, water consumption varies in an
approximately linear manner with an increase in production. Hence the planned production
capacity of an industrial plant is important in establishing the water consumption. The data in
the look up tables tend to concentrate on industries that use relatively high quantities of water.
A1.2 Examples of how to use the industrial water

consumption look up tables
The look up tables presented in Section A1.3 provide water resources managers with
indicative industrial water consumption figures. The look up tables aid decisions on water
abstraction licences, forecasting and demand management. Examples of how the look up
tables can be used are given below.
Example of water consumption for a new brewery
3
A proposed new brewery requires 780,000 m of water per year when fully operational
Information required
How many litres of beer are to be produced ? 10 million litres per month when fully
operational
Calculations
The quantity of beer produced per year = 10 x 12 = 120 million litres per year
3
= 120,000 m per year
3 3 3
The quantity of water consumed per m of beer produced = 780,000/120,000 = 6.5 m /m
Conclusions
The look up tables indicate that the typical range of water consumption for medium to large
3 3 3
breweries is 5 m to 15 m per m of beer produced. Hence the water consumption of the
3 3
proposed new brewery of 6.5 m per m of beer produced would appear to be reasonable.
Handbook for the assessment of A-1

Example of water consumption for a new beef processing factory
A new beef processing factory has been proposed. The new factory will to process 7,000
3
tonnes of beef per year. The plans indicate that the factory will consume 40 m of water per
hour.
Information required
What is the product to be produced ? Beef
What is the production capacity ? 7,000 tonnes of beef per year

3
How much water is required ? 40 m of water per hour
How many hours is the plant in operation ? 7,500 hours per year
Calculations
3
The quantity of water consumed = 40 x 7500 = 300,000 m per year
Water consumed per tonne of beef produced = 300,000/7,000

3
= 42.9 m of water per tonne of beef
produced
Conclusions
The look up tables indicate that the best available figure for water consumption for a beef
3 3 3
processing factory is 4.3 m and an average figure is 5.8 m of water per m of beef
produced.
3
The site’s request for 42.9 m of water per tonne of beef produced appears unreasonable
3 3
when compared with the average figure from the look up tables of 5.8 m of water per m of
beef produced. There would seem to be considerable scope for recycling water and
reducing the consumption of the site.
A1.3 Industrial water consumption look up tables

A1.3.1 Beverages (soft drinks)
A soft drink can be defined as a non-alcoholic carbonated or non-carbonated beverage. The

production involves the blending of concentrates and additives with water. In terms of water
consumption in some cases the washing of returnable bottles can use up to 50% of a plant’s
total water consumption.
Orange juice concentrates are made by extracting the juice from oranges and concentrating it
in an evaporator. Fruit juice is manufactured by the blending of juice concentrates with
water. Table A1.1 below gives water consumption figures based on a survey of 26 soft drink
factories in South Africa. Table A1.2 below gives water consumption figures for a range of
beverages based on information collected in Europe.

3 3
Table A1.1 Water consumption for soft drink factories in South Africa (m /m )
Average 2.6
Typical range 1.3 to 3.8
Note: The above figures are for all types of soft drink manufacture including carbonated and non-
carbonated drinks.
The degree of bottle washing has a significant effect on water consumption. The figure of
3 3
1.3 m of water per m of soft drinks produced is for a factory where no bottle washing takes
place.
Source: Reference A1.1
A study carried out of a bottling plant in Zambia indicated that the specific water
consumption for a soft drink bottling plant in Lusaka was 6.4 m3/m3. This figure includes the
water used for bottle washing (Reference A1.39).
Table A1.2 Water consumption for beverages
Carbonated soft Orange juice concentrate Fruit juices

3 3 3 3 3
drinks (m /m ) (m /tonne of oranges) (m /m )
Average 4.0 0.55 14.5

Best available 3.0 Not available 3.0*
Note: *This figure assumes re-circulation of the cooling water
Source: References A1.2 and A1.3
In many plants that produce bottled beverages it is bottle washing that is often the major
consumer of water. It has been estimated that for a relatively old beverage plant in Bulawayo
in Zimbabwe some 54% of the water consumed was used in the washing of bottles. However,
it should be noted for more modern plants this figure may be significantly lower (Reference
A1.4).
Methods by which water use efficiency can be improved include:
• Replacement of flow through systems with recirculation, recycling and re-use systems;
• Improvement of plant washing procedures;
• Controlling flow rates for spray, sealing and cooling water supplies especially in bottle
washing processes;
• Modifications to wastewater treatment systems to allow an increase in re-use and
recycling.
A1.3.2 Breweries
Modern larger breweries tend to be more efficient than small breweries. One of the most
efficient breweries in the world is located in Windhoek in Namibia. The brewery was
constructed with water efficiency in mind. The short lines in the beer production process
used in the brewery reduce the quantity of water needed for washing and reduces
consumption by around 10% (Reference A1.5). Table A1.3 provides typical water
consumption figures for breweries taken from a survey of eight breweries in South Africa.

Table A1.3 Water consumption for breweries in South Africa
Large breweries producing Medium breweries producing

3 3 3
greater than 10,000 m of beer between 2,000 m and 10,000 m of
per month beer per month
3 3 3 3
(m /m of beer) (m /m of beer)
Typical range 5.5 to 7.1 6.7 to 8.8

Mean 6.3 7.7
South African Breweries recently estimated that the water consumption for their breweries
worldwide was 6 m3 of water per m3 of beer produced. South African Breweries most
efficient clear beer brewery is Alrode in South Africa. This uses 4.19 m3 of water per m3 of
beer produced (Reference A1.7). The target water consumption for this brewery is 4.0 m3 of
water per m3 of beer produced. Table A1.4 provides water consumption figures collected for
breweries in Europe and compare them with the best available figure for southern Africa.
Table A1.4 Water consumption for breweries
Medium to large breweries Small breweries

3 3 3 3
(m /m of beer) (m /m of beer)
Typical range 5 to 15 Up to 22
Mean value 10 Not available
Best available 4* Not available
Note: *Figure is for the Windhoek Brewery, Namibia
Various measures that can be implemented to make water use in breweries more efficient
include:
• Utilising ammonia as a refrigerant for the main refrigeration system, as well as the air
conditioning. This reduces water cooling requirements;
• Recycling of caustic soda solutions reduces the water consumption of bottle washing
machines;
• Use of stainless steel product handling vessels and pipe work provide a self-sterilising
effect that reduces the water needed for cleaning;
• A separate effluent discharge system allows separate treatment of all wastewater thus
optimising water recovery;
• An incoming water filter backwash recovery filter system can allow re-use of water for
gardening thus avoiding the need for irrigation water.
A1.3.3 Sorghum beer and sorghum malting
Sorghum beer is often referred to as “traditional” or “African” beer. The process of

converting sorghum grain into sorghum beer involves two distinct processes. These are:
• The malting of the sorghum grain;

• The brewing of the beer.
Typical water consumption figures for these processes taken from a survey in South Africa
are given in Tables A1.5 and A1.6.

Table A1.5 Water consumption for sorghum malting for South Africa
Medium to large maltsters Small maltsters

3 3
(m /tonne) (m /tonne)
Typical range 2.5 to 10.0 Up to 12.3

Weighted average 3.3 Not available
Target 3.0 7.0
Best available 2.5 Not available
The water consumption for malting is related to the size of the plant as well. There are certain
operations that require the same amount of water regardless of the level of production
3 3
Table A1.6 Water consumption for brewing sorghum beer in South Africa (m /m )

Target 2.0
Mean 2.5
The results of the survey in South Africa indicated that large sorghum breweries are not
necessarily more efficient than smaller ones. South African Breweries recently estimated that
typical water consumption figures for their sorghum breweries are between 2 m3 and 3 m3 of
water per m3 of sorghum beer produced (Reference A1.8).
Water efficiency can be improved as follows:
• Improvements in the efficiency of use of washdown water. This is used to wash tanks,
floors and vehicles. The use of pistol grips on hoses can reduce water use;
• The use of compressed air for cleaning can reduce water use;
• A high pressure low volume system can reduce water consumption by 35%;
• In maltings mechanised water control systems reduce water use;
• In sorghum brewing the following measures can be undertaken to improve water use
efficiency:
• Install boiler condensate returns to reduce evaporation losses;
• Pressurised cooking reduces evaporation.
A1.3.4 Brick production
Bricks are manufactured by grinding natural clays and mixing them into a marl by the
addition of water. The exact composition of the marl determines the type of the brick. The
figures for water consumption given in Table A1.7 are taken from information collected from
a large brick factory in the UK.
Table A1.7 Water consumption for brick production

3
(m per tonne of bricks produced)
Typical range 15 to 30

Water use efficiency can be improved by monitoring the actual use versus the theoretical
usage (with respect to the marl-water content specification) and maintaining the water use as
close to the theoretical usage as possible.
A1.3.5 Cement and concrete products
Typical water consumption figures for various cement and concrete products are given in
Table A1.8. The figures are based on information taken from a variety of sources worldwide.
Table A1.8 Water consumption for various cement and concrete products
Type of product Water consumption
Cement manufacture Cooling water 0.41 kg/kg clinker

Product water 0.325 kg/kg raw materials
3
Cement production 3.8 m per tonne of product
3
Concrete products 1.0 m per tonne of product
3
Concrete blocks 1.0 m per 200 blocks
3 3
Reinforced concrete 0.63 m /m
A1.3.6 Ceramics manufacture
The main steps in the production of ceramic products are cooling of equipment, finishing of
the products, preparation of glazes and the washing of equipment. Typical water
consumption values for a range of ceramic products are given in Table A1.9.
3
Table A1.9 Water consumption for ceramics manufacture (m /tonne of product)
Ceramic Sanitary Stone ware Glazed Fine

whiteware ware tile porcelain
3 2
(m /m )
Typical range 15 to 20 6 to 15 2 to 10 0.5 to 8.0 5.5 to 14.0

(without closed
cooling water)
Typical figures with 1.6 2.4 2 0.04 Not
closed cooling water available
cycle*
*Note: These figures assume that water use for auxiliary needs is met by the water reclaimed from the
purified industrial wastes
There are a number of methods by which the water consumption of ceramics manufacture can
be improved. The re-circulation of water in a fully closed system and the reuse of reclaimed
water from industrial wastes can significantly reduce water consumption as illustrated by
Table A1.9.
A1.3.7 Manufacture of various specialist chemical products
There is a wide range of specialist chemical products for which typical specific water
consumption figures are given in Table A1.10. The main steps that use water include:
cooling, steam production, product washing, effluent dilution, plant and product washing.

Measures for water reduction can be split into three distinct areas: water management
techniques, good house keeping measures and plant/process modifications.
Table A1.10 Water consumption for various chemical products

3
(m /tonne)
Resins, adhesives, disinfectants, photographic solutions Less than 1

Sulphonic acids, detergents, rubbers, pigments, salts 1 to 2
Silicones, polyacrylics, water treatment chemicals, chelating agents, 2 to 5
surfants, amine products, synthetic organic polymers, esters,
imides, anhydrides, quaternaries, alkyl ethers, salts, soaps
Brightening agents, dyes, biocides, herbicides, insecticides, 5 to 10
phosphates, pharmaceutical, intermediates, polyacrylics, amine
products, esters, soaps
Esters biocides, fungicide intermediates, mercaptan gas, odorants, 10 to 50
carbonates, thioglycollates, thioureas
Pharmaceutical intermediates, acrylates, amine products 50 to 100
Liquid crystals, buffer solutions, pigments, chlorine and bromine Greater than 100
products
Water reduction measures for the production of specialist chemicals can be spilt into:
• Water management techniques;

• Good housekeeping measures;
• Plant and process modifications.
Table A1.11 gives an indication of the percentage reductions in water use that can be
achieved.

Table A1.11 Water reduction measures for the chemical industry
Measure Reduction
in water
use (%)
Water management techniques
Metering individual product areas and setting reduction targets 30

Production scheduling, improvements in plant washdown, use of trigger hoses 50
Good housekeeping measures
Improved pipe work to reduce leaks 10
Good housekeeping (e.g. taps on hoses) 8
Better housekeeping to avoid cleaning 3
Storm water prevented from entering effluent system 7
Leak detection and reduction 30
Flow restrictors on vessel cooling lines 5
Plant and process modification
Cooling water re-use 21
Installation of cooling tower 50
Improved cooling tower 50
Re-use of water for batch dilution 25
Installation of pressure washers for cleaning blending tanks 5
Re-circulation of water in liquid ring vacuum pumps 50
Replacement of water seal vacuum pumps with dry versions 24
Installation of air chillers 5
Reduction in process water wastage 5
Water minimisation project 20
Modernisation of manufacturing facilities 38
Improvements in process efficiency 5
A1.3.8 Chipboard and medium density fibreboard (MDF)

production
Chipboard and medium density fibreboard are manufactured from waste wood that is
shredded, mixed with resin and pressed into boards. Typical consumption figures are given in
Table A1.12.
Table A1.12 Water consumption for chipboard and MDF
Chipboard Medium density fibreboard

3 3
(m /tonne of board produced) (m /tonne of board produced)
Typical range 0.23 to 7.2 0.31 to 7.2

Best available 0.23 0.31
The higher figure of 7.2 m3 of water used per tonne of board produced is for a ‘once-through’
manufacturing plant where there is minimal re-use of water. The lower values of 0.23 m3 and
0.31 m3 of water used per tonne of board produced are for plants with the latest
manufacturing techniques and water re-use technology and procedures.
It should be noted that the water re-use is inherent in the process. Environmentally there is
often a balance between decreasing water use and increasing energy consumption.

A1.3.9 Cosmetics manufacture
Typical water use figures for cosmetic products are given in Table A1.13. The quantity of
water use is dependent upon whether the cosmetics are packed on the site of production.
Water use efficiency can be improved by the efficient use of water for cleaning and re-cycling
of steam.
Table A1.13 Water consumption for cosmetics manufacture

3
(m per tonne)
Typical consumption 4.18

A1.3.10 Electronic goods assembly
The data in Table A1.14 has been collected for the production of electronic goods worldwide.
It should be noted that the water consumption figures are given in terms of m3 per m2 of
production space.
3 2
Table A1.14 Water consumption for electronic goods (m per m of production space)
Low consumption Medium consumption High consumption
0.02 Not available 30.0

Photocopiers, domestic Digital test equipment Clean room operations e.g.
appliances, computers, instruments manufacturing of laser devices
data storage devices
A1.3.11 Fibreglass production
Fibreglass is produced by extruding glass fibres from a furnace. This fibre is wound to form
the final product. The quantity of water used is dependent on the type of binding agents used
and the final form of the product. Typical water use figures are given in Table A1.15.
Table A1.15 Water consumption for fibreglass production

3
(m /tonne of produced)
Average range 10 to 20
Best available 2.5
Water efficiency can be improved by:
• Improving the efficiency with which vessels are cleaned;

• Introducing closed loop cooling systems;
• Using a water consumption monitoring and management system.

A1.4.12 Fish processing
There are a number of steps involved in the processing of fish. The main water use step can
be the use of holding and thawing tanks. These can account for 30% to 40% of the water
used. Water is also used for butchering, chilling, glazing, cleaning, cooking and cooling fish.
The quantity of water used is dependent upon the final product i.e. fresh or frozen. Table
A1.16 gives typical ranges of water use for a variety of fish.
3
Table A1.16 Water consumption for fish processing (m /tonne of product)
Tuna Other fish Filleting Herring Shrimp Fish meal
Average 10 to 20 17 9 to 25 4 to 8 30 to 60 1 to 3
Best available Less than 10 4 9 4 30 1
Table A1.17 provides specific water consumption figures collected in the 1980s for fish
processing plants in South Africa.
Table A1.17 Water consumption for fish processing for South Africa
3
(m /tonne of product)
Freezing of Perlemoen Fish meal Fish meal (Walvis

white fish (abalone) (South Africa) Bay Namibia)
shellfish
Range 3.1 to 8.5 2.5 0.37 to 2.30 0.40 to 2.15

Average 6.5 - 0.84 0.84
Water use can be reduced by:
• Replacing water cooling systems with air cooling ones;

• Improving the efficiency of cleaning by using more efficient high pressure hose systems.
A1.4.13 Food processing – dairy produce
Pasteurised milk
There are a number of steps in the production of pasteurised milk. These include:
• Raw milk reception including chilling and storage;

• Pasteurisation to destroy harmful bacteria;
• Standardisation to remove excess cream;
• De-aeration to expel unwanted gases;
• Homogenisation to ensure the milk has a uniform fat distribution;
• Cooling and packing
Specific water consumption figures for pasteurised milk dairies in South Africa are given in
Table A1.18.

Table A1.18 Water consumption for pasteurised milk dairies in South Africa
3 3 3
Production capacity (m of pasteurised Water consumption (m /m product)
milk produced per month)
1000 or less 3.25

1000 to 2000 0.98
More than 3000 1.04
3 3
Notes: The maximum specific water consumption was found 5.5 m of water per m of pasteurised
3 3
milk produced, the minimum specific water consumption was found 0.75 m of water per m of
pasteurised milk produced
The water usage figures presented in the above table are for milk production until the end of
processing including reception and vehicle washing but not including packaging
The volume of milk produced is not necessarily sold, as part may be utilised as raw material for
further processes
Figures from Europe for specific water consumption for liquid milk processing plants range
from 0.5 m3/m3 to 12.9 m3/m3 (Reference A1.3). Table A1.19 gives details of water
consumption figures for dairies producing milk in Zimbabwe. The figures are based on a
survey of Zimbabwean dairies carried out in 1999.
Table A1.19 Water consumption for milk dairies in Zimbabwe

3 3 3
Production capacity (m of pasteurised Range of water consumption (m /m
milk produced per month) product)
More than 3000 1.15 to 6.00

The effect of packaging of milk on water consumption
The effect of different types of packaging of milk products on specific water consumption is
shown in Table A1.20. Inefficiencies in bottle washing operations can lead to a two fold
increase in the specific water consumption for pasteurised fresh milk.
Table A1.20 Water consumption for milk in different packages in South Africa
3
Unpackaged Bottled Sachets Cartons
Mean 1.6 3.0 1.7 2.2

Maximum 5.5 5.4 3.2 2.6
Minimum 0.75 2.0 1.1 1.5
Note: Produced from raw milk, the reception stage is included
Other dairy products
Specific water consumption figures for other dairy products collected in South Africa are
given in Tables A1.21 and A1.22. It should be noted that in the case of cheese and milk
powder significant quantities of milk are required to produce a relatively small mass of the
final product.

Table A1.21 Water and steam consumption for dairy produce in South Africa
3
Cultured Fruit Sterilised Butter Ice cream

3 3 3 3
products (m /m ) juices and /UHT (m /tonne) (m /tonne)
mixes products
3 3 3 3
(m /m ) (m /m )
Mean 10.2 2.7 3.7 1.5 2.5

Maximum 13.8 5.5 6.2 6.9 3.1
Minimum 6.3 0.75 2.0 1.3 1.9
Table A1.22 Water and steam consumption for dairy produce in South Africa
3
Skimmed milk Milk powder Cheese Condensed milk

3 3 3 3
(m /tonne) (m /tonne) (m /tonne) (m /tonne)
Mean 3.6 11.8 23.0 4.4

Maximum 5.0 16.6 29.0 5.3
Minimum 2.1 8.7 16.4 3.5
Other water consumption figures that have been collected from around the world are given in
Table A1.23.
Table A1.23 Water and steam consumption for dairy produce

3
Bottled Cheese Ice Yoghurt Powdered milk

pasteurised milk cream
3 3
Water 5.8 8.5 0.9 m /m 4.2 0.03
Steam 0.19 0.25 0.14 0.09
Note: Includes water used at the reception stage
Ways in which the specific water consumption can be reduced include:
• The use automatic shut-off nozzles on all water hoses;

• The use of high-pressure, low-volume cleaning systems;
• Minimising spills of ingredients and of raw and finished product on the floor.
A1.3.14 Food processing – flour products
Table A1.24 gives figures for a variety of flour products.

Table A1.24 Water and steam consumption for flour products
Bread Bakers yeast Corn starch Corn Pasta

snacks
3
Water (m /tonne) 2* to 6 25 6.3 0.03 0.3
Steam (tonne/tonne) 0.06 0.4 - - -
Note: *The lower value is based on an audit of a multi-product bakery in North Carolina, USA
The water consumption figures in Table A1.4.24 are based on the actual requirement with no
water recycling. As a consequence the figures should be treated as a worst case value.
Methods by which specific water consumption can be reduced include:
• The use of nozzles on all water sprays;

• Installation of automatic shutoff valves on all water hoses;
• The use of a high-pressure, low-volume cleaning system;
• Installation of pressure regulators on supply lines for hoses and pan washers;
• Reuse of water where permitted, such as in pan or tray washer;
• If flour, product, or debris collect and it requires excessive amounts of water to flush
them away, a redesign of the equipment should be considered to prevent the materials
from collecting.
A1.3.14 Food processing – fruits
Tables A1.25 and A1.26 provide specific water consumption figures for the canning and
juicing of a variety of fruit products. The figures are based on a survey of South African fruit
processing plants carried out in the late 1980s.
Table A1.25 Water consumption for the canning of fruit produce in South Africa
3
Typical range Average Target

Apples 4.3 to 11.3 6.6 3.6
Apricots 2.5 to 14.5 5.5 5.5
Citrus fruit 1.1 to 2.6 2.1 1.5
Guavas 4.0 to 10.0 6.4 6.0
Peaches 2.5 to 11.5 5.5 5.1
Pears 4.5 to 12.9 12.7 8.0
Pineapples 2.1 to 4.5 2.9 2.8
Strawberries 6.8 to 27.0 17.0 -
Tomatoes 2.0 to 3.0 2.4 2.2

Table A1.26 Water consumption for the juicing of fruit produce in South Africa
3
Apples - 1.8 0.5

Citrus fruit 1.1 to 2.6 2.1 1.5
Pears 1.4 to 1.9 1.5 0.6
Table A1.27 provides water consumption figures based on data collected worldwide.
3
Table A1.27 Water and steam consumption for fruit products (m /tonne
of product)
Frozen apples Fruit puree Fruit jam/preserve
Average water consumption 2.0 0.4 8.0

Average steam consumption - 0.18 0.28*
Note: This steam figure is included with the average water consumption figure shown in the top row
of the table.
The main ways in which specific water consumption can be reduced include:
• Optimising the layout of the plant in order to minimise spillage;

• Recycling of wash water;
• The use of air cooling systems rather than water cooled ones;
• The recycling of steam;
• The implementation of good housekeeping techniques.
A1.3.15 Food processing – miscellaneous
Table A1.28 provides data for a variety of miscellaneous food products. The information is
based on figures collected in Europe.
3
Table A1.28 Water consumption (m /tonne of product)
Refined sugar beet Baby food Canned pet food
Average 9.0 0.2 2.8

A1.3.16 Food processing – multi-product confectionary plant
The figures in Table A1.29 are based on a water audit that was carried out for a Cadbury
Schweppes plant in Australia that produces chocolate confectionery products. The table
indicates the reduction in water consumption after the audit was completed. These reductions
were achieved by using specially designed wash bays, automatic hoses on nozzles, automatic
taps and utensil washers.

3
Table A1.29 Water consumption (m /tonne of product)
Before audit 6.0

After audit 1.5
A1.3.17 Food processing – vegetable products
Frozen vegetables
Table A1.30 provides figures for the water consumption for frozen vegetables based on data
collected in South Africa.
Table A1.30 Water consumption for the freezing of vegetable produce in South
3
Africa (m /tonne of product)
Broccoli - 8.1 -
Cauliflower - 25.0 12.5
Carrots 6 to 26 6.1 5.8
Corn - 4.6 -
Green beans 10 to 25 25.0 16.0
Peas - 30.0 20.0
Potatoes - 25.7 1.3
The target water consumption figures in Table A1.30 were set after audits of several factories
in South Africa. The main ways in which water consumption can be reduced is as follows:
• Optimisation of the factory’s layout. Many old plants have been steadily extended
leading to long distances between process areas. This leads to long inter-stage
transportation systems with inherent spillage and excessive water use if wet systems are
employed;
• Improvements in the standards of house keeping;
• Use of water monitoring equipment;
• Hose pipe management;
• Efficient operation of boiler and steam plants.
Canned vegetables
Table A1.31 provides typical water consumption figures for the canning of vegetable products
based on data collected in South Africa. The main ways in which water consumption can be
reduced are detailed above.

Table A1.31 Water consumption for the canning of vegetable produce in South
3
Africa (m /tonne of product)
Beans in tomato sauce 20 to 70 20 15

Corn 6 to 11 9.8 5.0
Green beans - 7.4 5.1
Peas 19 to 25 22 10
Edible oil
Figures in Table A1.32 detail typical water consumption ranges for the two main stages of
edible oil production based on data collected in South Africa. It should be noted that most
plants that produce edible oil also produce secondary products such as margarine, peanut
butter and mayonnaise. An average water consumption of 1.4 m3/tonne was found for
margarine manufacture.
3
Table A1.32 Water consumption for edible oil in South Africa (m /tonne of
product)
Typical range Average
Milling 2.1 to 3.1 2.6

Refining 3.2 to 4.6 3.8
The survey of edible oil industry was carried out in 1989. At this time only some 65% of the
capacity was being utilised. It should be noted that many plant operations require the same
amount of water regardless of the amount of oil produced. This leads to higher specific water
consumption figures when a plant is operating below its design capacity.
Table A1.32 implies that the water consumption for a combined mill and refinery should be
6.4 m3/tonne. This agreed well with figures collected in South Africa, although there are
some economies of scale that tend to reduce this figure for an efficiently designed combined
plant. Table A1.33 illustrates the breakdown of water use at a typical edible oil processing
plant.
Table A1.33 Breakdown of water use at an edible oil processing plant in South
Africa
Mill Refinery Total
Process 2% 13% 15%

Boilers 10% 30% 40%
Cooling 25% 10% 35%
Washdown 1% 7% 8%
Domestic 1% 1% 2%
Total 39% 61% 100%
Specific water consumption can be reduced as follows:

• Maximising the condensate return to the boiler house;
• Ensure cooling towers are working efficiently;
• Use of high pressure, low volume equipment for floor washing;
• Installation of water meters.
Miscellaneous vegetable properties
Table A1.34 illustrates various other specific water consumption figures collected worldwide
for a variety of vegetable products.
Table A1.34 Water consumption for vegetable products

3
Frozen vegetables Tomato paste Margarine
Average 25.2 0.3 1.4

Table A1.35 gives typical water consumption figures for a mushroom farm. These figures
include the water used to prepare compost and the steam that is used to kill any bacteria in the
compost. The figures are based on a mushroom farm in Europe.
3
Table A1.35 Water consumption for mushroom farm (m /tonne of product)
Compost production Including steam
Average 2.1 25.0

A1.3.18 Fresh meat production
In the late 1980s a survey was carried out of abattoirs in South Africa. To enable
comparisons to be made between abattoirs where more than one species of animal is
slaughtered, the non-bovine species are counted in terms of cattle unit (cu) equivalents (e.g.
one cow is equivalent to 15 sheep). To allow a comparison of water consumption to be made
a water-related cattle unit (wrcu) was developed e.g. the quantity water used to process one
cow is equivalent to the water required to process six sheep. Details of these units are given
in Table A1.36.
Table A1.36 Comparison of cattle units and water related cattle units
Species Cattle units/head Water related cattle unit/head
Cattle 1 1
Calves 3 2
Sheep 15 6
Goats 15 6
Pigs 5 2.5

Water consumption figures for fresh red meat consumption for different grades of abattoir are
given in Table A1.37.
Table A1.37 Water consumption for fresh red meat production in South Africa
Grade of Maximum daily Minimum Maximum Median Mean

3 3 3 3
abattoir slaughter allowed (m /wcru) (m /wrcu) (m /wrcu) (m /wrcu)
(cattle unit/day)
A >100 or for export 0.71 2.88 1.12 1.42

purposes
B 99 1.25 3.15 1.73 2.04
C 50 1.01 4.64 2.36 2.33
D 15 0.70 4.71 2.45 2.38
E 5 1.31 4.19 2.35 2.28
Note: wrcu is a water related cattle unit see Table A1.36
The main water use steps and the breakdown of water use for these processes are given in
Table A1.38.
Table A1.38 Breakdown of water use in red meat abattoirs in South Africa
Activity Operations Percentage of Typical water use

3
total (m /wrcu)
Range Mean
(%) (%)
Processing Lairage 5 to 12 10 0.142

Slaughtering 8 to 20 12 0.170
Carcass dressing 5 to 13 8 0.114
Offal processing 11 to 60 25 0.355
Utilities Hot water generation 14 to 36 25 0.355
Cooling and refrigeration 5 to 11 8 0.114
Steam raising 2 to 9 5 0.071
Services Vehicle and yard washing 1 to 5 2 0.028
General washing 1 to 5 2 0.028
Laundry and ablutions 1 to 3 3 0.043
Overall 100 1.420
Water use in red meat abattoirs can be minimised through:
• Use of water metering;

• Use of high pressure, low volume jets for cleaning;
• Elimination of steam losses from steam lines;
• Minimisation of windage and drift losses from evaporative cooling towers and condensers
by the fitting of demisters;
• The fitting of self closing nozzles to hoses;
• Recovery and re-use wastewater where possible.
Table A1.39 below provides water consumption figures collected for abattoirs in the UK.

Table A1.39 Water consumption for fresh red meat production
3
Lamb only Beef and lamb only Beef only Pig meat only
Average 2.2 3.3 5.8 8.9

Best available 1.6 2.4 4.3 3.3
Note: No significant difference in the specific water consumption were noted due to the production
level.
Data for beef from the USA indicate that specific water consumption ranges from
568 litres/animal to 1703 litres/animal.
A1.3.19 Commercial laundries
The processes used in commercial laundries are similar to those used in the domestic
situation. A survey carried out of South African commercial laundries yielded the
information shown in Table A1.40.
3
Table A1.40 Water consumption for laundries (m /tonne of product)
Range Typical value Target consumption
Rinse water recycled 8.0 to 16.7 9.0 8

Rinse water not recycled 23.0 to 58.4 30.0 20
The survey detailed in Table A1.40 did not indicate that the size of the laundry or the type of
articles to be washed had an influence on the water consumption. Water efficiency can be
improved by recycling the rinse water. Table A1.41 gives water consumption figures for
laundries in the UK and USA.
3
Table A1.41 Water consumption for laundries (m /tonne of product)
UK USA Domestic washing machinery
Average 27 38 20 to 15
Best available 21 10 15
A1.3.19 Lead acid battery production
Table A1.42 gives figures for water consumption for the production of lead acid batteries
based on data collected in the UK. It should be noted that automotive and standby batteries
tend not to be charged hence the water requirement for cooling is eliminated.

Table A1.42 Water consumption for lead acid batteries
3
(m /tonne of lead processed)
Industrial Automotive and standby
Average 16 to 20 5 to 10
Best available 15 5
The main ways in which water use efficiency can be increased includes:
• Implementation of water monitoring and management systems;

• Re-use of water from slurry preparation;
• Re-use of cooling water from the charging operation;
• Re-use of plate washing water;
• The use of dry fume abatement systems.
A1.3.20 Leather tanning
Tanning is the process by which animal hides and skins are converted to leather. There are
three clearly defined stages in leather tanning. These are:
• The ‘wet-blue’ stage;

• Re-tanning;
• Dyeing.
In the last ten years in South Africa there has been a trend for animal hides to only be
processed to the wet-blue stage and then to be transported elsewhere for further processing.
Water consumption results from a survey of 11 tanneries in South Africa are shown in Table
A1.43 below.
Table A1.43 Water consumption for leather tanning (litres/hide)
Range Average
Full tanneries 320 to 744 432

Wet blue plant - 339
Re-tanning, dyeing and finishing plant - 389
There was some indication from the results of the survey that large tanneries have a better
water use efficiency than smaller tanneries. The water consumption for full tanneries
producing more than 900 hides per day was 356 litres per hide while for tanneries producing
less than 900 hides per day the water consumption was found to be 546 litres per hide. Table
A1.44 gives figures for leather tanning based on figures collected in the UK and USA.
Table A1.44 Water consumption for leather tanning

3
(m /tonne)
Average for chrome tanning 40 to 67

Average for vegetable tanning 16 to 27

• The re-cycling of effluent from the secondary treatment to the pre-tanning stage;
• The re-use of effluent from liming wash.
A1.3.21 Light industrial estate water consumption
Table A1.45 provides information on water consumption for light industrial estates. The
figures for light industrial consumption are based on typical figures collected for a variety of
light industrial estates worldwide.
3
Table A1.45 Water consumption (m /worker/day)
Usage Consumption
Basic factory requirements for cleaning and sanitation 0.05

Average consumption in light industrial estates with no large water 0.25 to 0.50
consuming factories
Average consumption in light industrial estates that include a proportion of 0.90 to 1.10
factories engaged in food processing, ice making and soft drink manufacture
A1.3.22 Metal finishing
Metal finishing is used to give a product a service that makes it suitable for its intended
service conditions as well as providing an attractive appearance. The term “metal finishing”
covers a wide range of techniques for the treatment of metallic articles. These techniques
include chemical and mechanical surface pre-treatments, electroplating, post-treatments,
stripping, and anodising and protective coatings. The electroplating, anodising and chemical
surface treatment processes are the processes that consume the most water.
Electroplating is the process by which a thin coating of metal possessing certain desirable
properties (e.g. corrosion resistance, appearance) is deposited onto a cheaper base metal via
an electric current. Anodising is the process by which aluminium components are treated to
provide them with a protective oxide layer against atmospheric corrosion. Phosphating is a
common form of chemical surface treatment. It is used to improve corrosion resistance under
paint. Typical water use figures for these type of metal finishes taken from a survey carried
out in South Africa are given in Table A1.46.
Table A1.46 Water consumption for metal finishing for South Africa
3 2 3 2
Electroplating Anodising (m /m ) Phosphating (m /m )
3 2
(m /m )
Range 0.03 to 1.25 0.03 to 0.96 0.03 to 0.42

Average 0.31 0.10 0.13
Target water consumption figures for the three processes in Table A1.46 should be 0.1 m3/m2
for operations treating in excess of 10,000 m2/month and 0.2 m3/m2 for operations treating
less than 10,000 m2/month (Reference A1.18). Table A1.47 presents typical water
consumption figures for metal finishing based on data from the UK, USA and Australia.

Table A1.47 Water consumption for metal finishing
Metal Metal finishing US Galvanisers

3 3
finishing (m /tonne) electroplating (m /tonne)
3 2 3 2
(m /m ) (m /m )
Average 0.2 to 0.5 2.5 to 11.7 0.045 0.025

Best available 0.07 0.5 0.0008 -
Note: Water use will be product, material and finish specific.
Water consumption is normalised against tonne of product or treated surface depending on the
form of the product e.g. plate material or door furniture.
• The re-use of rinse water, this can reduce water consumption by up to two-thirds;
• Counter-current rinsing. This can reduce water use by 90% to 97%;
• Correct design of the rinse tank to ensure that the rinse water is well mixed;
• Treatment and re-cycling of water;
• Vacuum devices to remove dragout liquid.
A1.3.22 Metal processing
After mining metal ores require processing. The main steps in the processing are as follows:
• Concentrating based on physical and chemical methods;

• Smelting where the metal is melted and oxidised to separate it from the ore;
• Refinement of the smelted metal into commercially useful grades.
Tables A1.48, A1.49 and A1.50 give water consumption figures for the processing of lead,
nickel and copper respectively.
3
Table A1.48 Water consumption for nickel processing (m /tonne)
Smelting Refining Iron castings
Average 0.6 5.0 0.4

3
Table A1.49 Water consumption for lead processing (m /tonne)
Refining
Average 0.4
3
Table A1.50 Water consumption for copper processing (m /tonne)

Table A1.51 gives figures for water consumption for metallurgical plants processing gold and
uranium ore based on data collected in South Africa.
3
Table A1.51 Water consumption for gold and uranium processing (m /tonne)
Gold Gold and Gold, uranium Gold, uranium, pyrite

recovery uranium and pyrite recovery and
only recovery recovery production of
sulphuric acid
Typical range 0.9 to 1.4 2.5 3.0 4.0 to 4.5

The main ways in which water efficiency can be improved are:
• Improvements in the efficiency of cleaning methods and coating make-up procedures;

• Re-cycling of cooling water;
• Adopting an efficient steam system, including a condensate return to minimise leaks.
A1.3.23 Mining
Water is used by the mining industry for the extraction of minerals (e.g. solids such as coal
and ore), liquids (e.g. crude petroleum) and gas (e.g. natural gas). The mining industry also
uses water for quarrying, crushing, screening and washing.
It is recommended that the water used in the mining industry is divided into four categories.
These are:
• Metal mining;
• Coal mining;
• Oil and gas extraction;
• Mining and extraction of non-metallic materials.
The quantity of water used by a mining facility is dependent on the size and the type of the
operation. It may be possible to develop a relationship between the water used per tonne of
material produced. In many cases mines will be self-supplied from surface or groundwater
sources. It is recommended where possible that mine water use is taken from metered
records. It should be noted that mines often have to be dewatered. Hence a mine may actual
be a net producer of water (Reference A1.22).
Coal mining
Typical figures for coal mining from data collected in the UK are given in Table A1.52 below
3
Table A1.52 Water consumption for coal mining (m /tonne)
Mining Open cast
Typical range 0.33 to 0.45 0.11


The factors that determine water use are specific to the type of mine. The water use efficiency
is highly dependent on the type of processes and the amount of water re-use that occurs at
each mine.
Diamond mining
Typical figures for diamond mining are given in Table A1.53 below. These figures are based
on data collected from the Orapa diamond mine in Botswana. The figures are based on the
quantity of water consumed per tonne of ore produced.
3
Table A1.53 Water consumption for diamond mining (m /tonne)

The Orapa diamond mine has significantly increased its water efficiency in the past decade by
improving the efficiency of its wastewater re-use.
A1.3.24 Plastic manufacturing
The basic processing steps in the manufacture of plastics include:
• Preparation of reactants;
• Polymerisation;
• Polymer recovery;
• Polymer extrusion;
• Supporting operations e.g. equipment cleaning.
Table A1.54 gives figures for the water consumption of a number of products based on data
collected in the USA.
Table A1.54 Water consumption for various plastic products

3
(m /tonne)
Polyethylene 2.5 to 10.0

Polypropylene 1.7 to 2.8
Elastomer 9.2 to 37.5
Polyvinylchloride (PVC) 9.2
Monomers 66.6
Resin 5.8
Water reduction steps that can be implemented are similar to those outlined for the speciality
chemicals sectors in Section A1.4.7.
A1.3.25 Power generation
There are two main methods by which power is generated in southern Africa. These are:
• Thermal power stations;

• Hydropower stations.

The water use for these two methods is discussed below.
Thermal power station water use
Thermal power station water use can be defined as the amount of water used in the production
of electricity generated by heat. The source of the heat may be fossil fuels (e.g. coal, oil or
natural gas), nuclear fission or a geothermal source. In order to estimate the quantity of water
used by thermal power stations the following information is required:
• Self-supplied surface and groundwater withdrawals;

• Deliveries from public suppliers;
• Consumptive use;
• Power generation;
• Number and type of facilities.
Information on water use for generating electricity in thermal power stations is best obtained
from each individual utility. The quantity of water used at both fossil fuel and nuclear power
stations depends primarily on whether or not the cooling water is re-circulated. Power
stations that use a “once through cooling system” withdraw the largest quantity of water
because the water is not re-circulated within the facility. This technology is commonly used
in older thermal power stations (Reference A1.23).
The more water efficient alternatives to once-through cooling include cooling ponds and
cooling towers. In some cases where it is not possible to get the quantity of water used
directly from a power plant facility it may be possible to use data from other similar plants to
estimate the quantity of water used per kilowatt or mega-watt hour of electricity generated.
Eskom, the electricity utility in South Africa, estimates that approximately 1208 m3 of water
is used to produce 1 GWh of electricity (Reference A1.24). Water consumption figures for
coal and oil fired power stations in South Africa are given in Table A1.55.
Table A1.55 Water consumption for power stations in South Africa

3 3
Fuel type Direct cooled (m /GWh) Indirect cooled (m /GWh)
Coal 220 to 330 1280 to 2550

Oil 320 to 4600 -
Dry cooling technology in power stations significantly reduces the water use. Eskom has
implemented dry-cooling technologies at their Kendal, Matimba and Majuba power stations.
Approximately 80% of the water consumed at wet-cooled power stations is lost due to
evaporation from the cooling towers (Reference A1.24). Dry-cooled technology reduces
these losses and hence the total amount of water consumed at power stations.
According to design specification, the total station water consumption of a dry-cooled system
will not exceed 800 m3 of water consumed per GWh of electricity produced. The Morupule
Power Station in Palapye Botswana is a coal fired power station that utilises dry-cooled
modern technology. In 1997 it was estimated that this power station consumes 1140 m3 of
water per GWh of electricity produced. These figures compare to 2500 m3 of water
consumed by wet-cooled systems to produce one GWh of electricity (Reference A1.25).
These numbers vary depending on the technology employed at the various power stations.

However, the dry-cooled system may result in a loss of overall efficiency, thus impacting on
emissions produced.
Hydropower water use
Hydropower water use can be defined as the quantity of water used by a power plant where
the turbines are driven by falling water. In most cases the hydropower facility will be located
in the channel of the watercourse, often in the form of a dam (e.g. Kariba Dam in Zimbabwe).
The water used by these types of facilities is considered to be an instream water use. It should
be noted that in some cases water is diverted away from the watercourse to generate
electricity (e.g. the hydropower plant that forms part of Lesotho Highlands Water Project).
These types of facilities are considered to be an offstream water use. Instream and offstream
hydropower facilities are shown schematically in Figure A1.1 below.
Dam incorporating
a hydropower facility
Watercourse
Instream hydropower facility
Diversion tunnel
Watercourse Hydropower facility
or weir
Offstream hydropower facility
Figure A1.1 Instream and offstream hydropower facilities
In order to estimate the water used by hydropower facilities the following data is required:
• Instream water use;

• Withdrawals from surface water sources;
• Power generation;
• Number of facilities.
There are two main techniques for establishing the quantity of water required by hydropower
facilities. These are:

• Direct method;
• Indirect estimation method.
In the majority of cases the instream and offstream water used by a hydropower plant will be
available directly from each individual facility. However, in some cases where information is
not directly available, water use estimates can be made using the effective head for the
hydropower plant, the generated electricity and the efficiency of the turbines (Reference
A1.25). The following equation can be used:
P = ηρgQH
Where: P is the power generated in watts;

η is the overall efficiency of the station expressed as a percentage;
ρ is the density of water in m3/kg;
g is the acceleration due to gravity in m2/s;
Q is the quantity of water flowing through the turbines in m3/s;
H is the effective head difference in metres.
Rearranging the above equation gives:
P
Q=
ηρgH
Using this equation the water use for both instream and offstream hydropower requirements
can be calculated. It should be noted that wherever possible the quantity of water used by a
hydropower plant should be taken from direct method figures. Overall efficiencies for
hydropower stations are given in Table A1.56.
Table A1.56 Overall efficiency of hydropower stations
Type Efficiency
Head losses 85% to 95%

Turbine 80% to 90%
Electrical generation 90%
Total 60% to 77%

Example of estimating the minimum instream flow requirement for a hydropower
facility
Kafue hydropower plant in Zambia has six 150 MW turbines each with an efficiency of 90%.
Head losses are estimated to be 95% and electrical generation efficiency is 90%. To
produce sufficient electricity to meet the base load requirements a minimum of three of
these turbines are required to be in operation. The effective head difference for the plant is
400 m.
The minimum instream flow requirement can be calculated as follows:
Overall efficiency of the hydropower station = 90% x 90% x 95% = 77%

6 6
Minimum power requirement P = 3 x150 x10 = 450 x10 W
P 450 x10 6 3
Minimum instream flow requirement Q = = = 149 m /s
ηρgH 0.77 x1000 x9.81x 400
A1.3.26 Poultry processing
Table A1.57 includes water consumption figures based on a survey of 17 poultry abattoirs in
South Africa. The surveys indicated that the abattoirs processing over 10,000 birds per day
tended to have a lower specific water consumption than those processing less than 10,000 per
day.
Table A1.57 Water consumption for poultry processing in South Africa (litres/bird)
More than 10,000 birds per day Less than 10,000 birds per day
Average 18.4 23.9

Typical range 15 to 20 15 to 30
Best available 15 15
Figures available from the USA indicate that for broiler chickens the specific water
consumption varies from 13 litres/bird to 38 litres/bird. The American data for the processing
of turkeys indicates that specific water consumption varies from 42 litres/bird to 87 litres/bird
(Reference A1.3).
Water use can be reduced using the following measures:
• Using dry clean up practices i.e. removing all dry wastes from the floor before cleaning
with water;
• Introducing a water monitoring and management system;
• Maximising condensate return to the boiler house;
• Using of trigger gun attachments on hoses.
A1.3.27 Pulp and papermaking
The pulp and paper mills in southern Africa can generally be categorised as integrated, non-
integrated and secondary fibre mills. An integrated mill is one where pulping and paper
making take place on a single site. Mills that manufacture paper but not pulp are termed non-

integrated mills. Mills that use waste paper as their main product are referred to as secondary
fibre mills. The processes involved in paper making are briefly described below.
Pulping is the process by which cellulose fibres are recovered from the raw material (e.g.
wood). This includes:
• Debarking;
• Mechanical pulping;
• Chemical pulping.
The processed pulp is then made in paper via the following stages:
• Bleaching;
• De-inking (wastepaper only);
• Steam raising;
• Cooling;
• Process.
Typical water consumption figures for various mills and processes are given in Tables A1.58
to A1.59.
Table A1.58 Water consumption from a survey of paper mills in South Africa
Mill Production Water intake Water consumption

3 3
(tonnes/day) (m /day) (m /tonne)
1 650 44,840 70
2 1,200 39,320 33
3 290 20,000 69
4 900 122,400 136
5 145 2,700 19
6 120 1,000 8
7 900 32,600 36
8 22 230 11
9 185 145 0.8
10 15 740 49
11 148 2,800 19
Note: Mills 1 to 5 are integrated mills with pulp and paper making
Mills 6 to 11 are non-integrated i.e. paper making only
Table A1.59 Water consumption for pulp and paper making in Europe
Comments Water
consumption
3
(m /tonne)
UK average Average over whole industry 35.0

UK well designed Production of paper from waste paper 3.7
Netherlands average Average of 18 paper mills 8.5
Newsprint Modern system with closed white water 13.6
system
Source: Reference A1.3 and A1.29

Table A1.60 provides typical water consumption figures for a number of different paper types
based on data collected worldwide.
Table A1.60 Typical water consumption figures for various paper types
3
(m /tonne)
Minimum Maximum Mean
Packing board 1 50 14
Corrugated casing 2 46 23
Newsprint - - 29
Printings/Writings 2 68 32
Tissue paper 44 75 60
Note: Water consumption is directly linked to paper quality and starting material (e.g. wood pulp or
wastepaper)
A1.3.28 Quarries
There are two major types of quarry. These are:
• Hard rock e.g. basalt, graphite, limestone;

• Sand and gravel.
Depending on the geology, depth and location of the quarry, dewatering is often required to
prevent flooding of the operation. Similar to mines quarries may be a net producer of water.
The type of quarry, the quality and quantity of the deposit, together with the local
hydrogeology will all effect the water requirements. As a consequence it is not possible to
determine water use figures for a particular type of quarry. Each quarry should be assessed
on its own merits. The process efficiency of water usage will be dependent upon the
efficiency of collection treatment (usually settling lagoons) and re-use of water.
A1.3.29 Semiconductor wafer fabrication
In order to produce semiconductors ultra pure water is required. This can mean that the actual
demand on the water supply system is 25% higher than the quantity of ultra pure water used.
This is due to the fact that there are losses in the water purification steps. The water
consumption figures given in Table A1.61 are based on data collected in the USA.
Table A1.61 Ultra pure water consumption for semiconductor wafer

3 2
(m /m )
Range 56 to 345
The main ways in which water use can be reduced are:
• Re-use of water in cooling towers and air scrubbers;

• Re-cycling of rinse water;
• Using more efficient rinse techniques.

A1.3.30 Steel manufacturing
Table A1.62 presents data for water consumption for steel manufacture. The data was
collected for a number of steel manufacturing operations in the UK. The wide range of
figures is due to a number of factors including the age, size and location of the plants together
with the grade and the final form of the steel product.
Table A1.62 Water consumption for steel manufacture

3
(m /tonne)
Range 3 to 62
Water use efficiency can be improved by:
• Recycling cooling water;

• Cascade use of water.
A1.3.31 Sugar cane refining
The processing of sugar cane takes place in two stages, milling and refining. Sugar cane
contains approximately 70% water by mass. The main process in both a mill and a refinery is
to extract sugar crystals from the solution. Theoretically utilising the water in the cane can
satisfy all the water requirements of a sugar processing plant. However, for practical reasons
sugar plants usually take their water from other surface or groundwater sources. Water from
these sources is usually used for cooling or domestic consumption. Table A1.63 below gives
water consumption figures for sugar mills in South Africa.
Table A1.63 Water consumption for sugar mills in South Africa
Sugar cane processed Water consumption

3
(tonnes/hour) (m /tonne)
Range 100 to 600 0.3 to 1.0

Mean 250 0.6
Figures for sugar mills in the USA indicated water consumption range of 0.4 to 1.3 m3 per
tonne of cane crushed (Reference A1.3).
A1.3.32 Textile manufacturing
The textile industry is extremely diversified, producing a wide variety of end-products

ranging from processed fibres to woven materials to finished garments. In terms of water
consumption extremes in the industry are represented by factories that employ a large number
of people but carry out mainly “dry” operations and factories operating as commission dyers
using large volumes of water for these “wet” processes. This section provides water
consumption figures for the wet processes. Many of the steps that use water in the textile
industry are common to a number of different fibre types. The main water using steps in
textile manufacture are described briefly below.
• Sizing – The coating of yarns with a film of agents to provide protection during weaving;
• Singeing – The protruding fibres are burnt to give a smooth finish;

• Desizing – The size solution is removed;
• Scouring – The fabric is cleaned using alkali solution to remove inherent or added
impurities in raw fibre or fabrics;
• Bleaching – The fabric is decolourised before printing;
• Mercerising – This process improves the dyeability of cotton goods;
• Dyeing – The addition of colour to the textile;
• Printing – There are a variety of techniques used including rotary screen, direct,
discharge, resist, ink-jet and heat transfer;
• Finishing – Chemical and mechanical treatments performed on the yarn or fabric to
improve appearance, texture or performance.
Typical water consumption for the various processes gathered from a survey carried out in
South Africa are given in Table A1.64.
3
Table A1.64 Water consumption for textile manufacturing (m /tonne)
Process step Low value High value
Singeing 1 2
Sizing – woven goods 1 2
Desizing 5 15
Scouring 4 25
Mercerising 6 17
Bleaching 10 100
Dyeing 9 330
Printing 3 33
Washing-off prints 29 505
Finishing 13 134
Overall (typical)* 100 300
*Note: For some types of fibres or product, one or more of the processing steps indicated may not be
carried out at all, in which case the lower case would be zero. The “overall (typical)” water consumption
values given are thus not obtained by summing the individual values in the above table.
Table A1.65 gives typical water consumption figures for a range of fabrics based on data
collected world-wide.

3
Table A1.65 Water consumption for textile manufacturing (m /tonne)
Fibre Process Consumption
Cotton De-sizing 3 to 9
Scouring or kiering 26 to 43
Bleaching 3 to 124
Mercerising 232 to 308
Dyeing 8 to 300
Wool Scouring 46 to 100
Dyeing 16 to 22
Washing 334 to 835
Neutralisation 104 to 131
Bleaching 3 to 22
Nylon Scouring 50 to 67
Dyeing 17 to 33
Acrylic Scouring 50 to 67
Dyeing 17 to 33
Final scour 67 to 83
Polyester Scouring 25 to 42
Dyeing 17 to 33
Final scour 17 to 33
Viscose Scouring and dyeing 17 to 33
Salt bath 4 to 13
Acetate Scouring and dyeing 33 to 50
Table A1.66 provides typical water consumption figures for various textile processes in the
USA.
3
Table A1.66 Water consumption for textile manufacturing in the USA (m /tonne)
Processing sub- Minimum Median Maximum

category
Wool 111.0 284.6 658.4

Woven 5.0 113.5 508.2
Knit 20.0 83.5 377.2
Carpet 8.3 46.7 162.7
Stock/Yarn 3.3 100.1 558.3
Non-woven 2.5 40.1 82.6
Felted fabrics 33.4 212.8 933.0
A1.3.33 Vehicle manufacturing
The water consumption figures for vehicle manufacture given in Table A1.67 are based on a
variety of data collected world-wide. The main water uses during the construction of motor
vehicles is for engine test bed cooling and during production. Water re-use options are to re-
cycle cooling water and to utilise cascade rinsing in metal finishing operations.

Table A1.67 Water consumption for vehicle manufacture
3
(m /vehicle or engine)
Passenger vehicles Commercial vehicles Engine

(i.e. cars) (i.e. lorries, buses) manufacture
Typical range 2.6 to 8.0 12.0 to 16.0 0.04

A1.3.34 Wallpaper manufacturing
Table A1.68 provides information on water consumption for wallpaper. The figures are based
on data collected for the UK. The figures in Table 6.68 do not include the water used in
manufacturing the paper.
Table A1.68 Water consumption for wallpaper manufacturing

(litres per roll of product)
Water use efficiency can be improved via:
• Re-cycling of cooling water;

• Introduction of cleaning methods and coating make-up procedures;
• The use of an efficient steam system including the return of condensate and reducing
leaks.
A1.3.35 Wine production
Tables A1.69, A1.70 and A1.71 give details of water consumption for wine making, distilling
and bottle washing. The information is based on a survey of South African wineries carried
out in the early 1990s. The variations in water consumption shown in the tables are to a
certain extent affected by factors such as the scale of the operation and the types of packaging
and bottling employed. However, there is no evidence to suggest that larger wineries or
distilleries are more water efficient than smaller ones.
Table A1.69 Water consumption for wine making in South Africa
Production Water intake Water consumption

3 3
(tonnes/year) (m /day) (m /tonne)
Range 8,000 to 22,000 12,000 to 33,000 0.7 to 3.8

Average 14,000 25,200 1.8

Table A1.70 Water consumption for distilling in South Africa
Production Water intake Water consumption

3 3
(tonnes/year) (m /day) (m per litre of absolute
alcohol)
Range 1,250 to 35,000 1,000 to 11,500 18 to 62

Average 14,000 1,400 35
Table A1.71 Water consumption for wine bottle washing in South Africa
Water consumption (litre per bottle)
Range 0.3 to 2.1

Average 1.5
Water use can be reduced by:
• Use high pressure low volume cleaning equipment;

• Sweeping up loose dirt before floors are hosed down;
• Clean cooling towers regularly.
A1.4 References
A1.1 Water Research Commission, South Africa (1987) Water and wastewater
management in the soft drink industry, WRC Project No. 145 TT 35/87
A1.2 Environmental Technology Best Practice Programme UK (1998) Water use in the soft
drink industry EG126
A1.3 Environment Agency England and Wales (2000) Optimum use of water of water for
industry and agriculture: Best practice manual, Research and Development
Technical Report W254
A1.4 Personal communication with Jeff Broome, Ncude Borrow, Zimbabwe (2002)
A1.5 IUCN (2000) Water demand and management: Towards developing effective
strategies for southern Africa IUCN – The World Conservation Union (Regional
Office for South Africa)
management in the sorghum malt and beer industries, WRC Project No. 145 TT
39/90
management in the malt brewing industry, WRC Project No. 145 TT 29/89
A1.8 South African Breweries Corporate Citizen Review (2002)

http://www.sabplc.com/pdfs/reducingwater.pdf

management in the pelagic fishing industry, WRC Project No. 97 TT 28/87
management in the dairy industry, WRC Project No. 145 TT 38/89
A1.11 Ikhu-Omoregbe and Masiiwa H. (2002) A postal survey of effluent generation and
disposal in the Zimbabwean dairy industry Water SA Vol 2 April 2002 pp 179 to 182
A1.12 Water Research Commission, South Africa (1987) Guide to water and wastewater
management in the fruit and vegetable processing industry, WRC Project No. 96 TT
30/87
management in the edible oil industry, WRC Project No. 145 TT 40/89
management in the red meat industry, WRC Project No. 145 TT 41/89
management in the red meat abattoir industry, WRC Project No. 145 TT 45/90
management in the laundry industry, WRC Project No. 145 TT 42/89
management in the tanning and leather finishing industry, WRC Project No. 145 TT
44/90
management in the metal finishing industry, WRC Project No. 145 TT 34/90
A1.19 Environmental Technology Best Practice Programme (1998) Water use in the metal
finishing industry Guide EG45 UK Government
A1.20 Personal communication (2002) with Chozi Vincent Lungu, The School of Mines,
University of Zambia, Lusaka, Zambia
A1.21 Water Research Commission, South Africa (1990) The water requirements and
pollution potential of South African gold and uranium mines, KV90
A1.22 Hadley, H.K. (1995) Water use guidelines: Mining water use, US Geological Survey
A1.23 Lindsey K.S. (1995) Water use guidelines: Hydroelectric water use, US Geological
Survey
A1.24 Lindsey K.S. (1995) Water use guidelines: Thermoelectric power water use, US
Geological Survey
A1.25 Eskom (2002) Eskom environmental data: Resource usage

http://www.eskom.co.za/Enviro%20data%202002/report01/resource.htm
A1.26 Novak, P., Moffat, A.I.B., Nalluri, C. and Narayanan (1996) Hydraulic structures, E
& F.N. Spon publishers second edition

management in the poultry industry, WRC Project No. 145 TT 43/89
management in the paper and pulp industry, WRC Project No. 145 TT 49/90
A1.29 Environmental Technology Best Practice Programme (1998) Practical water

management in paper and board mills Guide GG11 UK Government
management in the sugar industry, WRC Project No. 145 TT 47/90
management in the textile industry, WRC Project No. 145 TT 50/90
A1.32 Environmental Technology Best Practice Programme (1997) Water use in textile
dyeing and finishing Guide EG98 UK Government
management in the wine industry, WRC Project No. 145 TT 51/90
A1.34 North Carolina Department of Environment and Natural Resources (1998) Water
A1.35 Personal communication (1998) with Muller, J. General Manager of technology

development Abakor, Pretoria
A1.36 Hadley, H.K. (1995) Water use guidelines: Industrial water use, US Geological
Survey
A1.37 Water Research Commission, South Africa (1999) Water reclamation for direct re-
use in urban and industrial applications in South Africa and its projected impact
upon water demands, KV118/99
A1.38 Gumbo, B. Mlilo, S., Broome J., and Lumbroso D. (2002) Industrial water demand
management and cleaner production: A case study of three industries in Bulawayo,
Zimbabwe Third WaterNet Symposium: Integrating Water Supply and Water
Demand for Sustainable Use of Water Resources, Tanzania October 2002
A1.39 Mulenga, S (2002) Assessment of catchment water demand and use, unpublished
thesis for the Department of Civil Engineering, University of Zambia

Contact details:
Darren Lumbroso, HR Wallingford, Howbery Park, Wallingford, Oxon, OX10 8BA, UK

Tel: +44 1491 822217, Fax: +44 1491 826352, Email:
[email protected]
http://www.hrwallingford.co.uk

Handbook For Catchment Water Demand and

Uploaded by

Copyright:

Available Formats

Handbook For Catchment Water Demand and

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Handbook For Catchment Water Demand and

Uploaded by

Copyright:

Available Formats

Handbook for the

in collaboration with the Department for Water Development

Handbook for the assessment of

1.2 Scope of the Handbook 1

1.3 Definitions of water demand and use 2

2. AN INTRODUCTION TO THE PRINCIPLES OF MANAGING

2.2 Roles and responsibilities of catchment managers 7

2.3 Balancing supply and demand 9

2.3.1 The supply side 10

2.3.2 The demand side 10

2.4 The legal framework 11

2.6 Scales and boundary conditions 13

2.7 Issues in water allocation 14

2.8 Efficiency and equity 15

2.10 Water allocation between sectors 17

2.11 Recommendations on water allocation 18

3. ENVIRONMENTAL WATER DEMAND AND USE..................... 20

3.2 Environmental flow assessment methods 20

3.3 Hydrological index methods 22

3.8 Lake level requirements 40

3.9 Methodologies for assessing ecological flow requirements of

4. AGRICULTURAL WATER DEMAND AND USE ....................... 44

4.2 Overview of irrigated agriculture in the SADC region 44

Handbook for the assessment of ii

4.4 Approximate method for calculating irrigation water demand

4.5 Detailed method of estimating irrigation water demand and

4.8 Improving rain fed agricultural production 75

4.10 Case study of improved water use efficiency for an irrigation

4.11 Livestock water use 81

5. RURAL DOMESTIC WATER DEMAND AND USE ................... 85

5.4 Factors affecting rural domestic water demand and use 93

5.6 Direct methods of estimating rural domestic water demand

6. INDUSTRIAL WATER DEMAND AND USE ............................ 107

6.2 Industrial processes water use 107

6.3 Industrial water consumption statistics 109

6.4 Participatory methods for obtaining industrial water use

6.5 Water audits for industry 110

6.7 Case studies of industrial water demand management for

Handbook for the assessment of iv

7. URBAN WATER DEMAND AND USE..................................... 121

7.5.1 Background 137

7.8 References 156

8. FORECASTING WATER DEMAND AND USE........................ 158

8.2 Influences on water demand and use 158

8.3 Criteria for assessing forecasting methods 158

8.9 Forecasting industrial water demand and use 173

8.11 The effect of the HIV/AIDS virus on future water demand in

8.12 Case study for demand forecasting for city of Masvingo in

9. RIVER TRANSMISSION LOSSES........................................... 185

9.2 Nature of river transmission losses 185

9.3 Example of transmission losses from southern Africa 186

Handbook for the assessment of vi

A1 INDUSTRIAL WATER CONSUMPTION LOOK UPTABLES...A-1

A1.1 Background to look up tables A-1

A1.2 Examples of how to use the industrial water consumption

A1.3 Industrial water consumption look up tables A-2

Handbook for the assessment of viii

Handbook for the assessment of ix

Each of these sectors is covered in a separate chapter as detailed below.