Representative and Experimental Basins.: Unesco

Representative and
experimental basins.
An international guide fou research and practice I
Edited by C.Toebes and V. Ouryvaev

‘I
A contribution to the International Hydrological Decade
Unesco
Studies and reports in hydrology 4
TITLES IN THIS SERIES
Co-edition Unesco/IASH
i. The use of analog and digital computers in hydrology. Proceedings of the Tucson Symposium,
vol. 2. I L’utilisation des calculatrices analogiques et des ordinateurs en hydrologie.Actes du
colloque de Tucson, vol. 2.
2. Water in the unsaturated zone. Proceedings of the Wageningen Symposium, vol. 1. / L’eau
dans la zone non saturée. Actes du symposium de Wageningen, vol. 1.
3. Floods and their computation.Proceedings of the Leningrad Symposiurn,August 1967,vol.2. J
Les crues et leur évaluation. Actes du symposiiim de Leningrad, aJût 1967, vol. 2.
Published by Unesco
4. Representative and experimental basins - A n international guide for research and practice
(Willalso appear in French, Russian and Spanish).
5. Discharge of selected rivers of the world, vol. 1. / Débit de certains cours d’eau du monde,
vol. 1.
Published in 1970 by the United Nations
Educational, Scientific and Cultural Organization
Place de Fontenoy, 75 Parìs-P
Printed by Henkes-Holland, Haarlem
0 Unesco 1570
Prinled in îhe Neiherlands
SC NS.68/XX-l/A
Preface
The International Hydrological Decade (IHD)1965-74 was launched by the General

Conference of Unesco at its thirteenth session to promote international co-operation
in research and studies and the training of specialists and technicians in scientific
hydrology. Its purpose is to enable all countries to make a fuller assessment of their
water resources and a more rational use of them as man’sdemands for water constantly
increase in face of developments in population, industry and agriculture. In 1968,
national committees for the Decade had been formed in iW-of Unesco’s 122 Member
States to carry out national activities and to contribute to regional and international
activities within the programme of the Decade. The implementation of the programme
is supervised by a Co-ordinating Council, composed of twenty-one Member States
selected by the General Conference of Unesco,which studiesproposals for developments
of the programme,recommends projects of interestto all or a large number of countries,
assists in the developmentof nationaland regional projects and co-ordinatesinternational
co-operation.
Promotion of collaboration in developing hydrological research techniques,diffusing
hydrological data and planning hydrological installations is a major feature of the
programme of the IHD which encompasses all aspects of hydrological studies and
research. Hydrological investigations are encouraged at the national, regional and
international level to strengthen and to improve the use of natural resources from
a local and a global perspective. The programme provides a means for countries well
advanced in hydrological research to exchange scientific views and for developing
countries to benefit from this exchange of information in elaborating research projects
and in implementing recent developments in the planning of hydrological installations.
As part of Unesco’s contribution to the achievement of the objectives of the IHD
the General Conference authorized the Director-General to collect, exchange and
disseminate information concerning research on scientific hydrology and to facilitate
Preface
contacts between research workers in this field. To this end Unesco has initiated two
collections of publications: Studies and Reports in Hydrology, and Technical Papers
in Hydrology.
The collection Studies and Reports in Hydrology is aimed at recording data collected
and the main results of hydrological studies undertaken within the framework of the
Decade as well as providing information on research techniques. Also included in the
collection will be proceedings of symposia. Thus, the collection will comprise the
compilatibn of data, discussions of hydrological research techniques and findings, and
guidance material for future scientific investigations.It is hoped that the volume will
furnishmaterial of both practical and theoretical interestto hydrologists and governments
participating in the IHD and respond to the needs of technicians and scientists concerned
with problems of water in all countries.
The responsibility for the choice and presentation of facts and for opinions and views expressed
lies with the organizations and authors cited in the foreword to each publication in the collection.
The designations employed and the presentation of the material do not imply the expression
of any opinion whatsoever on the part of Unesco concerning the legal status of any country
or territory,or of its authorities,or concerning the delimitations of the frontiers of any country
or territory.
Contents
Foreword 17
List of contributors 19
1 Introduction
1.1 Scope and purpose 21
1.2 Definition of representative and experimental basins 22
1.2.1 Representative basins 22
1.2.2 Experimental basins 23
1.3 Purposes of representative and experimental basins 23
1.3.1 Representative basins 23
1.3.2 Experimental basins 24
1.4 Survey of research needs 24
1.4.1 Staffing for basins 25
1.4.2 Basic recommendations on standardization of methods of observation,
instrumentation and data processing 25
1.4.3 Data-reporting methods 21
1.4.4 Research co-operation and research observation programmes 21
I .5 Terminology 30
1.6 Measurement units and symbols 36
References 41
2 Selection and organization of basin networks

2.0 General 43
2.1 Use of maps 43
2.2 Selection of hydrological regions 44
2.2.1 Selection of hydrological regions in arcas where no detailed hydrological
data are available 45
2.2.1.1 Example in Brazil 46
2.2.1.2 Example in N e w Zealand 41
2.2.2 Selection of hydrological regions in areas where detailed hydrological data
are available 48
Contents
2.2.2.I Example in the U.S.S.R. 48

2.2.3 Delineation of soil-vegetationcomplexes 51
2.3 Selection of representative basins 51
2.4 Selection of experimental basins 53
2.4.1 Selection of run-offplots 54
2.5 Analysis of a basin’s representativeness 55
References 56
3 Planning of observations according to the

research objectives
3.1 General 51
3.1.1 Observational programme for representative basins 58
3.1.1.1 Observational programme for representativebasins used for fundamental
research 58
3.1.1.2 Observational programme for representativebasins used for the study of
the effect, on the hydrological regimen, of natural changes (benchmark
and vigil basins) 58
3.1.1.3 Observational programme for representativebasins used for hydrological
prediction 59
3.1.1.4 Observational programme for representative basins used for extension
of records 59
3.1.2 Observational programme for experimental basins 59
3.2 Observational asprcts of principal hydrological elements 60
3.2.1 Precipitation 60
3.2.2 Interception 61
3.2.3 Snow cover 61
3.2.3.1 Snow survey 62
3.2.3.2 Snowmelt 62
3.2.4 Condensation 63
3.2.5 Evaporation 63
3.2.5.1 Evaporation from the water surface 63
3.2.5.2 Evapotranspiration 64
3.2.5.3 Evaporation from snow 64
3.2.6 Surface water 65
3.2.7 Subsurface water 67
3.2.7.1 Water in the unsaturated zone 67
3.2.7.2 Water in the saturated zone 67
3.2.7.2.
I Determination of principal characteristics of aquifers 68
3.2.8 Infiltration 68
3.2.9 Glaciers 69
3.2.10 Erosion and sedimentation 69
3.2.11 Quality of water 70
3.2.12 Ice phenomena in streams 70
3.2.13 Climatological data and energy balance 71
3.2.13.1 Climatological observations 71
3.2.13.2 Measurement of energy-balancecomponents 71
3.3 Planning observations for the study of the effect,on the hydrological regimen,
of a natural and/or cultural change 72
3.3.1 The study of the influence of forest on the hydrological regimen 73
3.3.1.1 Selection of basins 74
3.3.1.2 Programme of observations 74
References 75
Contents
I 4 Methods of observation and instrumentation

General requirements 76
~ 4.1.1
4.' Sampling techniques 77
4.I .1.1 Sampling in time 77
4.1.1.2 Sampling in space 78
4.2 Climate 79
4 2.1 Precipitation 79
4.2.1.1 General 79
4.2.1.2 Networks 79
4.2.1.2.1 Network reappraisal 82
4.2.1.3 Precipitation gauges 82
I 4.2.1.3.1 Recording raingauges 82
82
4.2.1.3.2 Non-recording precipitation gauges
I 4.2.1.3.3
4.2.1.3.4
Errors in precipitation gauges
Installation methods
83
84
4.2.1.3.5 Selection of precipitation gauges 84
4.2.1.3.6 Measurement of dew and fog 84
4.2.2 Snow cover 84
4.2.2.1 Snowmelt 89
4.2.3 Interception of precipitation by vegetation 89
4.2.3.1 Interception of rain 89
4.2.3.1.1 Forest vegetation 69
4.2.3.1.1.1 General 89
4.2.3.1.1.2 Variables 90
4.2.3.1.I .3 Methods and instrumentation 91
4.2.3.1.1.3.1Gross rainfall,91; 4.2.3.1.1.3.2
Throughfall, 91;
4.2.3.1.1.3.3Stem flow, 91; 4.2.3.1.1.3.4
Litter interception,91;
4.2.3.1.1.3.5Sampling intensity,91
4.2.3.1.2 Herbaceous vegetation 92
4.2.3.1.2.1 General 92
4.2.3.1.2.2 Sampling methods 92
, 4.2.3.1.2.3 Gross interception loss 93
4.2.3.1.2.4 Net interception loss 93
4.2.3.2 Interception of snow 93
4.2.3.3 Interception of dew and fog 94
4.2.3.4 Future research 95
I
4.2.4. Evaporation pans 95
4.2.4.2 Evaporimeters and lysimeters 98
4.2.4.2.1 Evaporimeters 98
4.2.4.2.1.1 General 98
4.2.4.2.1.2 Soil evaporimeter GGI-500-100 99
I 4.2.42.1.3 Soil evaporimeter GGI-500-50 99
4.2.42.1.4 The small-typehydraulic soil evaporimeter 99
4.2.4.2.1.5 The big hydraulic evaporimeter (BGI) 99
4.2.4.2.1.6 Forest hydraulic evaporimeter (large type) 101
4.2.4.2.1.7 Soil-weighingevaporimeter 102
4.2.4.2.1.8 Soil-weighingevaporimeter (large) 103
4.2.4.2.1.9 Evaporimeter for swamps GGI-B-1000 103
I. 10
4.2.4.2. Lysimeter-compensating evaporimeter 103
4.2.4.2.1.11 Soil raingauge 103
4.2.4.2.1.12 Methods of observation of evaporation from the soil surface 104
4.2.4.2.1.13 Evaporation from snow 104
4.2.4.2.2 Lysimeters 105
4.2.4.2.2.1 Purpose 105
4.2.4.2.2.2 Principle of construction 105
4.2.4.2.2.3 Construction of the weighable lysimeter 107
~~
Contents
4.2.4.2.2.4 The non-weighable lysimeter 108

4.2.4.2.2.5 Lysimeters of the U.S.S.R. Hydrometeorological Service 108
4.2.4.2.2.6 Major disadvantage of the lysimeter 109
4.2.5 Other climatic observations, including energy balance 1 o9
4.2.5.1 Climate stations 109
4.2.5.2 Climate station requirements 110
4.2.5.2.1 Instrumentation 110
4.2.5.2.1.1 Fully equipped base station 110
4.2.5.2.1.2 Auxiliary climate station 110
4.2.5.2.2 Location 110
4.2.5.3 Instrumentation and observational techniques 111
4.3 Surface water 112
4.3.1 Objectives 112
4.3.2 Streams and gauging-site selection 112
4.3.2.1 The stage-measurement cross-section 112
4.3.2.2 The flow-measurement cross-section 112
4.3.3 Natural controls (high flow-low flow) 113
4.3.4 Precalibrated devices-weirs, flumes and orifices 113
4.3.4.1 Experimental basins 113
4.3.4.2 Representative basins 113
4.3.4.3 Desiderata for measuring structures i 14
4.3.4.4 Permanency of rating 114
4.3.4.5 Stage inaccuracies 114
4.3.4.5.1 Precalibrated artificial structures 115
4.3.5 Flow measurement-current meter, chemical, miscellaneous 127
4.3.5.1 Current-meter limitations 130
4.3.5.2 Chemical-gauging methods 130
4.3.5.2.1 Salt-velocity method 130
4.3.5.2.2 Salt-dilution method 130
4.3.5.2.2.1 Fluorescent dye-dilution method 130
4.3.5.2.3 Integration or ‘gulp’ method 130
4.3.5.2.3.1 Use of dyes 130
4.3.5.3 Miscellaneous methods 131
4.3.5.3.1 Head-rod method 131
4.3.5.3.2 Slope-area method 131
4.3.5.3.3 Volumetric gauging 131
4.3.5.3.4 Optical current-meter gauging 131
4.3.5.3.5 Measurement of îlow through culverts 131
4.3.5.3.6 Measurement of flow through contracted openings 131
4.3.5.3.1 Float measurements 131
4.3.6 Stage measurement 132
4.3.6.1 Staff gauges 132
4.3.6.2 Automatic stage recording 132
4.3.6.3 Float-recorderstations 132
4.3.6.4 Punched-tape recorders 132
4.3.6.5 Pressure-bulb recorders 134
4.3.6.6 Servo-manometer 134
4.3.7 Operation, maintenance and accuracy requirements of gauging stations 135
4.3.1.1 Winter operation 135
4.3.1.2 Effect of ice on the stage discharge relation 135
4.3.1.3 Freezing of stilling well 135
4.3.7.4 Summer or tropical operation 136
4.3.1.5 Blockage of control 136
4.3.7.6 Float-recorder structures 136
4.3.1.1 Intake pipes 136
4.3.1.8 Flushing 136
4.3.1.9 Accuracy requirements 131
4.3.7.9.1 Accuracy of calibrated structures 131
4.3.7.9.2 Accuracy of field-rated controls 137
Contenis
4.3.8 Run-off plots 137

4.4 Subsurface water 140
4.4.1 Water in the unsaturated zone 140
4.4.1.1 Soil moisture 140
4.4.1.1.1 Networks 141
4.4.1.1.2 Measurement methods and equipment 142
4.4.1.1.2.1 Electrical resistance units 142
4.4.1.1.2.2 Neutron-scattering technique 143
4.4.1.1.2.2.1Access tubing, 143; 4.4.1.1.2.2.2Measurement site, 143;
4.4.1.1.2.2.3Calibration, 143
4.4.1.1.3 Frequency of measurement 143
4.4.1.1.4 Laboratory methods 144
4.4.2 Water in the saturated zone 144
4.4.2.1 Hydrogeological assessment of a basin 144
4.4.2.2 Instrumentation 145
4.4.2.2.1 Geological subsurface techniques 145
~ 4.4.2.2.2 Geophysical subsurface techniques 145
4.4.2.3 Distribution of measuring points 147
4.4.2.3.1 Accuracy 148
4.4.2.3.2 Additional requirements 148
4.5 Infiltration 148
4.5.1 Infiltrometers i 48
4.5.1.1 Flood-type infiltrometers 149
4.5.1.2 Ring infiltrometers 149
4.5.1.3 Cylinder or tube infiltrometers 149
4.5.1.4 Rainfall-simulator infiltrometers 150
4.5.1.5 Portable rainfall-simulator infiltrometers 153
4.5.1.6 Operating sprinkler infiitrometers 154
4.6 Measurement of phytomorphological characteristics 156
4.6.1 General 156
4.6.2 Measurement methods of surface vegetation 156
4.6.2.1 Aerial photographs and field-sample plots 156
4.6.2.2 Sampling herbaceous species and low shrubs 157
4.6.2.3 Sampling trees and large shrubs 158
4.6.3 Special observations for evaporimeters and lysimeters 159
4.6.4 Root depth and root density 160
4.7 Soil physical measurements 160
4.7.1 Texture and structure 161
4.7.2 Water relations i 62
4.7.2.1 Moisture retention 162
4.7.2.2 Moisture-retention capacity and saturation-moisture capacity 163
4.7.3 Water movement 163
4.7.4 Vapour versus liquid movement 164
4.8 Measurement of soil frost and thaw 164
4.8.1 General 164
4.8.2 Measurement of soil frost and thaw 165
4.8.3 Effect of exposure on freeze-thaw cycles 166
4.9 Erosion and sedimentation 166
4.9.1 Scope and purpose 166
4.9.2 Erosion studies 167
4.9.2.1 Erosional processes 167
4.9.2.1.1 Hillslope erosional processes 167
4.9.2.1.2 Erosion studies on run-off plots 169
4.9.2.1.3 Stream-channel erosional processes 169
4.9.2.2 Measurement of erosion 169
4.9.2.2.1 Measurement techniques 169
4.9.2.2.1.1 Hillslope erosion 169
4.9.2.2.1.2 Stream-channel erosion 170
4.9.3 Sedimentation studies 171
-
Contents
4.9.3.1 Techniques of sediment measurement 172

4.9.3.1.1 Measurement of suspended-sediment discharge 172
4.9.3.1.I.1 Sampling equipment 1 72
4.9.3.1.1.2 Sampling procedure 172
4.9.3.1.2 Measurement of bed load 172
4.9.3.1.2.1 Samplers 172
4.9.3.1.2.2 Gauging methods 173
4.9.3.1.2.3 Bed-material observations 174
4.9.3.1.3 Determination of concentration and particle size 174
4.9.3.1.3.I Filtration method 174
4.9.3.1.3.2 Methods for determining particle-size distribution 174
4.9.3.1.4 Sedimentation surveys in small reservoirs 174
4.10 Measurement of water quality 175
4.10.1 Purpose of water-quality measurements 175
4.10.2 Frequency of measurements 175
4.10.3 Location of sampling 175
4.10.4 Sampling methods 176
4.10.5 Chemical and physical analysis 176
4.11 Use of radioactive tracers 176
4.11.1 Sediment-transport measurements 177
4.1 1.1.1 Space-integration method 177
4.11.1.2 Time-integration method 177
4.1 1.1.3 Equipment required 117
4.11.1.4 Suspended-sediment measurements 177
4.11.2 Measurement of the water content of the snow pack 177
4.11.3 Soil-moisture and density measurements I 78
4.11.4 Ground-water tracing, velocity and direction measurements, determination
of the effective porosity 178
4.1I .4.1 Simple point-to-point tracing experiments for the determination of the
hydraulic continuity and direction and velocity of ground water 179
4.11.4.2 Determination of the effective porosity and dilution techniques for the
determination of the direction and velocity of ground water 179
4.11.5 Environmental isotope techniques 179
4.11.5.1 Deuterium and oxygen-I8 180
4.11.5.2 Tritium and carbon-14 181
4.11.5.3 General 183
4.12 Minimum equipment i 83
4.13 Location of instruments and equipment in a basin 186
References 186
5 Data processing and publication

5.0 General 196
5.I Mapping of representative and experimental basins 196
5.1.1 Topographical mapping 197
5.1.1.1 Techniques 197
5.1.1.2 Features 197
5.1.1.2.I Contours 197
5.1.1.2.2 Basin divide 197
5.1.1.2.3 Drainage net and cultural features 197
5.1.1.3 Construction of maps 198
5.1.1.3.1 Base maps 198
5.1.1.3.2 Size of maps 198
5.1.2 Geological and hydrogeological mapping I 98
5.1.2.1 Mapping of surface geology 200
5.1.2.1.1 Techniques 200
Contents
5.1.2.1.2 Features 200

5.1.2.1.3 Cartography 200
5.1.2.2 Subsurface mapping 202
5.1.2.2.1 Techniques 202
5.1.2.2.1.1 Hydrogeological mapping 202
5.1.2.2.1.2 Geological engineering mapping 203
5.1.2.2.2 Features 203
5.1.2.2.3 Cartography 203
5.1.3 Pedological mapping 204
5.1.3.2 Features of a soil map 204
5.1.4 Vegetation maps 207
5.1.4.1 Vegetation surveys 207
5.1.5 Geomorphologicalmapping 209
5.1.6 Land-inventory mapping 209
5.1.6.2 Features 209
5.1.6.3 Cartography 209
5.1.7 Aerial and terrestrial photography 210
5.1.7.1 Terrestrial photography 210
5.1.7.2 Aerial photography 21 I
5.1.8 Cartographical description of physiographical characteristics 21 1
5.1.9 Inventory of observations and research made on representative and ex-
perimental basins 212
5.2 Recommendations on routine data-processingmethods 212
5.3 Climatic data 213
5.3.1 Precipitation 213
5.3.1.1 Consistency of records 283
5.3.1.2 Point rainfall record 214
5.3.1.3 Routine methods of determining mean basin rainfall 214
5.3.1.4 Other methods of determining mean basin rainfall 216
5.3.1.5 Other methods of processing precipitation data 216
5.3.2 Snow cover 216
5.3.2.1 Errors in snow data 217
5.3.2.2 Tests for consistency 218
5.3.3 Interception of precipitation by vegetation 218
5.3.4.1 Water-balancemethod 218
5.3.4.2 Pan evaporation 219
5.3.4.3 Energy-balance method 219
5.3.4.4 Aerodynamic method 220
5.3.4.5 Empirical equations 220
5.3.4.6 Evaporation from soil and snow cover 22 L
5.3.4.7 Actual evapotranspiration 222
5.3.5 Other climatic observations, including energy balance 222
5.3.5.1 Wind 223
5.3.5.2 Temperature 223
5.3.5.3 Humidity 223
5.3.5.4 Radiation 223
5.3.5.5 Energy balance 223
5.4 Surface water 225
5.4.1 General 225
5.4.2 Charts and tapes 225
5.4.3 Tabulation of time and stage heights 225
5.4.4 Stage-dischargerelations and ponding corrections 226
5.4.5 Mean discharge 227
5.4.6 Run-off calculations 227
5.4.7 Approximate checks on validity of streamflow 229
5.4.8 Lakes and ponds 229
Contents
5.5 Subsurface water 229

5.5.1 Water in the unsaturated zone 229
5.5.1.1 Soil moisture 229
5.5.2 Water in the saturated zone 230
5.5.2.1 General considerations 230
5.5.2.2 Errors of measurement 230
5.5.2.3 Methods of presentation of data 231
5.5.2.3.1 Ground-waterlevels 231
5.5.2.3.2 Fluctuation diagrams 231
5.5.2.3.3 Ground-watercontour maps 232
5.6 Erosion and sedimentation 233
5.6.1 Sediment-ratingcurves 233
5.6.2 Calculation of sediment yield 233
5.6.3 Relation to deposits in reservoirs 235
5.6.4 Interpretation of results 235
5.7 Water quality 235
5.8 Automated processing of data 236
5.8.1 Automated processes 236
5.8.1.1 Digital recorder 236
5.8.1.2 Pencil follower 236
5.8.1.3 Punchcard system 236
5.8.1.4 Computer programmes 238
5.8.1.4.1 Digital recorder programme 238
5.8.1.4.2 Examples of programmes available 239
5.8.2 Some considerations regarding data automation 240
5.9 Data storage and retrieval 240
5.9.1 General considerations 240
5.9.2 Microfilm aperture punchcards 241
5.10 Publication of summaries 241
5.10.1 Purpose of publishing 241
5.10.2 Publication requirements 243
5.10.3 Frequency of publication 260
5.10.4 Contents and format 260
5.10.5 Publication of summaries 260
References 260
6 Analysis techniques and interpretation of research results

6.1 General 264
6.1.1 Basin characteristics 264
6.1.1.1 Vegetation characteristics 265
6.1.1.2 Geomorphological characteristics 266
6.1.1.2.1 Area elevation curve- median elevation 266
6.1.1.2.2 Maximum and minimum elevations 266
6.1.1.2.3 Aspect 266
6.1.1.2.4 Maximum valley side slope 266
6.1.1.2.5 Mean slope curve 266
6.1.1.2.6 Hypsometric curve 267
6.1.1.2.7 Slope index 268
6.1.1.2.8 Drainage characteristics 269
6.1.1.3 Pedological characteristics 270
6.1.1.4 Hydrogeological characteristics 271
6.1.2 Analysis of climatic characteristics and energy balance 272
6.1.2.1 Definition of the climatic characteristics of a basin 272
6.1.2.2 Study of variations of climatic characteristics in relation to hydrological
characteristics or processes 272
Contents
6.1.2.2.1 Precipitation 273

6.1.2.2.2 Snowmelt 274
6.1.2.2.3 Temperature 274
6.1.2.2.4 Solar radiation 274
6.1.2.2.5 Humidity 274
6.1.2.2.6 Evaporation 274
1 6.1.2.3 Energy balance 274
I 6.1.3 Surface water 274
275
6.1.3.I Standard hydrological characteristics
6.1.3.2 Various forms of flow, recession studies 275
6.1.3.2.1 Master recession curves 276
6.1.3.2.2 Recession equations 277
6.1.3.2.3 Storage relations 277
6.1.3.3 Hydrograph analysis 277
6.1.3.3.1 Flow separation 277
6.1.3.3.2 Miscellaneous characteristics defining the hydrograph 280
6.1.3.4 Volume of flow (surface or direct run-off) 28 1
6.1.3.5 Infiltration analysis 281
6.1.3.5.1 General methodology 282
6.1.3.5.2 Sprinkling-plotanalysis 282
6.1.3.5.3 Run-off plot analysis 284
6.1.3.5.4 Natural basin analysis 284
6.1.3.5.5 Natural basin analysis using a flow-limit curve 285
6.1.3.5.6 Natural basin analysis by the time-condensationmethod 287
6.1.3.5.7 Infiltration equations 287
6.1.3.5.8 Storage-flowrelationships 289
6.1.4 Subsurface water 291
6.1.4.1 General 291
6.1.4.2 Principles of subsurface flow 29 1
6.1.4.3 Unsaturated-flowanalysis 293
6.1.4.3.1 Types of infiltration 293
6.1.4.3.2 Infiltration at a constant rate, capillary rise 293
6.1.4.3.3 Ponded rainfall infiltration 296
6.1.4.3.4 Ponding infiltration 298
6.1.4.4 Saturated-flowanalysis 301
6.1.4.4.1 Flow analyses from ground-water contour maps 301
6.1.4.4.2 Flow analyses based on piezometric heads near free water surfaces 303
6.1.4.4.3 Flow analyses in drained areas 307
6.1.5 Erosion and sedimentation 309
6.1.5.1 Analysis of study data 309
6.1.5.2 Relation of basin area to sediment yield 309
6.1.5.3 Relation of relief and length to sediment yield 310
6.1.5.4 Relation of drainage density to sediment yield 310
6.1.5.5 Other basin characteristics affecting sediment yield 311
6.1.6 Water quality 312
6.2 Relation between elements of the hydrological cycle and between these and
basin characteristics in approximately stationary condition 312
6.2.1 General 312
6.2.2 Time variability of basins 313
6.2.3 Uncertainty as regards system inputs and outputs 313
6.2.4 Non-linearityof hydrological processes 313
6.2.5 Types of model 313
6.2.5.1 Linear normal models 315
6.2.5.2 Graphical analysis 315
6.2.5.3 Models with a central linear element (unit hydrograph) 317
~ 6.2.5.3.1 General 317
6.2.5.3.2 Concept of unit hydrograph 317
6.2.5.3.3 Determination of unit hydrograph 318
6.2.5.3.4 Use of unit hydrographs 319
Contents
6.2.5.4 Non-linear functional models (general non-linear analysis) 320

6.2.5.5 Conceptual models (general synthesis) 321
6.2.5.5.1 Automatic parameter adjustment 323
6.3 Natural and cultural changes 324
6.3.1 Graphical analysis and linear normal models 324
6.3.1.I Test of representativeness of period of study 325
6.3.1.2 Single-basin technique 325
6.3.1.3 Comparative-basin technique 321
6.3.2 Hydrological characteristics 331
6.3.2.1 Example of working hypothesis 332
6.3.3 Other mathematical models 332
6.4 Water-balance studies 335
6.5 Translation of results to other basins 336
6.5.1 Flood prediction 336
6.5.2 Long-range prediction of streamflows 338
6.5.3 Short-term prediction of streamflows 338
6.5.4 S u m m a r y of prediction techniques 338
References 34I
Bibliography 341
Foreword
Since the inception of the IHD programme,it has been apparent that the development
of hydrological research should include the perfecting of research methodology on
specially chosen and equipped basins. Such basins, designated ‘representative’ or
‘experimental’basins, are of great value for hydrological research and for the training
of research workers when the standards of organization and functioning comply with
clearly defined research aims.
A Symposium on Representative and Experimental Basins was organized by Unesco
in collaboration with the Hungarian Government with the participation of the Inter-
national Association of Scientific Hydrology (IASH) in Budapest in September 1965
in order to study a wide range of experiments and the results obtained,so that general
guidelines could be formulated for these basins. These guidelines are intended mainly
for countries wishing to establish such research basins,but are useful also for improving
the effectiveness of research already in progress.
The Co-ordinatingCouncil ofthe I H D formed an ad hoc Working Group to draw up,
from the conclusions of the symposium,the principles for organizing representative
and experimental basins. This Working Group was asked to analyse the results of an
investigation undertaken on the current state of hydrological research being conducted
or contemplated in such basins in various countries. The growth of projects related
to the creation of representative and experimental basins in the countries participating
in the Decade, and the importance of independent experiments performed in several
countries,led theWorking Group to propose to the Co-ordinatingCouncilthe preparation
of a guide for international research and practice in representative and experimental
basins, based on existing documentation.
This technical guidance material outlines the methodological principles for the
organization and functioning of representative and experimental basins according to
the specific aim of the research undertaken.In order that it could be used for the setting up
of IHD national programmes,it was necessary to prepare the guide as soon as possible.
With this aim in view, the Co-ordinating Council, at its second session in April 1966,
established a permanent Working Group, which had as its main task the preparation
of this guide, and asked it to nominate from its members a Panel of Experts to
complete this task.
The Panel thus formed met for the first time in October 1966. T w o draft guides were
prepared, including contributions and material received from national committees for
the IHD. One was edited by Mr.Toebes,with Messrs. Hadley and Jacquet as sub-editors,
and the other was edited by Dr.Ouryvaev, and prepared in the U.S.S.R. The two guides
were combined at successive meetings of the chief editors, of the Panel and of the
17
Foreword
Working Group. The publication of the project was endorsed by the Co-ordinating
Council at its third session in May 1967.
The guide comprises six chapters:
The Introduction describes the aims of the publication and the purpose of representative
and experimental basins within the larger framework of hydrological research;
Chapter 2 gives the criteria for selecting sites and for organizing a network of basins;
Chapter 3 deals with the organization and planning of observations according to the
aims of the research being carried out on the basins;
Chapter 4 contains descriptions of methods of observation and types of instruments
to be used for the study of the various parameters of the hydrological cycle;
Chapter 5 gives a wide range of methods for data processing,synthesis,and preparation
for publication;
Chapter 6 deals with analysis techniques and interpretation of the results obtained.
The present publication,thus prepared in a very short period of time,obviously presents
imperfections and gaps. Nevertheless,it is hoped that it will serve as a valuable source
of guidance material, especially for countries in which this type of investigation is just
beginning, and will at the same time offer a framework for the international exchange
of experience in this field of hydrological research.
The Working Group which prepared this first version of the guide has also been
entrusted with the responsibility of collecting additional data and formulating various
improvements to the present text; particularly with reference to Chapters 3 and 6,
both of which are important prerequisites of successful research on representative and
experimental basins.
All IHD national committees are invited to send comments on the contents and to
forward complementary scientific material with a view to perfecting the guide. Such
contributions would be especially welcomed from countries where interesting new
results have been obtained from investigations.This information will be collected by
the Secretariat and passed on to the Working Group.
18
List of contributors
ALLENJr., W.H.,United States Department of Agriculture, Agricultural Research Service,

U.S.A.
AMOROCHO, J., University of California, Davis, U.S.A.
BERNIER, J., Electricité de France, CREC, France.
BLAKE,G.J., Ministry of Works, Water and Soil Division, N e w Zealand.
BLOEMEN,G.,Institute for Land and Water Management, The Netherlands.
BLOK,T.,Provinciale Waterstaat, Gelderland, The Netherlands.
BOCHKOV,A. P.,State Hydrological Institute, U.S.S.R.
BRANSON, F.A.,United States Department of Interior,United States Geological Survey,U.S.A.
COLENBRANDER, H.J., Rijkswaterstaat, The Netherlands.
DAGG, M.,East African Agriculture and Forest Research Organization, Kenya.
DAVIS, G.H.,International Atomic Energy Agency.
DINÇER, T.,International Atomic Energy Agency.
DOOGE, J., University College, Cork, Ireland.
ENGLAND,C. B., United States Department of Agriculture, Agricultural Research Service,
U.S.A.
FOURNIER, F.,Office de la Recherche Scientifique et Technique Outre Mer, France.
HADLEY, R.F.,United States Department of Interior,United States Geological Survey, U.S.A.
HARROLD, L. L., United States Department of Agriculture, Agricultural Research Service,
U.S.A.
HELVEY, J. D.,United States Department of Agriculture, United States Forest Service, U.S.A.
HERAS, R.,Instituto de Hidrología, Madrid, Spain.
HOLTAN, H . M.,United States Department of Agriculture, Agricultural Research Service,
U.S.A.
JACKSON, R. J., Department of Scientific and Industrial Research, N e w Zealand Soil Bureau,
New Zealand.
JACQUET, J., Electricité de France, CREC, France.
JOHNSON, A.I., United States Department of Interior,United States Geological Survey, U.S.A.
KELLY, L. L.,United States Department of Agriculture,Agricultural Research Service, U.S.A.
KOVZEL, A. G.,State Hydrological Institute, U.S.S.R.
KRAYENHOFF VAN DE LEUR,D.A.,University of Wageningen, The Netherlands.
KRESTOVSKY, O. I., State Hydrological Institute, U.S.S.R.
KUPRIANOV, V. V., State Hydrological Institute, U.S.S.R.
KUZIN, P. S., State Hydrological Institute, U.S.S.R.
KUZMIN, P. P.,State Hydrological Institute, U.S.S.R.
MCQUEEN, I. S.,United States Department of Interior,United States Geological Survey, U.S.A.
MILLER, R.F.,United States Department of Interior,United States Geological Survey, U.S.A.
MORRISSEY, W . B., Ministry of Works, Water and Soil Division, N e w Zealand.
NASH,J. E., University College, Galway, Ireland.
O'DONNELL, T.,Imperial College of Science and Technology, London, United Kingdom.
19
List of contributors
OURYVAEV,V. A., State Hydrological Institute, U.S.S.R.

PACKER,P.,United States Department of Agriculture, United States Forest Service, U.S.A.
POPOV,O.V.,State Hydrological Institute, U.S.S.R.
PUSHKAROV,V. F., State Hydrological Institute,U.S.S.R.
ROCHE,M., Office de la Recherche Scientifique et Technique Outre Mer, France.
RODDA,J., Natural Environment Research Council, Hydrological Research Unit, United
Kingdom.
RODIER,J., Office de la Recherche Scientifique et Technique Outre Mer, France.
SHOWN,L.M., United States Department of Interior,United States Geological Survey, U.S.A.
SPENGLER,O.A.,State Hydrological Institute, U.S.S.R.
SUBBOTIN,A. I., U.S.S.R. Hydrometeorological Centre, U.S.S.R.
TAKENOUCHI, T.,Water Resources Development Public Corporation, Japan.
TISON,
L. J., International Association of Scientific Hydrology.
TOEBES,
C.,Ministry of Works, Water and Soil Division, N e w Zealand.
UBELL,K.,Institute for Water Resources, Hungary.
VISSER,
W. C.,Institute for Land and Water Management, The Netherlands.
VOLFTSUN, I. B., State Hydrological Institute, U.S.S.R.
WESSELING, J., Institute for Land and Water Management, The Netherlands.
In addition, material was provided by the Secretariat of the Canadian National Committee
in the form of proceedings of the National Workshop Seminar on Research Basin Studies held
in Ottawa, 1966.
20
Introduction
1.1 Scope and purpose

This guide deals with the establishmentand operation of representativeand experimental
basins and aims at suggesting desirable methods and possible avenues of research.
Hydrological research on representative and experimental basins has suffered not only
from a lack of basic data and a lack of standardization of observation and processing
techniques, but also from research methods giving only a limited understanding of
the physical processes occurring in basins. Research in representative and experimental
basins has frequently been treated as a statistical experiment but, apart from the fact
that classical statistical methods are not always applicable to hydrological data, the
results are of limited value since no methodology has been evolved to translate the
research findings to other basins. Moreover, the aim of hydrological research is not
only the collection of data, but rather an interpretation of these data for use in the
solution of management problems.
All basin studies,regardless of the classificationevolved, should be planned so as to
further understanding of the mathematical and physical relationships between the various
components of the hydrological cycle. Principal objectives of hydrological research
on representativeand experimental basins are the prediction and quantitativeestimation
of these components.For these purposes it is necessary to organize special experiments
and to carry out complex hydrologicalresearch on individual aspects of the hydrological
regimen in experimental basins (which are under the influence of man) as well as studies
in representative basins (which are in their natural state). It is evident that the scope
of experimental investigations should be based on the practical needs, taking into
account physiographical conditions, of a given country.
It is very difficult to conduct such investigations.In every case an individual approach
is required,and these investigationsshould therefore be considered only when objectives
are clear and funds and trained personnel are available.
The primary aim of this guide is, therefore, to foster more detailed research into
physical processes occurring in natural basins and at the same time to attack the elements
of uncertainty in hydrology. Since this type of research is both time-consuming and
expensive,the basin studies should be planned in detail and should allow for accurate
standardized data observation and data processing.Immediate analysis of the data is
necessary to establish needsforadditionalobservationsor forindicatingthoseobservations
which may be discontinued. Concurrent research in mathematical and statistical
techniques may be utilized to test hypotheses and to develop models for part of or for
the whole of the hydrological cycle.
21
Representative and experimental busins
The present work is aimed at serving those w h o are responsible for the organization
and conduct of hydrological research on representative and experimental basins. It is
based on the collective experience of various countries and contains general recommen-
dations on the selection of basins and the organization of observations on them, o n
the programme and types of observation, on instruments and observation methods,
on data-processing analysis and interpretation of observational results.
Taking into consideration the variability of natural conditions under which investi-
gations will be carried out on representative and experimental basins, only those
recommendations which are most suitable for definite conditions should be used.
The text which follows contains some detailed descriptions of observational methods
and data processing.These examples should not, however, pre-empt the use of methods
and existing procedures valid in many countries,
Numbers in brackets inserted in the text refer to references listed at the end of each
chapter. There is also a general bibliography at the end of the volume.
1.2 Definitionof representative and experimental basins

1.2.1 Representative basins
Representative basins are basins which are selected as representative of a hydrological

region, i.e., a region within which hydrological similarity is presumed. They are used
for intensive investigations of specific problems of the hydrological cycle (or part thereof)
under relatively stable, natural conditions. Thus a sparse network of representative
basins may reflect general hydrological features of a given region and their variations
over large natural zones.
Observations on representative basins should be long term and, if possible, combined
with the study of climatic, pedological, geological and hydrogeological characteristics.
They should be orientated to fundamental hydrologicalresearch and/orthe determination
of the effect of natural changes of the hydrological regimen and/orhydrological prediction
(see section 1.3.2) and/or the formation of a basic network of stations to which short-
term records from so-called temporary stations can be correlated.
For these purposes, representative basins should have minimal natural or cultural
changes during the period of study and, if changes occur, they should be carefully
recorded. The size of a representative basin depends on natural conditions and on
the objectives of the study. It lies ordinarily between 1 and 250 k m 2 and rarely exceeds
1,000km2. If the area of representative basins is larger, it may cause certain difficulties
in the homogeneity of certain basin characteristics and in the organization of
instrumentation and observations.
Benchmark basins are representative basins which are still in their natural states and
which have soil and vegetation conditions which are not expected to change for a long
time. In some countries benchmark basins could exceed 1,000km2 provided the principal
requirements of representativeness, homogeneity and stability are met. The principal
hydrological characteristics should be measured in these basins to permit interpretation
of the long-term trends. In some countries vigil basins are a special type of
representative basin.
In principle, after a long-term observation period on the basin, the whole represent-
ative basin or a part thereof may be transferred to the experimental basin category,
if on such blisiris the analysis shows that experimental investigations m a y be usefully
carried out,
22
Iniroduction
I 1.2.2 Experimental basins

Experimental basins are basins which are relatively homogeneous in soil and vegetation
and which have uniform physical characteristics.On such basins the natural conditions,
i.e., one or more of the basin characteristics, are deliberately modified and the effects
of these modifications on the hydrological characteristics are studied. This general
objective makes it imperative that the research organization has the right to manipulate
the land at will. Because more detailed studies are required on experimental basins
than on representative basins, and also because of the necessity of owning or leasing
experimental basins, these basins are normally restricted in size to a maximum of
about 4 km2.
Basins where detailed studies of the hydrological cycle are undertaken, but where
no actual cultural change is carried out, have also been called experimental basins
~ in some countries.
Research on experimental basins is normally a study of comparisons and therefore
they are sometimes operated in groups of two or more basins. The important aspect
of any study on an experimental basin is that any cultural change is preceded by a
pre-calibration period.
1.3 Purposes of representative and experimental basins

1.3.1 Representative basins
Representative basins are set up to suit the purposes specified below. As mentioned
in section 1.2.1,it is not always possible to utilize any representative basin for all purposes
and it is desirable that the primary purpose is specified for each basin [6,14, 17, 21,
23, 28, 32, 341.
(a) Fundamental research. Representativebasins are frequently used for fundamental
research.This can take the form of a study of all the physical processes of the hydrological
cycle or of any part of it or of any specific hydrological characteristics. It includes
research on observational techniques.
The primary motivation of this research is the study of the physical phenomena
and, as such, is ideally suited for the training of staff. It can be carried out on individual
basins or on networks of basins.
(b) Eflects of natural changes. Some representative basins, especially benchmark
and vigil basins, may be used to study the effects on the hydrological regimen of a
natural change.Natural changes may, for instance,be changes in climate,in vegetational
characteristics because of natural growth, in pedological characteristics such as erosion,
podzolization, etc.
(c) Hydrological prediction. Most representative basins will be used for the
development and improvement of methods of hydrological calculation and prediction
and for the assessment of water resources in a region or area. This will involve a detailed
analysis of hydrological phenomena and the direction of research towards solving
methodological problems in hydrology.
(d) Extension of records. Ideally representative basins can serve as a basic network
of hydrological stations in areas where river basins are generally small (say < 3,000km2).
In some countries the representative basin network does serve as the basic network
of hydrological stations.
One main purpose of this basic network is to provide a long-term record of basic
data to which short-term records observed on roving or investigation stations m a y
be correlated.
23
Representative and experimental barins
1.3.2 Experimental basins

(a) Effects of cultural change. The most pres-ing and practical objectives in many
countries for the setting up of experimental basins are the study of the effects of cultural
change on the hydrological regimen. Cultural changes involve the artificial change of
one or more basin characteristics, with a resulting change in some hydrological
characteristics and include any changes in land use [12](such as afforestation) and/or
land management (such as contourbanking) and the influence of the use of water
resources (e.g., pumpage, artificial recharge).
This type of research can be carried out on individual basins or on networks of basins.
It is sometimes thought that research on the effect of a cultural change on the hydrological
regimen is speeded up if duplicate basins are set up, on one of which the basin charac-
teristics are kept as constant as possible during the investigation period. This basin is
called the control basin. In some cases multiple replication is advocated [5]. This is
not always desirable; the control basin is frequently a changing base, as it is difficult
to keep the basin characteristics constant for any period of time and the utilization
of one or more control basin(s) fosters the tendency to calibrate the experimental basin
after a few years of observation.This does not necessarily give a representative record
of the typical local climate.The climate during the calibration period,may, for instance,
have been abnormally wet and it is not likely that research findings for this period
are directly applicable to drier periods. Moreover, the use of replication may give
statistically valid results, but it does not necessarily contribute to the problem of
translating research findings to other basins (see section 2.4).
(b) Hydrological prediction and extension of records. Experimental basins can be
used also for this purpose during their calibration period, provided no direct conflict
arises with the principal aim of the manipulation action. Control basins which are left
in their natural condition can also be used as representative basins [29,301.
(c) Fundamental research. Experimental basins are,like representativebasins,ideally
suited for fundamental hydrological research.Moreover, the extensive instrumentation
available in experimental basins provides excellent opportunities for staff training.
1.4 Survey of research needs

Requirements for establishing representative and experimental basins in any country
should take into consideration the following: (a) availability of hydrological data and
hydrological research results; (b) estimation of present water resources and their use
(with special reference to the plans of future economic development of the country);
(c) hydrological problems which have to be solved in the country and availability of
data specifically necessary for solving these problems.
The need for hydrological research of any kind and for a study of hydrological
processes to establish hydrological laws may be estimated from a consideration of the
above conditions.
Having established research needs it is necessary to decide whether representative
basins or experimental basins or both should be installed [is].
Nations with a currently sparse network of hydrological stations should establish
representativebasins beforeexperimentalbasins are attempted,for the followingreasons.
First,some or all of the representative basins may fit into the basic network and this
would simplify staffing and financial needs. Secondly, experimental basins, because
of ownership or lease requirements and because of the need for much more extensive
instrumentation and observation,are more costly to operate than representativebasins.
Thirdly,research on representative basins is more defined at present since a completely
successful methodology for experimental basins is not yet available.
24
Introduction
The objectives and research methods should be clearly stated for each basin or group
of basins. Fundamental scientific research for any problem demands that a hypothesis
be formulated and that methods be stated to test this hypothesis.
If changes in land use and/or land management are planned, approximate times
of change should be stated at the onset as a working plan. The approximate period
of investigations (normally indefinite for representative basins and for experimental
basins over a period typical of the local climate) should be stated.
To increase the usefulness of the data which is collected in experimental basins,
consideration should be given during the planning stage to the inclusion of companion
studies. Such studies, which could be carried out on sites within basins, plots or
sub-basins,would not only permit assessment of short-termmanipulative actions which
could guide the planning of larger modifications, but would also permit detailed
investigations of basin hydrological processes.
1.4.1 Stafing for basins

Potential staffing requirements must be estimated. It is imperative to have a trained
hydrologist in charge of each basin or group of basins in order to make one research
worker familiarwith questionsand items which are not easily recorded.The qualifications
of the person in charge and the supporting staff should depend on the scope of the
research and on the degree of automation of the observations.
Generally the staff should consist of three types of specialist:
1. Scientists or engineerstrained as hydrologists in hydrologicalinstitutes or universities.
2. Hydrological assistantswho have had specialhydrologicaltraining at technicalcolleges.
3. Hydrological observers who have had special hydrological observation courses.
For any basin or group of basins where advanced research is carried out, the person
in charge should be a professional hydrologist.
Hydrological assistants are generally used for complex observations and data analysis;
observers make regular hydrologicalobservations and carry out primary data processing.
It is extremely desirable for every observer on representative and experimental basins
to be trained to make all kinds of observation taken on the given basin so that he may
take the place of another observer if the necessity arises.
It is frequently necessary, especially in the first stage of the research, to employ
specialists such as surveyors,geologists,soil physicists,botanists,etc.(see section 1.4.6).
With automation of observations on basins or the application of more complex
measuring devices,it may be possible to operatewith a smaller staff.The ratio of scientists
to sub-professionalstaff varies from country to country,but a ratio of one professional
to three sub-professionalsmay be taken as satisfactory.
Table 1.1 gives information on the staffing used in the U.S.S.R.for observations
on representative and experimental basins of different sizes. Table 1.2 is given as
an example from the U.S.S.R. of the number of observation points on basins of various
sizes in flat country. In other countries numbers of staff and observation densities
may be different.
1.4.2 Basic recommendations on standardization of methods

of Observation, instrumentation and data processing
Standardization of methods of observation, instrumentation and data processing is
desirable on a national,regional and world-widebasis so that results may be interpreted
and compared outside the research organization studying the problem.
Standardization,if possible,is essential in the following:(a) selection of hydrological
25
Representative and experimental basins
TABLE 1.1. Approximate number of staff employed on representative basins in the U.S.S.R.
according to the size of the basin and scope of observations
Scope of Small basins Medium basins Large basins
observations (up to 50 km2) (200-500 kma) (SOO-1,000 kmz)
Minimum 1 hydrological 1 hydrological 2 hydrological

assistant assistant assistants
2-3 observers 2-3 observers 4-6 observers
Average 1 hydrological 1 hydrological 1 hydrologist
assistant assistant (for 2 basins)
3 4 observers 4-6 observers 6-8 observers
Maximum 1 hydrological 1 hydrologist 1 hydrologist
assistant 1 hydrological 2 hydrological
4-6 observers assistant assistants
6-8 observers 10-12 observers
TABLE 1.2. Approximate number of points for different kinds of observation on representative
basins of various sizes in the U.S.S.R.zone of temperate climate and plain relief
Size of basin
scope of Programme of
observations observations Small Medium Large
(up to 50 kmz) (200-500 kma) (SOO-1,000 km2)
;d Precipitation
,# (monthly) 5-8 8-10 10-15
Snow surveys up to 10 10-12 12-15
Streamflow 1-2 2-3 3-5
Evaporation from soil
and water surface1 1 1 1-2
Ground-water levels 5-10 10-15 15-20
Soil moisture 1 1-2 2-3
Erosion and suspended
sediment discharge 1 -2 2-3 3-5
Chemical composition
of water 1 1-2 2-3
Energy balance 1 1 1
Temperature of river
and ground water 2 2-3 3-5
Snow melt 1 1 1
Soil frost 1 1 1-2
1. Evaporation from the water surface is measured only on a specified number of basins.
regions, representative and experimental basins and run-off plots ; (b) selection of
instruments and methods of their installation;(c) methods of observation;(d) accuracy
of measuring and units of measurements;(e) methods of data processing;(f) publication
forms for observed data; (g) terminology.
Standardization of instrumentation and methods of observation,as specified in this
guide,may be carried out from country to country when compatiblewith physiographicai
conditions and the financial situation.These standardization methods can be used on
an internationalbasis to aid the fruitful interchange of data and results. O n a national
basis, the use of as few different types of instrument as possible will simplify servicing
and maintenance problems. Moreover, standards of installation, such as that for the
26
Introduction
height of precipitation gauges above the ground,are essential to obtain full value from
the national research.
It is inevitable and desirable that in future most observations and data-processing
methods be automated and it is important to simplify and standardize methods now
to ease the change-overto automated methods. It is also important to study the problem
of accuracy of data and indicate in publications, if possible,the standard of accuracy.
1.4.3 Data-reporting methods

All data collected on representative and experimental basins should be processed and
tabulated to make them readily usable for scientific analysis.
Apart from the observation of basic data it is necessary to collect information that
is not easily recorded by instrumentation (e.g., grazing density, occurrence of fire,
failure of equipment, etc.). Such information is best recorded in a diary which is kept
daily for experimental basins and monthly for representative basins.
Some observed data (monthly, annual and extreme values) are to be published in
hydrological yearbooksissued by national committees for the IHD,according to formats
adopted by the Unesco Co-ordinatingCouncil.Data not published in these yearbooks,
in particular daily values and values for shorter time intervals, should be tabulated
(see Chapter 5) and kept in the files of the hydrological service of the country. These
should be easily accessible to scientists who wish to use these data for hydrological
research.
The next stage of data reporting is regular publication of preliminary results of
research on representative and experimental basins during the investigation period and
of publication of the final data after the whole programme has been completed.
Papers on preliminary investigations may be published every one or two years in
the form of research reports to give general research information on representative
and experimental basins.
Such synopses will help hydrologists of various countries,making similarinvestigations
in similar physiographical conditions,to avoid duplication,to save time and expenditure
and to exchange experiences concerning the organization of the research.
The final stage of data reporting is the preparation of scientific papers, monographs
and atlases of observation and research data on representative and experimental
basins. In cases where investigations of the same problem are carried out under a single
programme, this work requires international co-operation of scientists from various
countries.
1.4.4 Research co-operation and research observation

programmes
Hydrological research on representative and experimental basins is, to a large extent,
co-operative research between hydrologists, engineers, soil physicists, statisticians,
botanists, geologists,etc., and has to be carried out by a number of research groups.
Close co-operationbetween workers and research groups is necessary and it is desirable
also to maintain contact with similar projects in the same country or elsewhere.
To ensure this co-operation,a research observation programme is of great assistance.
Such a programme is prepared on an annual basis stating the work to be done on a
I particular basin, the person or group to do it and the methods to be used.
A research observation programme can, for instance, have specific headings such as
management, surveys,instrumentation and observation,data processing,data analysis,
and data publication.
27
Representative and exper-imentalbasins
Where variousprojects are involved,these should be correlatedto ensure that maximum

attention is given to all problems and that overlapping or duplication of work does not
become excessive.
This is particularly important in technique studies where particular organizations
may be charged with carrying out particular studies on one basin,which are applicable
to other basins.
Country
JAPAN
Organisation in
Ministry of Agriculture and Forestry
charge of
Forest Experiment Station
activity
4-770 Shimo-megum , Meguuro-hi , Tokyo
tame of region Yamagata
ieology : hasement complex ( t d f , schistose tuff)

Physiogrophlc ropoqraphy :middle of the Tôhoku RIOUntainOUS area
soit . -
description oí
Vegetation : forest land
region
ClIrnate : oceanic climate with abundant snowfall
(2390 rmn average areclpitation)
Area : o*082 K m 2 (composed of 4 sub-areas)
Location of Longitude :140' 16' E
Basin Latitude :38' 56' N
-
Altitude : extrene : 160-250 rn
Median Altitude :
Physloqraphic
description of
Bacin
~
1 raingauge and 1 V-not& weir equipped vi:

'lesent stage recorder for each sub-area
Equipment
Future -
To study the effect of change8 in forest
conditions, especially cutting, on flow.
ObjectIves
FIG.1.1. Basin card for Shione Experimental Basin, Japan [13].
28
Intuodiiction
Co-operationis desirable not only on the national level, but also on a world-wide
regional level (i.e.,with countries that have similar problems or similar physiographical
areas) and on a general international level.Basin cards should be completed for each
basin showing research under way or planned in the various countries and these cards
transmitted to Unesco for international distribution. A n example of such a card is
shown in Figure 1.1.
A useful summary of present and proposed representative and experimental basin
studies by all Member States of Unesco is given in reference [13].
1939 - basin established

Hlstory and
present wor II
Dota col1ecte.d:daily precipitation anü f'iovr
Analyses made : -
Research into methods of' maiiagei-neiit of head-

water forest, arrangement of water-spreading
works for flood control, water resources and
conservation in mcuntain districts.
Future
developments
Report on observations made at experirnontal

-
basins for forest hgai.oloo.g daily precipitat:
(0y publicatlons and daily flow. Government Forest Experiment
Station. March, 1963. (in Japanese)
29
1.5 Terminology
The establishment of a uniform scientific terminology for use on research basins is
of great significance to aid international co-operation in such research. Some terms
which have been used in the text are given below [l-4,7-11, 15, 16, 19, 20, 22, 24-7,
31-3, 35, 37-42].
GENERAL TERMS
Hydrological model. Mathematical formulation to simulate natural hydrological phenomena

which are considered as processes or as systems.
Hydrological process. Hydrological phenomenon which undergoes continuous or periodical
changes particularly with respect to time (time series).
Deterministic process model. The chance of occurrenceof the variables involved in such a process
is ignored and the model is considered to follow a definite law of certainty but not any law
of probability (e.g.,flood routing is a deterministicprocess,a unit hydrograph is a deterministic
model).
Probabilistic process model. The chance of occurrenceof the variables involved in such a process
is taken into consideration and the concept of probability is introduced in formulating the
model. The sequence of' occurrence of the variates is ignored (e.g.,flow duration).
Stochastic process model. The chance of occurrence of the variables involved in such a process
is taken into consideration and the concept of probability is introduced in formulating the
model. The sequence of occurrence of the variation is taken into account.
Hydrological system. Aggregation or assemblage of objects united by some form of regular
interaction or interdependence. The system is dynamic if there is a process taking place in it.
If the process is considered probabilistic or stochastic the system is said to be stochastic.
Otherwise it is a deterministic system.Furthermore,the system is called sequential if it consists
of input,output and throughput (e.g.,the hydrologicalcycle is a sequential,dynamic system).
For simplification most systems are treated as deterministic.
Stationary process. The probability distribution of the variable x remains constant throughout
the process.
Non-stationary process. The probability distribution of the variable x varies with time.
Normal hydrological values. Mean values of hydrological characteristicstaken over a period of
such length that the mean over any longer period does not significantly affect the vaIues
obtained.
Water balance. A balance of input and output of water within a given defined hydrological
area such as a basin, lake, etc.,taking into account net changes of storage.
Ecosystem. The interacting complex of soils,plants and animals which,in response to climatic
and other environmental conditions, forms a varied covering over much of the earth.
Region of deficient moisture. A region where the annual (potential) evapotranspiration exceeds,
on the average, annual precipitation.
Region of excessive moisture. A region where annual precipitation exceeds, on the average,
annual (potential) evapotranspiration.
Zonality of hydrological phenomena. Regimen changes of hydrological characteristics with alti-
tudinal and latitudinal location.Altitudinal changes occur in mountain regions (alpine belt,
forest belt, etc.); latitudinal changes are more marked in plain regions (tundra belt,forest
belt, steppe belt, etc.).
Azonalify of hydrological phenoniena. Regimen changes of hydrological characteristics caused
by local changes (aspect,prevailing wind direction,etc.). These conditions are superimposed
upon zonal Characteristics.
Basin and climatic characteristics. Variables (climatic,vegetational,pedological,geological and
geomorphological)that determine the hydrologicalregimen in a basin.Some of these charac-
teristics can be modified either naturally or culturally and result in changes in hydrological
characteristics.Sometimesbasin characteristicsare classified as zonal(mainly climatic), azonal
(basin area,stream profile,etc.), and interzonalfactors (factors associated with, but varying
within a given zone, such as percentage of forest, swamp,etc.).
Run-offplot.A part of a slopebordered with artificialwater divides and instrumented to measure
overland flow and total sediment load (and in some cases interflow also).
30
Introduction
CLIMATIC TERMS
Macro-climate. The general climate prevailing over a large area considered as a unit.
Meso-climate.The climate of a valley or system of valleys which differs from the macro-climate.
Micro-climate.The climate of a very small area with special reference to local variations from
the general climate of a region.
Precipitation.The discharge of water in liquid or solid statefrom the atmosphere onto the surface
of the earth. Precipitation includes rainfall,snow, hail, sleet and dew.
Rainfall. The quantity of precipitation that falls as rain only.
M a s s rainfall. The volume of rainfall.
Rainfall intensity. The rate of rainfall (depth or volume of rainfall per unit time).
Rainfall excess. The rainfall in excess of interception loss,depression storage,infiltration and
evaporation. It is the rainfall available for surface flow or direct flow.
Gross rainfall. The total amount of rainfall measured in the open or above the forest canopy.
Net rainfall. The water that enters the mineral soil after penetrating the forest canopy and litter.
Interception.The process by which rainfall is caught on forest vegetation and redistributed as
throughfall, stem flow, and evaporation.
ThroughfaZí. The rainfall which directly reaches the litter through spaces in the forest canopy
or as drip from leaves, twigs and stems.
Stem flow.The rainfall which,having been caught on the canopy,reaches the litter or mineral
soil by running down the stems.
Net interception loss. The portion of precipitation which is returned by the aerial parts of the
vegetation and which has no effect on the soil-waterconsumption of the plant.
Litter interception loss. The rainfall retained on the litter layer and evaporated without adding
to moisture in the underlying soil.
Gross interception loss. The rainfall evaporated from canopy and litter.
Snow. A form of precipitation composed of ice crystals-most are branched but some are star-
shaped.
Snow course. A line or series of connecting lines along which the depth of snow is measured
and snow samples are taken at regularly spaced points.
Snow depth. The snow accumulation on the ground (depth of snow-packcoverage).
Snow survey. Measurements of depth and density of snow pack along a fixed course to determine
water equivalent in snow pack over large areas.
Snow density. The ratio between the volume of melt water derived from a sample of snow and
the initial volume of the sample.This is numerically equal to the specific gravity (relative
density) of the snow.
Water equivalent (os snow). The amount of water that would be obtained if the snow were
completely melted.
D e w . Water drops deposited by condensationof water vapour from the air,mainly on horizontal
surfaces cooled by nocturnal radiation.
Vapour pressure. The pressure exerted by a vapour when it is in a confined space.When several
gases or vapours are mixed together in the same space each one exerts the same pressure
as it would if the others were not present.
Mixing ratio. (Humidity mixing ratio). The relative proportionsby weight of water vapour and
dry air in a given specimen of damp air.
Absolute humidity. The mass of water contained in a given volume of moist air.
SpeciJc humidity.The amount of water vapour (in grammes) containedin 1 grammeof moist air.
Relafive humidity. The ratio (expressed as a percentage) of the water vapour actually present
in unit volume, to that which the air would contain if saturated at the air temperature.
D e w point. The temperature to which air can be cooled without causing condensation.It is the
temperature at which the saturation vapour pressure is identical with the pressure of the
vapour in the air.
Radiation. Energy emitted in the form of electromagnetic waves.
Intensity of radiation. The rate at which energy crosses a given area.
Insolation. A term applied to the solar radiation received by terrestrial or planetary objects.
Long-wave radiation. Radiation received as heat.
Short-wave radiation. Radiation received with a smaller wavelength than that of heat, i.e.,
infra-red,ultra-violet,X-ray,gamma-ray,etc.
31
Albedo. Theproportion of radiation fallingupon a non-luminousbody which is diffusely reflected

by this body.
Evaporation. The process of conversion of water or ice into aqueous vapour. The amount of
water or ice so converted.
Transpiration.(a) The quantity of water absorbed and transpired and used directly in the building
of plant tissue in a specified time.It does not include soil evaporation.
(b) The process by which water vapour escapes from the living plant,principally the leaves,
and enters the atmosphere.
Evapotranspiration. The sum of water lost from a given land area during a specified time-
by evaporation from water surfaces,soil and snow,and by transpiration from vegetation
and in building of plant tissues. Evapotranspiration normally includes interception loss.
Potential evapotranspiration.The evapotranspiration that could occur if there were an adequate
moisture supply at all times.
SURFACE WATER TERMS
Perennial flow. Flow in streams containing water at all times except during extreme droughts.
Intermittentflow.Flow in streamsthat carry water most of the time but ceaseto flow occasionally
because evaporation and seepage into their bed and banks exceed the available stream flow.
Ephemeral flow. Flow in streams carrying water only after rains or periods of snowmelt.
Depression storage. The volume or depth of precipitation stored in natural depressions in the
land surface. Water in depression storage ultimately infiltrates,evaporates,or is drained
artificially.
Surface retention (Initialabstraction). That part of precipitation that does not infiltrate or appear
as a surface flow during,or immediately after the period of rainfall or snowmelt.Surface
retention includes interception storage and depression storage.
Detention storage. The volume or depth of precipitation on the ground surface or in channels
during or shortly after rainfall or snowmelt which is available for surface flow and/or
infiltration during or shortly after rainfall ends.Detention storage includes surface detention
and channel detention.Detention storage does not include depression storage.
Channel. An open,natural or artificial watercourse which periodically or continuously contains
moving water.
Surface detention. The volume or depth of detention storage on the ground surface,excluding
channels.Jt is available for overland flow and infiltration.
Channel detention. The volume or depth of detention storage in channels.Channel detention
is available for surface flow.
Infiltration.The movement of a fluid into a substance through pores or small openings; in
hydrology it is the movement of water into the soil.
Mass infiltration. The volume of infiltration.
Infiltration rate. The rate of infiltration (depth or volume of water infiltrating per unit time).
Infiltration capacity. The maximum rate at which water can flow into a soil in a given condition.
Initial infiltration rate. The initial rate of infiltration at the beginning of storm rainfall.
Ultimate infiltration rate. The final, approximate,constant rate of infiltration.
Flow. The movement of water or other mobile substance.As a term it may mean run-offor
discharge.
Run-08. Normally the flow of water derived from precipitation considered as a volume or as
depth of water.
Discharge. The flow of water considered as a rate.
Surface flow (Surface run-08). The flow of water over the land surface.Surface flow includes
overland flow and channel flow and is normally water derived from rainfall excess.
Overland flow. The flow of water over a land surface in transit towards a permanent stream
channel. Overland flow is derived from surface detention.
Channel precipitation flow (Channel precipitation).That part of the flow of water in permanent
stream channels that is derived from direct precipitation in channels.
Subsurface flow.Any flow below the surface of the ground which may contribute to interflow,
base flow or deep percolation.
Interflow (Prompt subsurface flow). The flow of water from ephemeral zones of saturation.It
moves through the upper strata of a formation at a rate much in excess of normal base-flow
seepage.
32
Introduction
Basefiow. The flow of ground water from beneath a permanent ground-water table.
Direct flow. The flow of water entering stream channels promptly. It includes surface flow and
interflow and is used where interflow cannot be separated in hydrological analyses.
Snowmelt spring fiood ('polovodie'). Considerable flood rise in rivers occurring every spring and
caused by melting of snow pack, accumulated during winter period. It is a typical feature
of rivers of plains in countries with a cold winter and heavy snowfalls.
Specific discharge.Water yield per unit drainage area,per unit time (l/sec/km2or ma/sec/km2).
SUBSURFACE WATER TERMS (see Figs. 1.2, 1.3 and 1.4)
Deep percolation. The loss of water from a basin by subsurface flow to other basins.
Basalflow. The volume of water appearing as base flow plus the volume of water lost by deep
percolation.
Zone of saturation. A zone in the lithosphere which is saturated with water.
Ephemeral zone of saturation. A zone in the lithosphere which is temporarily saturated with
water.
Ground-water table. The upper surface of a zone of saturation.
Water table (Phreatic surface). The upper surface of a continuous zone of saturation except
where that surfaceis formed by an impermeablebody.N o water table exists where the upper
surface of a continuous zone of saturation is formed by an impermeable body. Base flow
is supplied from water underneath the water table.
Perched ground-water table. The upper surface of a zone of saturation separated from other
bodies of ground water by impermeable strata.
Ephemeral ground-water table. The boundary of an ephemeral zone of saturation where this
cannot be termed a perched ground-watertable.
Zone of aeration (Unsaturated zone). A zone above the water table where unsaturatedconditions
occur.
Intermediate belt. That part of the zone of aeration that lies between the belt of soil moisture
and the capillary fringe overlying the water table.
Ground water. The water in the ground that is in a zone of saturation.
Intermediate wuter. The water in the intermediate belt.
Soil moisture (Soil water). The water in the soil if not all pores are filled with water (unsaturated
conditions). If all pores are filled with water (saturated conditions) soil moisture behaves
as ground water. (N.B. The base of the soil is regarded as reaching at least to the lower limit
of plant roots.)
Hygroscopic water (Hygroscopic moisture). The water adsorbed to the surface of soil particles
when in equilibrium with an atmosphere of 98 per cent relative humidity.It is unavailable
to plants.
Gravitational water. The water which moves into,through,or out of the soil under the influence
of gravity.
- WATER
GRAVITATIONA L CAPILLARY
I
I- WATER WATER l
a
+ I * - , =3
IQ
-
v>
-FIELD
CAPACITY
+
WILTING
POINT
Represetitative und experimeniul busitis
- LAND SURFACE
SOIL
MOISTURE EPHEMERAL GROUND-
. . _'WATER TABLE.
',
SUSPENDED INTER- ZONE

WATER MEDIATE OF
WATER AERATION
I LENS I
.-. .,,.+-:
i..:....:
i
:
:.::..:.....:,,-...:
.
.I
GROUND .'.:.I..::'.. .. ZONE OF

WATER iATüRATION
FIG.1.3. Subsurface water zones.
-10'
-10
O IO 20 30 40 50
Moisture content per ceni by weight
FIG.1.4. Relationship between moisture potential and moisture content.
34
Introduction
Capillary water (Capillary moisture). The water held by surface tension in the capillary spaces
and as a continuous film around the particles, free to move under the influence of capillary
forces.
Field capacity. The amount of moisture held in the soil after excess gravitational water has
drained.
Wilting point. The soil moisture content at which permanent wilting of the plant occurs. This
may not apply to arid-land plants.
Available soil moisture. The water in the soil available to plants. It is normally taken as the
water in the soilbetween wilting point and field capacity.In this contextwater-holdingcapacity
of a soil is used and is identical to the available water.
Percolation. The movement, under hydrodynamic pressure, of water through the interstices
of a rock or soil, except the movement through large openings such as caves.
Capillary conductivity. (a) Qualitafive. The physical property relating to the readinesswith which
unsaturated soils or rocks transmit water; or:
(b) Quantitafive.The ratio of the water flow velocity to the driving force in unsaturated
soil or rock. The calculation is valid under conditions where flow velocity is proportional
to driving force,e.g.,in practical units when the driving force is expressed in terms of the
hydraulic gradient,capillary conductivity is the ratio of flow velocity to hydraulic gradient
and has the dimensions of velocity. As saturation is approached, capillary conductivity
approaches the hydraulic conductivity.
Hydraulic conductivity (Coeficient of permeability). (a) The ratio of the flow velocity to the
driving force for the viscous flow under saturated conditions of a specified liquid in a porous
medium. Physical dimensions will depend on the equation selected to express the flow.
(b) Practical unifs. The ratio V/grad 4 = K,where Vis a fictivemean velocity of a specified
liquid under saturated conditions,and grad 4 is the hydraulic gradient in the Darcy equation
V = K grad 4. In this case K = LT-l.
Potential level (of soil moisture). The level of soil moisture at which evapotranspiration depends
only on evaporative factors of the atmosphere and not on soil moisture conditions. Some
workers consider this level approximately equal to field capacity.
Water suction. ([Soil]water-moisture suction-tension). (a) A numerical measure of the energy
by which suspended water is held. When expressed as the common logarithm of the head
in centimetres of water necessary to produce the suction corresponding to the capillary
potential, it is called pF.
(b) The negative gauge pressure relative to the external gas pressure on the suspended
water, to which a solution of identical composition must be subjected in order to be in
equilibrium through a porous permeable wall with this water.This quantity may be identified
with the capillary potential defined below.
Capillary potential. The amount of work that must be done per unit quantity of pure water in
order to transportreversibly and isothermally,to the suspended water,an infinitesimalquantity
of water from a pool containing a solution identical in composition to the suspended water
at the elevation and the external gas pressure of the point under consideration.For a definition
of total and other potentials of suspended water, see reference [16].
EROSION A N D SEDIMENTATION T E R M S
Suspended sediment load. Sediment which remains in suspension in flowing water for a con-
siderable period of time without contact with the stream bed.
Sedinient concentration.The ratio of weight of dry solids to the weight of the sample of water-
sediment mixture.
Bed load. Sediment which moves in almost continuous contact with the stream bed,being rolled
or pushed along the bottom by the force of water.
Rill and sheet erosion. The removal of a fairly uniform layer of soil or material from the land
surfaceby the action ofraindropimpactand the product of the kineticenergy and the duration
of rainfall. When the overland flow concentrates in microchannels, sheet erosion becomes
rill erosion. Although rill and sheet erosion are difficult parameters to measure in the field,
it has been shown that, for the areas investigated, it is the dominant source of sediment,
accounting for over 90 per cent of the eroded material in many basins.
Mass movement (includes soil creep, landslip). The term applied to unit movement of a portion
35
Represen fative and experimental basins
of the land surface as in soil creep,landslide or slip.Mass movement may be classified in a

continuous series from slow,relatively dry flow,in the case of creep, to sliding as well as
flow,in the case of slips, and becomes more rapid with increasing water content in earth
flows and mud flows.
Wind erosion. Wind erosion is the detachment,transportation and deposition of soil by the
action of wind. The removal and redeposition may be in more or less uniform layers or as
localized blow-outsand dunes. The problems caused by wind erosion are generally localized
in areas having sandy soils or fine, dry loess or pumice.
GulZy erosion. Gullies are transitional channels between rills and larger stream channels and
have been associated with accelerated erosion and result from concentration of rill wash.
The flow is generally ephemeral.
Bunk erosion. Bank erosion is the removal of the channel banks by lateral shifting and under-
cutting by the stream. In extreme cases large areas of fertile flood plain can be destroyed
by bank erosion. There are several factors that influence the rate of bank erosion. These
are composition of bank material and soil moisture content.There are indications that there
is a correlation between precipitation and erosion during selected periods. The precipitation
not only increases discharge,but also increases moisture content of bank material making
it more susceptible to removal.
Channel scour. Most alluvial stream channels scour their beds during periods of high flow.
This lowering of the channel floor is not always apparent because deposition of fine sediment
during flood recessions often refills the scoured reaches. Scour is, however, an important
erosional process because of the contribution it makes to the sediment load during floods.
Observations on natural rivers with sandy beds are often misinterpreted.The position of
thestream bed may not be well defined;there may be too few points to determinea satisfactory
cross-section,or too few cross-sections.Moreover,when the bed is low at one point it may
be high at another point during a flood.Therefore observations often represent maximums
of scour or fill instead of averages.
ISOTOPIC TERMS
Fructionation factor. Ratio of the vapour pressures of the lighter isotopic species of water to
the heavier species.
SMO W . Standard Mean Ocean Water; an arbitrary standard from which relative deviations
of deuterium and oxygen-18 concentrations are determined.
Tritium Unit (T.U.). Unit used to express tritium concentrations.1 Tritium Unit is a concen-
tration of 1 tritium atom in 1018 hydrogen atoms.
Deviafion per thousandfrom SMO W . Unit of deuterium and oxygen-18concentration of water
relative to SMOW.
Isotopic species of water. Water molecules formed by the combination of different isotopes of
hydrogen and oxygen, i.e., H@O, HDO,HPO.
Environmental isotope technique.A method of studying the concentrationsof deuterium,oxygen-
18, tritium and carbon-I4in natural water.
Zsofopic composition of water. The concentrations of isotypes of hydrogen and oxygen in water.
Halfilife.The period of time in which a radioactive species loses half of its original activity.
Radioactive dating. A dating method based on the property of radioactivedecay of radioisotopes.
Pulse dating. A dating method based on the monitoring of a pulse introduced in a system.
1.6 Measurement units and symbols

A list of measurement units and symbols employed in research on representative and
experimental basins is given in Table 1.3.
36
Introduction
t-
3 W
w mmmnmNmmmm, b b
rnrn
m m m m m ~ m m m m r nm m
%8
c?t
rnmmrnrn~~rnrnrnrnm
0 0 0 0 0 “ 0 0 0 0 ~ õö
O 0 I I I 0 0 0 0 0 N 0 0 0 0 0 loo
N
37
.-o
I
-B
B
O
.--
.
I I I I I I I I I I
a
B
.-
aO
Fl-F
mwrnmrn
w - m m m
c mclrnrnrn
c \ 1 \ 0 0 0 0
.B
I I I I I I I I I I I I I I lhioooo
E
I I I
.
2.2.
I 'E.'E.i.i.c'
I I I I I I I I I I
I I
q"
I I sz" s I ks I i&
o
--aa
fi
.-
-a
3 T
,CI:
.-8
I
m
E
38
00
t ?g
I 21 I I40
rn
t.
QI
I % i I
I .i I I
3 3 3 3 3
L 4
444444
L L4 I4 I . .
. .
. .
I*
39
Ft-r-r-r-t-r-r-t-r-
3 m m m m m m m m m m
- m m m m m m m m m m
~ m m m m m m m m m m
9999999999
I~oooooooooo I I I
I I
40
Introduction
References
1. ACADEMYOF SCIENCES OF THE U.S.S.R. (ed.) 1952. Terwiinologia mekhaniki zhidkosti
(gidromekhaniki) [Terminology of fluid mechanics]. Moscow.
2. ALTOVSKY, M. E. (ed.) 1962. Spravochnik gidrogeologa [Manual for hydrogeologists].
Moscow, Gisgeoltehizdat.
3. BARKOV,A. S. 1958. Slovar-spravochnik po jìzichesko i geograjìi [Glossary of physical
geography]. 4th ed. Moscow, Uchpedgiz.
BAUR,A. J. 1952. Soil and water conservation glossary. J. Soil Water Cons., 49 :41-52.
~ 5: BEFANY, A. N.et al. 1959. Experimentalnye issledovania poverkhnostnogo stoka i metody
obobschenia opytnogo materiala [Experimentalinvestigationsof overland flow and methods
I
of interpreting observational data].Proc. Third All-Union Hydrological Congress,p. 44451.
I 6. BLIDARU, S. 1965. Emploi des bassins représentatifs et des stations expérimentales à l'étude
des phénomènes hydrologiques. IASH publ. no. 66, 1 : 107-15.
7. BRITISHHYDROGRAPHIC DEPARTMENT, ADMIRALTY. 1938. Admiralty weather manual.
London, HMSO.
8. BRITISHMETEOROLOGICAL OFFICE. 1939. The meteorological glossary. London, H M S O .
9. CAMPBELL, D.A. 1951.Types of soil erosion prevalent in New Zealand.IASH publ. no. 35,
p. 82-95, Brussels.
10. CHEBOTAREV, A. I. n.d. Gidrologichesky slovar [Glossary of hydrology].Leningrad, Gidro-
meteoizdat.
11. COLBY, B. R. 1964. Scour and 511 in sand-bed streams. (USGS Prof. paper 462-D.)
12. CORMARY, Y. 1965. Le bassin expérimental en tant qu' outil de recherche en matière de
conservation des eaux et du sol. IASH publ. no. 66, 1 :201-16. (Symposium of Budapest.)
13. COSTA, J. DA; JACQUET, J. 1965.Presentation of the results of the Unesco/IHD questionnaire
on the representative and experimental basins in the world. IASH Bull. Paris, no. 4.
14. DUB,O. 1965. Experimental and representative basins in Czechoslavakia. IASH publ.
no. 66, 1 : 131-5.
15. GLYMPH, L. M. 1957. Importance of sheet erosion as a source of sediment. Trans. Amer.
Geophys. Un.,38(6) :903-7.
16. INTERNATIONALSOCIETY OF SOIL SCIENCES. 1963.Soil physics terminology.ISSS Bull. no. 23.
(News of Commission I.)
17. JACQUET, J. 1962. Les études d'hydrologie analytique sur bassins versants expérimentaux.
Bull. du CREC Chatou, No. 2.
18. ; CORMARY, Y. 1965. L'étude du cycle de l'eau sur un bassin d'investigation.
L a houille blanche, no. 3.
19. KAYE,G.W.C.;LAHY,T. H.1953. Tables of physical and chemical constants. London,
Longmans, Green.
20. KHROMOV, S. P.;MANONTOVA, L. I. 1963. Meteorologichesky slovar [Glossary of meteo-
rology]. 2nd ed. Leningrad, Gidrometeoizdat.
21. KOVZEL, A. G. 1959. Sostoyanie experimentalnykh issledovaniy protsessov formirovania
vesennege polovodia i perspektivy ikh razvitiya [State of experimental investigations of
spring floods and probability of their prediction]. Proc. Third All-Union Hydrological
Congress, 2 : 566-72.
I
22. LANGBEIN, W . B.; ISERI, K. T. 1960. General introduction and hydrological definitions.
(USGS Water supply paper 1541-A.)
23. LASZLOFFY, W. 1965. Terrains d'études dans les recherches hydrologiques hongroises.
IASH publ. no. 66, 1: 185-9.
24. LEOPOLD, L.B. et al. 1966. Channel and hillslope processes in a semiarid area, p. 193-252
(USGS Prof. paper 352-G.)
25. LINSLEY,R. K. et al. 1949. Applied hydrology. New York, McGraw-Hill.
26. MAKKAVEEVE, A. A. 1961. Slovar po gidrogeologii i inzhenernoy geologii [Glossary of
hydrogeology and engineering geology]. Moscow, Gostoptehizdat.
27. MEINZER, O.E. 1923. Outline of ground-water hydrology. (USGS Water supply paper 494.)
28. OURYVAEV, V. A. 1953. Experimentalnye gidrologicheskie issledovania na Valdae [Experi-
mental hydrological investigations in Valdai]. Leningrad, Gidrometeorizdat.
~ 29. .1956.L'étude expérimentale des éléments du bilan d'eau et des procès de formation
de l'écoulement.Addit au Reczieil des articles pour le Congres Intern. Geogr. Leningrad,
Ed. Hydrometeorologiques.
41
30. POPOV, O.V.; KUZNETSOV, V. I. 1960. Experimentalnye issledovania elementov vodnogo

balansa sushi [Experimental investigations of water balance components]. In: Teplovoi i
vodny rezhim zemnoi poverkhnost, p. 48-61. Leningrad,Gidrometeoizdat.
31. RODE,A. A. 1965. Osnovy uchenia o pochvennoi vlage. Tom 1. Vodnye svoisrva pochv i
peredvizhenie pochvennoi vlagi [Principles of soil-moisture study. Vol. 1. Soil physical
properties and soil-moisturetransfer]. Moscow, Academy of Sciences of the U.S.S.R.
32. SALAMIN,P.;GODA, L. 1965. Les terrains d'études comme moyens de la recherche hydro-
logique. IASH publ. no. 66, 1 :75-83.
33. SOILSCIENCESOCIETYOF AMERICA. 1960. Supplemental report of definitions approved by
the Committee on Terminology.Proc. SSSA,24(3) :235.
34. SZESZTAY, K. 1965. On principles of establishing hydrological representative and experi-
mental areas. IASH publ. no. 66, 1 : 6474.
35. TOEBES, C. et aí. 1965. Glossary of terms. Procedure No. 19 (Handbook of hydrological
procedures). Wellington,New Zealand, M.O.W.
36. UBELL, K. 1966. Hydrology of groundwater. Research Inst. for Water Resources. (Series
of manuals, no. 4, for international post-graduatecourse.)
37. VEN TECHOW.1964. Statistical and probability analysis of hydrologic data. In: Ven Te
Chow (ed.). Handbook of applied hydrology, section 8-1.
38. WOLMAN, M. G. 1955. The natural channel of Brandywine Creek, Penn. (USGS Prof.
paper 271.)
39. WORLD METEOROLOGICAL ORGANIZATION. 1965. Guide to hydronieteorological practices.
(WMOno. 168, T.P.82.)
40. ANON.1960. Geofogechesky slovar [Glossary on geology]. Moscow, Gosgeoltekhizdat,
2 vols.
41. ~ . n.d. Slovar-spravochnik gidrotel hnika-meliovatora [Manual of hydraulics and
reclamation]. Moscow, Selkhozgiz.
42. ~ . 1959. Spravochnoe rirkovodstve gidrogeologa [Guidefor hydrogeologists].Lenin-
grad, Gostoptehizdat.
42
2 Selection and organization
of basin networks
2.0 General
Selection of basins is the first and most importantstage of the organizationof observation
and research on representative and experimental basins. A n incorrect selection of basins
may reduce the value of long-term observations and expensive instrumentation and
throw doubts on the results of the scientific research. It must be realized also that
selection of basins is very difficult.
Difficultiesare caused by the specific requirementsfor representativeand experimental
basins and the lack of clear quantitative criteria for proper selection.
The principal requirement for representative basins is representativeness (the
correspondence of their physiographical characteristics to those of the hydrological
region). It is very difficult to select a basin which is representative of all hydrological
features and to satisfy simultaneously all other requirements. Aspects of selection of
representative basins are discussed in section 2.3.
The greatest problem in selecting experimental basins is the difficulty of finding a
completely homogeneous basin in all respects (e.g.,a basin completely covered by forest
or entirely composed of sandy or clay soils, etc.). Details of selection are given in
section 2.4.
In general, the selection of basins should follow a preparatory and a final stage.
The preparatory stage should be the selection and analysis of topographical,historical
and other reference information for the following purposes : survey of research needs
(see section 1.4);selection of hydrological regions (see section 2.2); selection of basins,
either representative or experimental according to the research objective (see
sections 2.3 and 2.4).
2.1 Use of maps

A n initial requirement for the selection of hydrological regions and representative
and experimental basins is an adequate map coverage.The mapping needs will,in part,
depend on the hydrological research objectives,although a selection of maps for field
and office work is generally essential.
At the preparatory stage all cartographical material may be of use. Maps showing
hydrological and climatological networks should be used to evaluate the data available
on hydrology and climate of the area and this data should be analysed to determine
reliability,length of observation period, etc. Areas which have insufficienthydrological
43
Representative and expeuinlental basins
data and on which specific hydrological research is of great scientific and/or economic
interest should be delineated. In these cases factors such as the needs of the country
forhydrologicaldata on given hydrologicalregionsinconnexion with present or proposed
development, access, density of population, etc., should be taken into account.
For the selection of hydrological region maps (if available) on hydrological,climatic,
pedological, vegetational and hydrogeological characteristics should be used (see
section 5.1).
Representative and experimental basins are finally chosen from a collection of possible
sites. This is, in part, achieved by using a small-scaletopographical map to consider
features such as aspect and topography.
Geological, pedological,climatic and land-inventory maps of comparable scale are
also required to check the general physiographical characteristics of the basins. Details
of maps are to be found in section 5.1.
Once a basin has been selected, a most detailed topographical map of the area is
required as a base map. It must be sufficiently accurate to minimize analysis errors.
The map may be simplified (e.g. to show only drainage patterns) but its main use is
in the location and storage of data in the field. Completed survey maps should never
be discarded.
Maps of basin geology, pedology, climate and observational sites, if not available,
must be obtained before the major flow and precipitation sampling positions can be
determined. Additional maps of vegetation, land use and historical record may also
assist in establishing patterns within the basin. Historical maps are extremely valuable
in building up the historical background to the basins.
As the studies develop, similar,but more detailed maps are required and new maps
of basin characteristics will become available.Master sites,for example,require detailed
hydrological soil surveys.
2.2 Selection of hydrological regions

Hydrological regions are relatively homogeneous subdivisions of an area or a country
and in each of such regions hydrological similarity is assumed. Accurate classification
of hydrological regions is complex, since hydrological classification of an area with
regard to one purpose or in respect of one phase or aspect of the hydrological cycle
may be quite different from the classificationfor another purpose or aspect;classification
of erosional characteristics will, for instance, involve factors such as rainfall intensity
and duration,infiltration capacity,vegetative cover,erosional resistivity of the soil and
sediment rating. Provided the soils are not impermeable, a low-flow classification is
governed mainly by the characteristics of geological strata and rainfall. Moreover,
even within a small basin, small areas with different hydrological characteristics may
be found. For this reason every hydrological division is approximate only and is based
on the most important factors.
A classification is, however, essential so that representative basins may be selected
and whether or not the regionalboundaries are accurately drawn is not of great moment
except when representativebasins are established with the limited purpose of hydrological
prediction of one hydrological variable (e.g.,low flow), or of a few closely related ones.
Selection of approximate hydrological regions provides that basins which are
representative of some larger area may be selected and,if required,the selection of the
region and basin could be directed to the particular purpose for which the representative
basin is required.
For the initial selection of hydrological regions which are used for the establishment
of representative basins, natural physiographical regions should be used. As a rule,
natural physiographical regions are sufficiently characteristic in many hydrological
44
Selection aiid organization of besitz rietworks
aspects. It should be noted, however, that principles of selection of physiographical

regions may vary from country to country [12].
The size of a region depends somewhat on the purpose for which a representative
basin is to be established and on the physiographical conditions of the country. Large
areas with similar physiography, for instance,could be subdivided into regions as large
as say 20,000km2.Smallmountainouscountries,especially thosewith maritimeinfluences,
should consider some regions as small as 1,000km2.
A delineation of soil-vegetationcomplexes will aid in the selection of experimental
basins and in the selection of sites within basins.
I 2.2.1 Selection of hydrological regions in areas where no

I detailed hydrological data are available
I
Where detailed hydrological data for the selection of hydrological regions are not
available, selection must be on climatic, vegetational, geomorphological,pedological
and geological characteristics. Regions are basically delineated by making maps of
characteristics such as mean annual rainfall, potential evapotranspiration, rock
permeability, land slope, etc., and placing these on top of one another to define
common boundaries.
This is not always feasible, since coincidence of boundaries is highly dependent on
the relative classificationof the characteristics used. For instance,annual rainfall could
be classified into groups of say perhaps < 1,000mm;1,000-1,500mm;1,500-2,000mm,
etc., but no guarantee could be given that such classes form common boundaries with,
say,selected rock-permeabilityclasses.It may be that rainfall classes of say < SOO mm;
SOO-1,200 mm; 1,200-1,600mm,etc. give better results, but the solution is normally
somewhatmore complex,and an approach to find the classificationof all characteristics
selected which will lead to maximum coincidence of boundaries would be a major task.
There is, in addition, the problem of obtaining quantitative estimates of all the
characteristics selected.
The simpler solution is to use one or two characteristics as a base and to adjust
boundaries or draw sub-boundariesif other characteristics have a locally dominant
influence.
The most important characteristics to use are climate and/or geological-pedological
ones. Geomorphological features correlate reasonably with the geology, but the soil
and vegetational characteristicsare too difficult to handle unless the ‘natural’vegetation
is considered.
If climate is used as the base, average annual rainfall, intensity-frequency-duration
relations, average annual temperature, precipitation-potential evapotranspiration
relationshipsor conventionalindicessuch as those proposed by Köppen or Thornthwaite,
are generally satisfactory. Small States within large land masses may, however, have
a relatively uniform climate and mountainous countries under extreme maritime
influencesmay have local climatic variations which defy a simple classification.Moreover,
climatic characteristics may be insufficiently known.
In such cases it is better to base delineation on geological-pedologicalcharacteristics.
Soils are usually based on geological features and will be found to correlate well except
where the natural vegetation varies; in such cases entirely different soils may have
formed on the same parent material and sub-boundariesmay be necessary.
Geological characteristicsmust be based on lithological aspects only and classification
must be either quantitative (in the form of a permeability index) or semi-qualitative.
Having drawn a base map of either climatic or geological-pedologicalfeatures,other
characteristics may be introduced, and sub-boundariesdrawn.
45
2.2.1.1 Example in Brazil

In the 80,000k m 2 Rio Jaguaribe basin (Brazil), it was considered necessary to establish
representative basins with the specific purpose of predicting mean and flood flows.
The basin has a homogeneousclimate,humid,tropical with dry winters and an isothermic
tendency (less than 5” C variation between the mean temperature of the warmest month
and the mean temperature of the coldest month) and falls under type Awi of Köppen’s
classification.
In this homogeneous climate slight variations in the mean annual precipitation were
classified as follows and boundaries drawn accordingly:
PI Mean annual precipitation < 600 mm (this covers an area of about 11,500km2 at the
extreme west of the basin);
PZ Mean annual precipitation 600-800 mm (this covers nearly 80 per cent of the basin);
P3 Mean annual precipitation > 800 mm (this covers about 5,000km”.
Next the lithological conditions were studied from the geological map. Apart from
three limited sectors of sedimentary terrain covering a small area in the extreme south,
atthecentreand in theextremenorth-east,the whole of the basin was made up of normally
impermeable pre-Cambrian terrain. Three classes of ascending permeability were
adopted as follows:
KI Impermeable terrain with a pre-Cambrian base;
KZ Slightly permeable terrain with localized ground water;
K3 Permeable terrain.
Topography was subsequently considered, using a mean slope factor computed from
I=Lx/L, where Iis the mean slopefactor and a the variation in basin elevation(considering
only those altitudes which have approximately 5 per cent of the superficial area of
the basin above and below them) and L the length of the equivalent rectangle of the
basin. This slope factor was computed for basins from 300 to 500 k m 2 within the
Rio Jaguaribe from a 1 :500,000map, making no allowance for the larger rivers. This
resulted in four classes of relief for an area of 500 km2:
SI Very slight slopes, < 1.25m/km.;
SZ Slight slopes, between 1.25and 2.5 m/km.;
S3 Moderate slopes,between 2.5 and 5 m/km.;
S4 Reasonably steep slopes, between 5 and 12.5 m/km.
No other factor was introduced; the vegetation was uniform, partly savannah with
scattered scrub (caatinga) and partly cultivated vegetation which did not effectively
protect the soil during the rainy season.
The various boundaries of climate, geology and geomorphology were traced on a
1 :100,000map and the pairing of the factors PIand P3, KIand K3, and SI and S4
revealed eighteen hydrological regions as follows:
1. P~K~SI 7. P2KzSi 13. PzKiS4
2. P3KzSa 8. PzKzS3 14. PIKYSI
3. P3K2S4 9. PzKzS4 15. PIK&
4. P3KiS3 10. PzKiSi 16. PiKiSi
5. P~KIS~ 11. PzKiS2 17. PIK&
6. P2K3Si 12. P2KiS3 18. PiKiS4
For the above study it was considered that a representativebasin established in P3KzS4
would also represent all groupings of KZwith S3 or S4 irrespective of the value of P.
A basin within P~KISI could be used for the groupings of Ki with Si and SZirrespective
of the value of P.The K3Si combination stands on its own and the remainder (the
larger area of the basin) can be covered by KiS3 or KiS4 [I, 21.
46
Selection and organization of basin networks
2.2.1.2 Example in New Zealand

In N e w Zealand it was desired to set up a network of representativebasins,all of which
would serve the purposes of hydrological prediction and extension of records and
some of which would serve for studying the effect of natural changes on the hydrological
regimen and for fundamental research. N e w Zealand is a steep, mountainous country
of 256,000 kmzand is under strong maritime influences.Consequently a basic division
into hydrological regions on the basis of climate is impossible because of the extreme
variability. Most valleys have their own peculiar climate (meso-climate) and adequate
records are not available to make any approximate classification.Since the topography
is extremely broken in many places and the natural vegetation for a large part removed,
the basic classificationwas done on the basis of lithologicalfeatureswith local adjustment
to soil groupings which showed a specific impermeability (podzolized heavy clays)
and this map was initially drawn on a scale of 1 :200,000 [8].Because of variations
in precipitation range, geological type and age, it was necessary to classify the North
and South Islands separately in both precipitation and permeability classes. The
lithological classes were defined as follows:
South bland North Island
Ki Gneiss, granite Greywacke, argillite,mudstone
K2 Alluvium (impermeable), mudstone, Schist, basalt, dacite, gabbro
greywacke (massive), shale,argillite
K3 Schist,basalt, sandstone, greywacke Rhyolite, sandstone, ignimbrite,
(shattered), rhyolite andesite
K4 Andesite, alluvium, conglomerate, Alluvium, conglomerate,limestone
limestone
K5 Gravels, sand, pumice Gravels, sand, pumice
Further divisions were subsequently drawn to show mean annual precipitation.Classes

were defined as follows (in millimetres):
South Island North Island
Pi 400- 700 < 1 100
P2 700-1 O00 1 100-1 300
P3 1000-1 300 1300-1 600
P4 1 300-2 O00 1 600-2 O00
P5 2 000-4 O00 2 000-2 500
PS > 4000 > 2 500
These isohyetal construction lines had somewhat different values in different parts of
the country as they appeared to coincide with topographical and/or land-usefeatures.
For instance, in some parts of the country Pi,P3 and P4 were used and in other parts
Pz,P4 and P5.
Topographical features were characterized by means of slope classes. The slope is
taken as the average slope of the dominant topographical form. A dominant form is
one which dominates 80 per cent of the region.
If major topographical form did not constitute 80 per cent of a region,an average-fall
form was taken.The classes defined are: Si,0"-5";SZ,5"-10";S3, 10"-15";S4, 15"-25";
S5, > 25".
Finally the map was redrawn on a scale of 1 :500,000 on which small separate
hydrological regions of somewhat less than 100 kmz were ignored. Where boundaries
appeared to follow closely a basin divide the latter was adopted as the regional boundary.
Eighty-eight hydrological regions were defined. Figure 2.1 shows one of the simplest
47
areas of N e w Zealand [9],and for these regions the classificationand dominant land
use are:
Region Dominant land rise
Peninsular region Grazing

Canterbury Plains Cropping
Canterbury Foothills Grazing
Eastern Alps Tussock/native forest
Glacial region Bare rock/glaciers
Western Alps Native forests
West Coast Lowlands Grazing
FIG.2.1. Map showing hydrological regions of part of the South Island,New Zealand.
2.2.2 Selection of hydrological regions in areas where detailed

hydrological data are available
The establishment of a network of representative basins in an area or country where
a hydrologicalnetwork is already in operationoffers the opportunityto selecthydrological
regions with greater accuracy. Better and more detailed indices could be used in the
selection,and it is especially desirableto use those indiceswhich are proposed as variables
in subsequent hydrological research [6,81.
A statistical treatment of streamflow and precipitation data in the selection of
hydrological regions could be carried out to fix the precision required for observations
on representativebasins and to lay down approximate periods for which representative
basins should be operated.
2.2.2.1 Example in the U.S.S.R.

Divisioninto hydrologicalregionshas been carried out with respectto small and medium
basins representing zonality features and other typical landscape elements. That is why
I for such basins interpolation and extrapolationof hydrologically similar characteristics

were used.
When dividing into hydrologicalregionsit was taken into accountthat the stream-flow
regimen is affected on one hand by climatic elements (precipitation, temperature) in
48
&ekctioi~ end organization of basin networks
combination with other physiographical factors (relief, geology, soils, vegetation) and
on the other hand by works carried out for economic reasons in dra'inageareas and
river channels. The division was based on hydrological observation data obtained at
base stations [4,li].
In dividing the territory, the following characteristicswere used :
1. Hydrological and climatic features, taking-into account the genetic type of hydro-
logical regimen and character of the stream recharge:
Type I-rivers with floods caused by snowmelt;
Type II-rivers with floods caused by snowmelt and rainfall;
Type III-rivers with floods caused by rainfall.
Periods of low water with ground-water recharge were also taken into account.
2. Orographical featurestaking account of the relief of basins and surroundingterritory.
3. Zonality features taking account of the hydrological regimen of rivers in plains
and in mountains of the main geographical zones.
In dividing into hydrological regions the following features were considered:(a) homo-
geneity of conditions of run-off(annual, seasonal), amplitude of flood peaks (snowmelt
and rainfall) and minimum (summer and winter) discharges; (b) variations of annual
discharge; (c) precipitation distribution in the region; (d) mean annual and annual
discharge variations; (e) duration of ice effects, dry and cold periods, periods of no
flow for intermittent and wholly frozen rivers.
Indices of drainage density, basin shape, river slope, lake and swamp type, karstic
features, etc., were also used.
The division was made by analysis, carried out in the following order, of annual flow
hydrographs:
1. The river regimen was studied by comparing flow hydrographs for the whole period
of observation, first those of individual rivers and then those of adjacent ones.
2. Rivers with homogeneous regimens were distinguished and were then classified into
separate types (groups) or subtypes (subgroups) of hydrological regimen.
3. Areas with riversof homogeneoushydrologicalregimenwere indicated on hypsometric
maps and their boundaries plotted. (Care was taken to ensure that regions were
neither too large nor too small so that regimen features could be relatively stable.)
4. Boundaries of regions were compared with distinct physiographical features (relief,
climate, geological structure, soils and vegetation).
5. Hydrological indices and their correlation with physiographical factors within the
region and between individual regions were indicated.
After the division was finished,a list of regions with brief, characteristic main features
of the hydrological regimen and river discharge was made.
To illustrate hydrological division,that of the Upper Volga territory (including the
Volga basin up to the town of Cheboksary) is given. In accordance with the present
hydrologicalmacro-division,the riversof the regionare referred to four largehydrological
regions (macro-regions), i.e.,to rivers of forest zone (Ly1 and Lyz) and to rivers of
steppe-temperatezone (Cy1 and Cy,).
Zonal differences between these regions are caused by the fact that, while rivers
of the first two (northern) regions (Ly1 and Lyz) are characterized by the presence of
spring snowmelt floods and summer and autumn rainfall floods, rivers of the second
two (southern) regions (Cy1 and Cyz) are characterized only by spring snowmelt floods
and summer rainfall floods.
Interzonal differences between the eastern and western regions (i.e., between Lyi
and Lyz and between Cy1 and Cy,) are caused by the fact that the boundary of winter
thawing weather lies between them (i.e.,January mean monthly isotherm -10" to -12").
This is why, in the more western regions (Lyz and Cyz), there are not only summer and
autumn floods due to rainfall but also winter floods due to thaw which are rarely if
ever found in the eastern regions (Lyl and Cy,).
49
Representative and experiniental basins
Insolvingmany hydrologicalproblems it is expedientto perform a detailed hydrological

division of the territory according to river regimen, taking account not only of zonal
conditions but also of local peculiarities.
In the region in question there are hills, plains and lowlands and each river is
characteristicof the surrounding features.Naturally,the more distinctthe physiographical
differences, the more obvious the differences in flow regimen. Thus, the Upper Volga
region may be divided not only into the four above regions,but also into several smaller
subregions where rivers can be characterized by more homogeneous hydrological
conditions (see Fig. 2.2).
Water-balance elements calculated for the given hydrological regions are tabulated
on the next page.
Boundary between forest and steppe belt
---_- Limit of winter floods
Boundary of hydrological regions
1, 2, 3, 4, 5, 6, 7 Hydrological sub-regions
Hills
FIG.2.2. Map of hydrological regions of the Upper Volga territory.

Selecrion arid üïganizution of basin networks
Index of the region Precipifalion (mm) Run-off (mm) Evaporalion (mm) Run-ofl coeficienf
LY2 750-810 210-250 540-560 0.27-0.32

CY2 700 150 550 0.21
LY1 680-760 140-210 540-550 0.21-0.28
CY1 600-650 100-120 500-530 0.17-0.18
2.2.3 Delineation of soil-vegetation complexes

Soil and vegetation are so closely interlinked that they form a complex.O n the one hand,
soil influences the vegetation through site characteristics of mineral materials, climate
and drainage (the inorganic cycle); on the other,vegetation influences soil by the action
of plants and associated organisms in controlling the rates of uptake and return of
elements to the soil through the organic cycle. When the correlations between soil and
vegetation are established and understood, field observations on one factor can be
used to indicate conditions in the other. Thus certain associations of broad-leaved
trees or grasses indicate free-draininggranular soils with a thin surface litter of organic
matter,whereas certain associations of fine leaves,trees or grasses indicate the opposite
properties. There are many intermediate patterns which require examination of the
geological and vegetational history before correlations can be relied upon to extend
site observations to larger areas. Nevertheless, the widespread value to be obtained
from the use of soil-vegetationcomplexes as indices of site conditions makes such
correlations highly desirable.
For an approximate delineation of soil-vegetationcomplexes for use in representative
and experimental basin research, delineations should be differentiated according to
whether sites or basins are being selected. The basic unit of classification is the soil
type,which has a narrow range of properties. The vegetation can be similarly grouped
in small units on the basis of phytomorphological characteristics (see section 4.6).
These combinations of soil types and vegetation units are the best criteria for the
selection of sites within basins.
For the selection of experimental basins, the variety of the soil characteristics for a
broad pattern of secondary classification can be demonstrated by describing them in
groups. Each group consists of soils which are formed by similar processes and which
are approximately in the same stage of development. In this way the properties of soil
types can be linked into a progressive order for interpretation [3].
The combination of such soil groups with vegetation classes based on species surveys
will give a simple grouping of soil-vegetationcomplexes.
2.3 Selection of representative basins

Having delineated hydrological regions it is necessary to select a representativebasin for
each region.Countries which have great physiographical variations need a dense network
of basins,while countries which are rather uniform in character need only a few basins.
The distribution of these basins over large areas will therefore be non-homogeneous
in space.
The selection of representativebasins is governed to a certain extent by their purposes
(see section 1.3) and either a compromise must be made in the selection or a representative
basin must be selected for one or two specific purposes only. In some countries the
representative basin network can serve partly or wholly as the basic network of
gauging stations [131.
(a) Representativeness. It is most important to consider whether the basin selected
representsthe hydrologicalregion.As a guide,the type and range ofclimatic,vegetational,
51
geomorphological, pedological and geological characteristics of the basin should be

compared with those of the hydrological region.A simple way of doing this is to select
a basin which has the same classification as the region;for instance, if a hydrological
region has been defined as P ~ K zwhereS ~ P3 represents a class of mean annual rainfall,
Kz a rock permeability class and S4 a relief class (see section 2.2), then a basin with
the same classification should be selected.
It must be realized, however, that successful basin selection depends, a priori, on
the relative success of the definition of hydrological regions (see section 2.5).
(b) Basin divide. The basin divide should be as distinct as possible for the exact
determination of the basin areas.
In some cases, if a basin is suitable in all other aspects but the basin divide is not
clear, an artificial divide can be constructed by means of small dams, etc.
The ‘surface’basin divide should also be coincidentwith the ‘subsurface’basin divide.
Where possible, stream piracy should be prevented by artificial means. Note that with
a decrease of the basin area, the relative influence of non-coincidence of surface and
subsurface basin areas will increase.
Basins with coincident surface and subsurface divides are difficult to find but, if
this requirement is not met, at least for the upper aquifers, great problems occur in
water-balance calculations.
(c) Constancy of condition. The cultural changes in land use, land management,
s t r e d o w utilization, etc.,should be minimal during the period of study and, where
they are inevitable,should be carefully recorded (see section 5.1.6). Where representative
basins are designated benchmark basins, it is imperative that cultural changes be
prevented entirely and that natural changes be minimal.
(d) Deep percolation and channel infiltration. The loss of subsurface flow by deep
percolation, or the gain of this flow from neighbouring basins, must be as small as
possible unless the study of the representativebasin is,per se, the study of such leakages
(see section 4.3.2.1).
(e) Quality of flow-measuring station. It is essential that the stage-dischargerelation
is relatively constant. For this purpose, the site for a gauging station should have a
natural control or, if this is not available, an artificial control should be constructed
(see section 4.3.4).Site selection should be carried out according to standards laid down
in the literature [13].It is important that flow characteristics as they relate to channel
features are typical of the region except in arid or semi-arid zones where it may be
advantageous to select, if possible, a basin that has perennial or intermittent flow,
or at least has some subsurface flow during and shortly after storms.
(f) Access. Access to the gauging station should be available for every streamflow
condition. Access in the basin should be such that precipitation and other climatic
observations can be carried out. In very difficult terrain,consideration should be given
to wholly automated recording of these variables or to the use of helicopters.
(g) Coincidence of representative basin with economic development of basins. Provided
that this condition does not clash with the proviso of (c) above, representative basins
can be basins which are being investigated for economic development. It is important
to consider, however, that measurement and analysis must, in the first instance,be
geared entirely to the purpose of the representative basin. If this is not possible then it
is better to establish a separate representative basin.
(h) Size of representative basins. The size of the representative basin depends on
the purpose for which the basin is being established, subject to the proviso given in
section 1.2.If hydrological prediction or extension of records within the region is the
purpose,a basin of a size typical of basins in the region is desirable.If several basins
are selected to serve concomitantly for hydrological prediction (for instance to develop
flood-predictionmethods), a range in size of basins is desirable [lo].If the study of the
effect of natural changes on the hydrological regimen or fundamental research is the
52
purpose, a small basin will afford a better instrumentation provided that this does not
lead, at least in arid or semi-arid zones, to basins where subsurface flow cannot be
calculated.
In general, the representative basin should be so small that its sensitivity to high-
intensity rainfalls of short duration is not suppressed by channel characteristics.
2.4 Selection of experimental basins

The selection of experimental basins is governed,to a certain extent, by the methodology
of research to be used (see sections 6.2and 6.3).It should be decided beforehand whether
a single basin, replication of basins or a nesting arrangement of basins is most desirable.
(a) Single basin. In this case, selection is subject only to the conditions set out
below. The size should be as large as possible to increase the likelihood of measurable
subsurface flow and to approach more closely a natural basin, consistent with the proviso
given in section 1.2 that it must have a maximum size of about 4 km2. Single basins
are not recommended because no possibility exists of comparing results.
(b) Replication of basins. Normally two basins are selected, one of which is to
serve as a control during the period of the experiment. The two basins should be as
similar as possible with respect to climate, geomorphology, pedology and geology.
The most important geomorphological factor is the aspect of the basins; similarity
of aspect will increase the likelihood of similarity of climate. The maximum sizes of
the individual basins are somewhat less than those for single basins; in some areas
it may be difficult to find similar basins of a reasonable size. Two'relatively large basins
will, moreover, extend instrumental and observational problems. The two basins need
not be the same size. M e a n monthly discharges of two similar basins may correlate
reasonably if the difference in size is no greater than a factor of ten E].
A disadvantage of this method is that the control basin m a y itself be subject to
significant changes and, apart from the difficulties of handling this non-stationarity,
results of the statistical analyses may be invalid.
A variant of duplicate basins is multiple replication in which normally a multiple
of two similar basins is selected.Although statistically this method may appear to have
advantages, it is frequently difficult to select a set of adjacent basins so that each basin
is of a reasonable size and that all are continuously satisfactory from a hydrological
point of view. Multiple replication is, in addition, still hampered by changing controls
and the expenditure of carrying out the research is frequently so great that the use
of this method should be considered carefully.
(c) Nesting arrangemeni ofbasins. This is a variant of the single basin. O n e or more
sub-basinsis instrumented and regarded as a separate entity. Ideally, two or more basins
on the main stream are chosen and the same cultural change is applied to basin and
sub-basin(s).
This method provides for a better statistical control than the paired-basin approach,
since basin and sub-basin(s), being treated alike, can be subjected to the same approxi-
mating procedures for describing the relative stationarity before and after treatment.
A nesting arrangement has the additional value that differences between small and
large basins can be studied. This is invaluable for translation of the results. In some
cases the nesting arrangement can be combined with basin replication. This may be
done in one basin where the nesting arrangement will then provide a summary control.
Important items for selection of basins are listed below.
(a) Ownership of basin. The first requirement of an experimental basin is that it
can be manipulated at will by the research organization since the purpose of an
experimental basin is, a priori, the artificial change of one or more of its basin
53
characteristics. Unless the basin can be wholly owned or leased for the duration of
the experiment, it is not worth while to proceed with the research.
(b) Operation of basin. The topographical and access conditions must be such that
the required land-useand land-managementpractices to be used can be carried out.
Access must be good for detailed research observations and in some cases opportunity
must exist for buildings,either temporary or permanent,to be erected for management
and/or research staff. Since experimental basins are, by definition,small basins, roads
within the basin can have a very large effect on the hydrology and should be kept to
an absolute minimum and be installed before the experiment is started.
(c) Uniformity of soil, vegetation and geomorphology. Where possible a basin should
be a simple soil-vegetationcomplex based on broad soil-vegetation groupings (see
section 2.2.3). While a definite methodology for research on experimental basins is
not clear (see section 6.3), the selection of a basin on a soil-vegetationcomplex which
is typical of a large area will aid in the ultimate translation of the research results.
To facilitate research,a basin with simple geomorphologicalcharacteristics must be
selected if possible.Many depressions (small swamps) will,for instance,create storage
problems which are difficult to handle in analyses.Basins with many ‘cones’(sometimes
found in volcanic areas) are likewise difficult to handle analytically.
(d) Deep percolation and channel infiltration. This requirement is similar to that
for representativebasins.Deep percolation is,however,relatively large on small basins,
especially on the more permeable soils, and must be accepted as a variable factor.
The prevention of channel infiltration is relatively more important on small basins
and, if channel infiltration occurs, either an alternative basin must be selected or the
channel sealed.
(e) Aspect and climatic variability. In mountainous regions it is desirable that the
basin aspect is such that very shady and sunny faces do not exist simultaneously in
the basin, since moisture conditions may be entirely different on these different faces
and make analyses of data very complicated.Where replicationof basins is used,basins
should, if possible, have a similar aspect.
(f) Flow measurement. Where possible perennial or intermittent flow is preferred
to provide a measure of subsurface flow during periods of run-off.Where conditions
are such that only ephemeral streams are encountered, it is important that the basin
size be large enough to ensure that during and shortly after stormsthe totalflow measured
at the gauging station includes some interflow. A sharp distinction between surface
and subsurfaceflow is artificial and some interflow must be measured to carry out basic
research and to aid in the ultimate translation of the research results to other areas.
Basins should be provided with a self-rated measuring structure providing for
continuous, accurate measurement of flow (see section 4.3.4).
(g) Provisional observations. Before a basin is finally selected it is advisable to make
provisional topographical, geological and pedological investigations (see section 5.i).
It is also often useful,before the basin is finally selected,to carry out some provisional
flow and precipitation measurements to estimate possible floods and minimum flows
and to estimatewhether flow and precipitationcan be measured with sufficient accuracy.
2.4.1 Selection of run-of plots

Run-offplots are plots on slopes isolated by artificial walls from the surrounding area.
O n such plots the surface run-offregimen in its initial phase may be studied and the
influence of individual factors of run-off,such as steepness, length and exposure of
slopes, variations in soil and vegetative cover and in soil cultivation, etc., may be
determined.
Run-offplots permit the study of run-offprocesses. If the objective is the study of
54
the effect of a cultural change on the hydrological regimen to aid the translation of
results to natural basins, it is desirable to set up a number of plots of varying sizes
(these could even include natural basins). The establishment of run-off plots is not
recommended except when there are operating experimental basins available.
Plots should be organized on slopes within an experimental basin and should occupy
a whole slope or part of a slope. They may differ in area from tens of square metres
up to several hectares, depending on the experimental programme.
In arid zones where the water yield per unit area is very small, the dimensions of
run-off plots may be greater than those in humid zones. The increase in area of plots
provides an increase of accuracy of the results obtained, because of the increase of
measured volume of flowing water and, consequently, decrease of the relative error.
Very small run-off plots, especially those with artificial boundaries, should not be
used, because their necessary isolation from their surroundingsprevents their conforming
to a natural state.
In some cases the number of plots is considerable; this depends on the number of
factors studied, on the number of experiments and on the finance available.
From an economic point of view it is expedient to locate plots in pairs or groups
with mutual boundary walls, thus reducing the cost of their construction.
Principal requirements for the selection of sites for run-off plots are :
1. A natural slope surface not disturbed by soil work and roads and free of hillocks,
saucer-form depressions and other kinds of relief roughness causing a distortion
of natural flow formation.
2. The possibility of locating boundary side walls, perpendicular to contour lines.
3. Even slope, free of sudden breaks over the whole plot.
4. Equal bedding depth of a confining stratum.
5. The absence of ground-water outflow to the soil surface.
6. Homogeneity of vegetal cover, soils and subsoil over the whole plot.
7. The possibilities of collecting all the water flowing from the plot, of passing it to
the gauge and of draining the water flowing to the plot from adjacent slopes.
In solving individual problems the requirements need to be specified and extended.
While selecting sites for run-off plots, it is necessary to provide for the possibility of
performing observations on them under both natural and experimental conditions,
e.g.,using artificial sprinkling,varying soil moisture content, cultivating soil by modern
methods and performing other agrotechnicai and forest-improvement arrangements.
2.5 Analysis of a basin’s representativeness

At present the problem of criteria for basin representativeness cannot be solved in
detail and, numerically, its solution may be only approximate.
Basin representativeness should be estimated with regard to the problems for the
solution of which they are selected.
If a basin where observations have not previously been made is selected to obtain
data on the hydrological regimen and water balance of a certain hydrological region,
representativeness may be estimated only by means of a comparison of the basin’s
physiography with that representative of the entire region. The main characteristics
which should have the same or similar values for the basin and the region are climatic
conditions, the origin of slope formation (e.g., glacial, tectonic, etc.), structure and
erosion channel, depth of drainage network (in plains regions), geological structure,
ratio of aerea1 distribution between various natural and artificial zones (including arable
land). Principal land characteristics which are recommended for a study of the
representativeness are kinds of weeds, meadows, steppes, deserts, salt marshes and
saline soils, peaty moors, agricultural crops and their various combinations and
sequences (see section 2.3).
55
Maintenance of similarity of these characteristics between the basin and the region
it represents must guarantee the future satisfactory representativeness of the basin.
Wrong selection of representative basins is often due to insufficient knowledge of the
characteristics of both the entire region and the selected basin. For this reason reliable
estimation of basin representativeness must be based on more detailed data of these
characteristics or on special field investigations.
In selecting and operating basins to study the effect of individual catchment
characteristics on the hydrological regimen, estimation of representativeness is based
on the utilization of observations already made during definite time intervals. The
principal criteria determining basin representativenessare the degree of possible influence
of the given catchment characteristic on the unknown hydrological parameter (established
on the ground of Observationaldata) and a comparison of this influencewith the probable
error of this parameter.
References
1. DUBREUIL, P. 1964. The layout of basins representing homogeneous hydrological regions.
Trans. Inter-Regional Seminar, Bangkok. (UN Water resources series, no. 27.)
2. ___ . 1965. Contribution a I’étude d’implantation de bassins représentatifs de régions
hydrologiques homogènes. Paris. (Cahier Orstom d’hydrologie, no. 2.)
3. GIBBS, H.S. 1963. Soils of New Zealand and their limitations for pastoral use. Proc. N.Z.
Inst. Agric. Sci., 9.
4. Kuzm, P. S. 1960. Klassifikatsia rek i gidrologicheskoe raionirovanie SSSR [River classi-
fication and territorial division of the U.S.S.R.into hydrological regions]. Leningrad,
Gidrometeoizdat.
5. LANGBEIN, W . B.;HARDISON, C.H.1955. Extending streamflow data. Proc. ASCE, paper
826, vol. 81, 13 p.
6. LÁZLÓFFY,W.;SZESZTAY, K.1963. Organization of a hydrologic service. UN Conference
on the Application of Science and Technology for the Benefit of the Less Developed Areas.
Geneva.
7. SZESZTAY,K. 1965. O n principles of establishing hydrological representative and experi-
mental areas. IASH pubi. no. 66, 1 :64-75. (Symposium of Budapest.)
8. TOEBES, C.;NEEF, G.1962.Regional hydrology.Hydrology and land management, p. 76-80.
Wellington, S C and RCC.
9. TOEBES, C.1965. The planning of representative and experimental basin networks in N e w
Zealand. IASH publ. no. 66. (Symposium of Budapest.)
10. TOUCHEBEUF DE LUSSIGNY, P. 1965. Etude des bassins versants représentatifs de lo00 k m 2
en Afrique tropicale.ZASHpubl. no. 66,p. 320-4. (Symposium of Budapest.)Paris,Orstom.
56
‘ 3 Planning of observations
according to the research
obJectlves
3.1 General
The programme of observations o n representative and experimental basins depends
on the objectives of the research and on the natural conditions of the region.
It is naturally impossible to define a standard, uniform programme of observations
that would suit every hydrological region. Moreover, the observational programme
for representative basins generally differs from the programme for experimental basins.
Observational programmes for similar types of basin located in similar hydrological
regions may differ because of differences in research objectives, availability of finance
and personnel, etc.
A few general principles are given below which will aid in formulating a programme
for any one basin. It should be noted that the programme must be clearly defined
before any basin is established (see section 1.4).It may not be necessary,or even possible,
to study all the items listed in the examples given below. For instance, if snow does
not fall on the basin the need for snow measurement is obviated or if erosion is negligible
no erosion and sedimentation studies are required. In general, hydrological research
o n representative and experimental basins leads to two types of result: the estimation
of characteristic parameters, or comparisons by means of statistical tests. Consequently,
the value of the result and the degree of accuracy obtained depend in the first place
on the length of the observational period.
A reduction of the observational period may be made in many cases if correlations
can be established between hydrological variables from the research basin and other
basins in the same hydrological region (for instance, correlations of rainfall data from
research basins with rainfall data from the national network). If such correlations
are sufficiently high, the accuracy of the estimation of the relevant parameters (as
characterized, for instance, by a confidence band) may be considerably improved.
Representative and experimental basins normally have different observational periods.
O n representative basins observations are, as a rule, long term. They should include
at least several years (or periods) with high and low flows to establish the natural run-off
pattern typical of the area. O n experimental basins, the observational period m a y be
as short as a few years or seasons. This would apply, for instance, if the effect of soil
cultivation on the flow pattern were studied. In other cases, such as the study of the
effect of afforestation on the hydrological regimen, the observational period is perforce
of longer duration. In all cases, the need to obtain reliable results is of prime importance
and particular reference should be made to section 6.3 [7].
51
3.1.1 Observationalprogramme for representative basins

Representativebasins should,as far as possible,have identical programmes and methods
of observation to facilitate a comparison of the research results.Individualprogrammes
must be adjusted,however,to the particular purpose or purposes as specifiedin section 1.3.
It is necessary to distinguish between minimum and special equipment requirements.
It is essential to carry out a minimum number of observations on any one basin and
these include at least observations on streamflow and precipitation to study basic flow
processes.Details of minimum equipment requirements are given in section 4.12.After
certain minimum studies, a progressive increase in the observations can be made as
dictated by the early research findings.For instance,it may be found that,for a proper
rainfall-run-offstudy, information on soil moisture and infiltration is required. Such
special observations can then be added to the initial programme.
Such a progressive development of observations will ensure a better adaptation of
measurements to the problems studied and to the particular conditions in the basin.
Moreover,observations on representative basins must be carried out to suit the analysis
techniques. If, for instance, rainfall-run-offrelations are studied on a basin with a
relatively short record, storm rainfalls should be used rather than annual totals to
obtain a statistically valid sample and this aspect must be reflected in the observational
technique (for instance,rainfall amounts should be measured at the end of each storm).
For any observations which involvean element of sampling in space (see section 4.1.1)
a dense sampling programme for a limited period can be very useful. For example,
to study the representativeness of a given basic rainfall network in a basin, a large
number of gauges are installed for a limited period. After a short period the data is
analysed by correlation techniques and, in many cases, a representative small basic
network for the given level of accuracy required can be established (see section 4.2.1.2.1).
In addition to the observations of streamflow and precipitation,it is necessary to
make a preliminary survey of the basin to consider the physiographical conditions
which affect the hydrology. This survey should include the following: (a) an exact
determination of the basin divide and basin area;(b) a survey of geological,soil and
vegetative-cover characteristics of the basin; (c) description of the basin relief;
(d) description of the hydrological installation (for details, see section 5.i).
It is desirable to have a climatological station within the representative basin, but
it is possible to use a climatological station located outside the basin, provided its data
are representative for the given basin [7].
3.1.1.1 Observational programme for representative basins used for fundamental

research
Such programmes may need to be fairly complete observational ones but are naturally
dictated by the objectives. If, for example, the hydrological differences between shady
and sunny slopes are to be studied,the observational programme must include at least
the following for both slopes [i, 21: flow (frequently with the aid of run-off plots),
precipitation,micro-climate,soil moisture, and infiltration.
3.1.1.2 Observational programme for representative basins used for the study of
the effect, on the hydrological regimen, of natural changes (benchmark and
vigil basins)
The study of the effect, on the hydrological regimen, of a natural change (for instance,
climatic or geomorphological) is extremely difficult.The basin, and consequently the
hydrologicalcharacteristics,is in a continuous change and this ‘non-stationary’
condition
58
Planning of observations according to the research objectives
of the basin is difficult to analyse (see section 6.2.2). The observational programme
must therefore allow for rather precise observations of certain basic elements.
Climatic changes are usually long term and changes over a shorter period are within
the natural variability of observed phenomena. For this reason, a study of micro-climatic
changes may be more productive. As an example of the observational requirements
for an expected geomorphological change, the following programme should be regarded
as a minimum: flow, precipitation, climate, soil frost, erosion and sedimentation, and
geomorphology.
3.1.1.3 Observational programme for representative basins used for hydrological

prediction
Most representative basins will be used for this purpose and the programme should
specify whether long-term or short-term prediction studies are intended so that the
observational programme can be adapted accordingly. Moreover, different programmes
m a y be required according to whether or not the prediction requirements will affect
the period of observational and processed data, e.g., whether daily or weekly mean
discharges are required.
As an example, the minimum observational programme for a representative basin
used for long-term prediction of ground water could be as follows (unconfined aquifer) :
delineation of aquifer ; precipitation; ground-water observations ; temperature and
ground-waterquality;flow (base-flowstudies) especially at springs; infiltration (especially
the rate of percolation from the surface to the aquifer).
3.1.1.4 Observational programme for representative basins used for extension of

records
This is strictly a special case of hydrological prediction and the principles are similar
to those given in section 3.1.1.3. The important requirement is that observations are
made in a similar fashion to that for the short-term station for which extension of
records is required.In general,the programme can be restricted to flow and precipitation
observations [S, 101.
3.1.2 Observational programme for experimental basins

Experimental basins are established to study the effect of cultural changes on the
hydrological regimen (land-useand/or land-management changes). Such studies have
some similarity with the study of the effects on the hydrological regimen of natural
change on representative basins (although experimental basins may frequently be
regarded as being in a ‘stationary condition’ before and after the change, simplifying
the analysis methods) (see section 6.3).
The combined analysis of the cultural changes and expected change in hydrological
characteristics or processes leads to a definition of the experimental design (see section 6.3)
and to choice of the statisticaltests which have to be used to control the level of significance
of the observed changes. This task has to be carried out before the definition of the
observational programme, otherwise it may be difficult to interpret the data correctly.
It will, moreover, aid in an optimization of results with the level of instrumentation
and observational period. Observational programmes are normally fairly extensive for
experimental basins and are generally dictated by the problems to be investigated rather
than by the variation in the physiographical conditions of the region as is the case for
representative basins [3, 5, 91.
The first requirement is a statement of which main parameters are to be controlled.
59
The study of the effect of various soil cultivationmethods on the hydrological regimen,
for instance,has as main parameters firstly the principal elements of the water balance
(irrespective of whether the basin is in a natural forest area, etc.), specific parameters
which are dictated by the problem (in this case, e.g., soil physical, micro-climaticand
similar characteristics) and the local physiographical conditions which may be
characterized by variations in soils,vegetation, climate,etc. Cultural changes are many
and the importance of each may vary from country to country. Important ones
recommended by the Unesco Working Group on the Influence of M a n on the
Hydrological Cycle are:
1. Afforestation, deforestation and modifications to forest vegetation.
2. The drainage of swamps by the control of surface water and the substitution of
useful vegetation for relatively useless vegetation.
3. Changes in vegetational cover imposed by the deliberate production of rain-grown
or irrigated crops.
4. Changes in surface water distributionby controland diversion for power production,
irrigation and other purposes.
5. Vegetational changes causing soil erosion and soil conservation works.
6. Imposition of urban and industrial development on natural conditions.
7. Changes in water quality due to a deterioration in vegetative cover.
8. Exploitationof ground-waterreservesleading to changesin surface-waterdistribution.
Section 6.3 describes general analysis techniques from which observationalprogrammes
may be derived (see also section 2.4, ‘Selection of experimental basins’). Particular
methods used in the U.S.S.R. to establish observational programmes on experimental
basins are given in section 3.3.
3.2 Observational aspects of principal hydrological

elements
Considering that the composition of observational programmes in a very important
item in the planning of representative and experimental basins, the determination of
the programme, especially in the initial stages, may be very difficult.
For this reason some examples on programming individual observations are given
below; these examples should not necessarily be included in every programme,neither
are they complete. However, they do illustrate the general ideas and may be helpful
in planning the most economical observational programme for any one basin [lo].
3.2.1 Precipitation
Precipitationis one of the most importanthydrologicalelements.The exact determination
of the amount of precipitation, its type (hail, snow, rainfall, dew, etc.), its origin
(convectional, cyclonic,orographical) and the knowledge of its spatial and temporal
distribution,governs the reliability of water-balancecalculations and the various relations
with hydrological characteristics (e.g., surface and subsurface flow,evaporation, soil
erosion, etc.).
Types of precipitation gauge and numbers required in any one basin are discussed
in section 4.2.1. The following measurements should be made on representative and
experimental basins :
1. Precipitation measurement daily, or twice daily, by a manual gauge with a visual
determination of the precipitation type (rainfall,snow,hail,etc.). The characteristics
of falling precipitation (storm, drizzle, etc.) and an indication of the continuity
(in hours) of the precipitation should also be recorded.
60
The use of manual gauges only on a basin provides basic data for: (a) the
determination of the mean basin precipitation for periods of one day or more;
(b) the developmentofrelations between precipitationand hydrologicalcharacteristics
for relatively long periods (month,season,year); and (c) the study of the formation
of floods on relatively large basins where the time lag exceeds one day.
2. Continuous recording of precipitation by means of recording rain gauges. Recording
rain gauges are essential on small basins for the determination of the mean basin
precipitation for periods of less than one day; for studies of flood formation on
basins where the time lag is in the order of hours or less; and for the determination
of intensity-frequency-durationrelations for various hydrological regions (for
instance, for estimation of peak flows for various frequencies for the design of
hydraulic structures).
3. Long-term measurements by storage gauge. These are manual gauges used in
mountainousregions and observations are made at intervals of more than one day.
A mixture ofthe three types of gauge is often employed. Utilization of the data available
for any particular purpose is described in section 5.3.1.
3.2.2 Interception
Interceptionobservations are made on vegetated areas to determine the net precipitation
(see section 1.5) that partakes in the infiltration or overland flow processes.Interception
is particularly important on small basins and may also be a major factor on experimenta1
basins where changes in vegetation are planned.
The amount of interception depends on the type of vegetative cover,its density and
stagein evolution,and also on the type,rate and falling characteristicsofthe precipitation.
Interception storage has a definite limit. Low-intensity precipitation, the amount of
which does not exceed the interception storage, is generally intercepted completely.
With high precipitation intensities,a relatively smaller amount is intercepted.Stem-flow
observations may be important in certain types of vegetation.
Snow is intercepted mainly by crowns of trees, especially conifers. Shrubby and
herbaceous vegetation greatly influences the interception of snow.
Interception involves complex sampling problems, especially in forested areas. A
failure of past investigations has been the inadequate description of the phytomorpho-
logical characteristics of the plots sampled.Details are given in sections 4.2.3 and 4.6.
3.2.3 Snow cover

In countries with heavy snowfalls and in high mountains, snow is one of the most
significant water-balancecomponents.In some regions,e.g.,in semi-aridand arid zones
with a continental climate,up to 80-90 per cent ofthe total volume of run-offis caused
by snowmelt. Snow investigations are of great importance for studying the formation
of spring floods, which are a major feature in streams which are snow-fed.
The principal problems of snow investigations on representative and experimental
basins are the determination of the following relations: (a) relations between spring
flow characteristics (run-off,instantaneous discharges and peak discharges) and snow
accumulation in winter; (bj relations between spring flood hydrograph characteristics
and snowmelt processes; (c) relations between spatial snow distribution and geo-
morphological and vegetational characteristics.
For the solution of such problems it is essential that programme planning of snow
investigationsinclude a considerationof the following:snow survey (see section 3.2.3.1);
61
snowmeltinvestigationsand water yield from snow (see section 3.2.3.2);and evaporation

from snow surface (see section 3.2.5.3).
3.2.3.1 Snow survey

Snow surveys are required basically to determine the maximum water equivalent of
the snow pack at the beginning of the snowmelt period.
By using either data on the maximum amount of accumulated snow over the basin
or an index, it is possible to calculate and predict the spring flood run-offand peak
discharges. In addition, snow survey data are used for the study of the regimen of the
accumulated snow during the winter and for the study of spatial snow cover distribution
as influenced by various physiographical features (vegetation, relief, etc.).
The length and location of the snow course depend on the size of the basin and the
detail required for the study of the snow distributionwithin the basin (see section 4.2.2).
In regions where the snow cover is stable during the entire winter period,snow surveys
are carried out once a month, beginning from the development of a settled snow cover
early in winter up to the maximum depth of snow cover late in winter. When the
snow reaches its maximum depth, snow surveys should be carried out 1-3 times a
month to ensure that the maximum amount of snow accumulated at the beginning of
spring snowmelt can be calculated. During spring snowmelt,when snow still covers
the ground, snow surveys should be made more frequently (say once a day), but
with fewer observation points. Note that such extensive snow surveys should be made
only when the conditions of spring-flood formation and the snowmelt intensity are
studied.
In regions with an unstable snow cover, where snowfall and thaw cycles occur
frequently during the winter, snow surveys are carried out several times each month,
but the frequency of observation depends on weather conditions,access and character
of the snow-coverregimen [7].
3.2.3.2 Snowmelt
To calculate and predict spring flood hydrographs it is essential to obtain data on
snowmelt intensity and on the rate of water yield from the snow cover.
A study of snowmelt can be made by any of the following methods: energy balance;
water balance (including simplified variants of this method) ; and empirical methods
using certain climatological factors, e.g., air temperature (see section 4.2.2.1).
Depending on the method used and the accuracy required,the programme planning
for snow investigations during snowmelt should include observations as follows.
(a) When the energy-balancemethod is used, actinometric observations are made to
determine the radiation balance of the snow cover and climatological observations
are necessary to determine the heat exchange in the atmosphere (turbulent heat and
moisture exchange). This requires observations on the following: downward all-wave
radiation (hemisphericalradiation of both short-and long-wavelengths),upward short-
wave radiation (albedo), net exchange of all-waveradiation (difference between upward
all-waveand downward all-wave)(see section 4.2.5.3), snow surface temperature,wind
velocity, air temperature and humidity at 200 c m above the snow cover. In regions
with a shallow depth of snow cover (lessthan 20 cm) and dominant solar type of snowmelt
(radiation snowmelt), additional observations of radiation penetrating into the snow
and absorbed by the soil surface are required.
(b) When the water-balance method is used the following observations should be
made: depth and density of snow cover,amount of liquid water content in snow pack,
precipitation, evaporation from snow surface, percentage of snow-cover area over
the basin.
62
(c) Standard climatological observations on wind velocity, air temperature and

humidity, total and low cloudiness,and snow-coversurface temperature are sufficient
to determine approximate values of snowmelt intensity using the simplified version
of the energy-balance method.
(d) Precipitation,depth and density of snow cover are sufficient if a simplified version
of the water balance method is used [7].
3.2.4 Condensation
Condensation is important only in desert and semi-desert regions with a continental
climate and great daily variations in temperature and, frequently, with sandy soils.
Some workers consider that the condensation on the surface.of frozen land is a significant
part of the total amount of moisture input in basins located in the permafrost zone.
In some mountain regions,intense hoar-frostformation is observed on trees and rocks
which are exposed to the wind.
D e w gauges (see section 4.2.1.3.6)are used for observation of condensation but
quantities observed lie frequently within the natural variability of the observations
and problems of the representativeness of the observations have not yet been solved.
3.2.5 Evaporation
Evaporation from the water surface and from the soil and snow cover and evapo-
transpiration are not necessarily to be studied on every representative and experimental
basin. In the simplest case, with the minimum observational programme for a
representativebasin,evapotranspirationmay be determined by the water-balancemethod
as a difference between precipitationand run-offwithout taking into accountsoil-moisture
conditions. This method is suitable only for long periods.
The study of evaporation for shorter periods requires a more complex observational
programme and in some cases may need detailed measurements of evaporation from
the soil and/or evapotranspiration,taking into account the dynamics of soil moisture,
and, in some cases, the utilization of the energy-balance method.
If several vegetative covers occur on any one basin, special investigations can be
carried out (especially on experimental basins) to determine the transpiration of various
types of vegetative cover by means of weighing evaporimeters and by determining
the difference between evapotranspiration and evaporation from the soil.
Observations of evaporation from the water surface and the snow cover are not
necessarily required on all basins, but on only a selected network of representative
and experimental basins, where such observations are carried out according to specific
problems and specialprogrammes.Climatologicalobservations,which usually accompany
evaporation measurements, include observations of precipitation, evaporating surface
temperature, air temperature, humidity and wind velocity. For the sake of economy
such observations are made at the same climatological station, located on the given
basin,or close to it. Similar principles may be used to carry out observations of energy-
balance components.
3.2.5.1 Evaporation from the water surface
Measurements of evaporation of water surfaces should aim at the following objectives:

estimationof the rate of evaporation,estimationof the temporalvariations,improvement
of evaporation measurements, and the study of evaporation processes.
Evaporation from reservoirs depends on the climate,the physiographical conditions
of the immediate surroundings and the geometrical characteristics. For this reason,
63
Repuesenlaiive and experimental basins
the following relations should be held in mind when measuring evaporation from
water surfaces:
1. Climatic conditions determine the duration of the evaporation period, the long-term
evaporation value, the interseasonal distribution and long-term evaporation
fluctuation.
2. Evaporation intensity depends on a combination of climatological factors.
3. The surroundings of the reservoir (relief, vegetative cover, height of shores and
their wind protection, towns and villages in the neighbourhood, etc.) govern the
transformation of the air flux coming from the land. This factor influences greatly
the intensity of evaporation from smallreservoirswith a surfacearea ofless than 5 km2.
4. Geometricalcharacteristicssuch as the area,configurationand depth of the reservoir
influence the evaporation on various parts of the water area.
To elucidate the above relations it may be necessary to carry out the following:
1. Observations by 20 m2 tank (to be organized only on selected representative and
experimental basins,if there are lakes and reservoirs within the area of these basins
and if they appear particularly suitable).
2. Observations by GGI-3000 pan (used on the station network in the U.S.S.R.),
by Class A pan (used on the station network in the U.S.A.), or by any other standard
pan, installed on land or in the middle of a lake on a raft. Observations by any of
these pans should be made until the establishment of an international standard
evaporation pan.
3. Observations, with either 1 or 2 above, of: (a) preCipitation; (b) water-surfacz
temperature in the tank or pan, and in the reservoir;(c) air temperature,humidity
and wind velocity.
3.2.5.2 Evapotranspiration
Evapotranspiration is one of the most important elements of the water balance. The
process of evaporation from land surfaces is complex and depends on the interaction
ofa relatively largenumber of variables,theimportantonesbeing climaticand vegetational
characteristics and subsurface water conditions.
The important methods used in estimating evapotranspiration are discussed in
section4.2.4.All methods have their drawbacks and few have proved to be very accurate,
although a most extensive estimation by evaporimeters has proved successful in the
U.S.S.R. Where the measurements are carried out correctly,the estimation procedure
is simple and gives a direct value of the evaporation. Observations should include
climatological variables such as air temperature, humidity,cloudiness,wind velocity,
precipitation and soil temperature and, in addition, soil-moisturemeasurements and
observations on the state and growth of the vegetation.
In some cases the energy-balancemethod can be used as a check and in such cases
complete actinometric observations are required as well.
Evaporation values estimated by evaporimeter are useful for the comparison of
evapotranspiration from various types of vegetation under various moisture conditions.
3.2.5.3 Evaporation from snow
In certain regions of the world (e.g.,in arid regions and in countries with a continental
climate), evaporation from snow may be relatively large,thus constituting a significant
part of the water balance.
In such regions it is necessary to obtain data on evaporation from the snow pack;
such data are useful not only for solvingwater-balanceproblems but also for investigating
spring snowmeltby means of the water-balancetechnique.
64
~
Evaporation measurements from snow may be carried out by means of weighing

evaporimeters with snow monoliths installed in such a way that the surface of the snow
cover and the surface of snow in the evaporimeter are on the same level. In addition,
observations of surface temperature of the snow cover, air temperature,water vapour
pressure and wind velocity at the snow surface should be made.
The estimation of evaporation from snow is of some interest for regions with a
comparatively thin snow cover and many sunny days in winter. In countries with
thick snow cover and considerable cloudiness in winter the evaporation value from
snow cover is relatively small and has no practical importance for water-balance
calculations.
I 3.2.6 Surface water

Surface-watermeasurements on representative and experimental basins are aimed at:
1. The determination of discharge and run-offvalues for given periods as an indication
of the water resources.
2. The study of the flow distribution and extremes for given periods.
3. The study of the effect of climatic and basin characteristics on surface-water
characteristics.
4. The study of the effect of natural or cultural changes on surface-waterquantities
and distribution.
5. The determination of water quality and its possible changes, either naturally or
under the influence of man.
Surface water includes overland flow, streamflow and water in lakes and reservoirs.
The most important aspect is streamflow and it is well to realize that,being an integrated
measurement, streamflow is possibly the only hydrological characteristic that can be
observed with a given level of accuracy. It is important that gauging stations on
representative and experimental basins are installed to the highest accuracy, with
sensitiveand stablemeasuring sections.All observations should be made on a continuous
basis with automatic recorders (see section 4.3).
In determining the duration of the observation of streamflow, consideration should
be given to the facts that streamflow has a high natural variability and that a sufficiently
long period of observation is required to establish normal values.
In the U.S.S.R.the determination of normal run-off with a given accuracy is
characterized by its coefficientof variation.Table 3.1 gives mean values of mean quadratic
error of normal run-off depending on the number of years of observation and the
coefficient of variation C, (error expressed in percentage from arithmetic means of
run-offvalues for II years).
This table shows that for the determination of normal run-off with a mean quadratic
error of about 5 per cent, a 5-year period is required if the coefficient of variation of
annual run-offequals 0.10;25 years if Cu=0.25and 100 years if Cv=0.50.
TABLE
3.1
Number of observation years, n

CIJ
5 10 20 40 60 80 100
0.10 4.5 3.2 2.2 1.6 1.3 1.1 1 .o

0.25 11.1 7.8 5.6 4.0 3.2 2.8 2.5
0.50 22.2 15.8 11.2 7.9 6.4 5.6 5.0
1 .o0 44.5 31.6 22.4 15.8 12.9 11.2 10.0
65
Estimates of the coefficient of variation for basins to be established may be made by

analogy to similar basins.
Dischargemeasurementson representativeand experimental basins are not necessarily
carried out only at the outlet, but also on a number of discharge sites located on the
main stream and its tributaries in order to obtain data on the influence of various
parts of the basin on the flow conditions.
Methods of measuring stream flow vary considerably according to geomorphological
and climatic conditions. Streams may be perennial or ephemeral;produce flashy or
flat hydrographs; have stable or unstable channels; have problems of frost, aquatic
vegetation, etc. Details of measurements are discussed in section 4.3.
Overland flow is studied on some experimental and representative basins for the
following reasons:(a) to establish an understanding of hydrological processes (erosion,
run-off,infiltration, etc.), taking into account effects of natural factors (e.g., slope,
soil frost,ice crust on soil under the snow cover, type of vegetative cover,etc.); (b) to
provide a quantitative evaluation of the effect of cultural changes (see section 2.4.1).
The general aim should be to measure run-offand erosion for given conditions of
soil, vegetation, rainfall, etc., and the variation of run-offwith time.
Overland flow may be studied on small natural basins with areas of up to several
hectares, when approximately homogeneous conditions occur. More detailed studies
should be made on run-offplots with areas of up to several hundred metres square;
such plots are isolated from the surrounding surface and can be selected to be almost
completely homogeneous.
In some countries a variant of the run-offplot called a water-balance plot is used.
While a run-offplot is isolated from the rest of the slope only by divides to a depth
of about 0.3 m,water-balance plots are isolated by means of waterproof walls to a
depth of at least 2-3 m.They provide not only overland flow measurements (as on
run-offplots) but also measurements of subsurface flow which occurs no deeper than
the walls of the divide.
Such plots are common in the U.S.S.R. and provide data for the,water balance of
a slope to estimate evapotranspiration, and for the determination of interrelations
between hydrological characteristics.
Obviously, for purposes of comparison, it is desirable to establish run-off and/or
water-balance plots within the limits of an experimental basin. Observations should
include precipitation and overland flow. In countries where snow occurs in winter,
snow survey and observations on the depth of soil frost and thaw are also essential.
Soil-moistureobservations and water-tablemeasurements should be made if applicable.
In the U.S.S.R., batteries of plots are frequently installed to determine the effect
of cultural changes on the hydrological regimen, by means of the following types of
experiment: methods of soil cultivation on slopes; artificial sprinkling or soil drying;
the increase of snowmelt by means of darkening the snow surface;the reduction of the
depth of soil frost by means of heating;the creation of an ice crust on the soil surface
under a snow cover [3].Such experiments are normally carried out after an initial
calibration period of about 2-3 years.
When run-off plots are specifically installed for an understanding of hydrological
processes, results of the studies are of interest for comparison with over-all basin
studies and it is absolutely necessary to standardize the following: measuring devices,
observational methods, definition of basin characteristics,and experimental design for
comparison of results.
3.2.7 Subsurface water

Subsurface water includes water in both the unsaturated and the saturated zones (see
section 1.5 [4,61.
66
Planning oj observations according to the research objectives
~ 3.2.7.1 Water in the unsaturated zone

Water in the unsaturated zone, generally referred to as soil moisture, is studied for its
importance in relation to water-balance,surface and subsurface phenomena,estimation
of evapotranspiration,etc.
Soil-moisturemeasurements are generally restricted to experimental basins, although
some observations on the smaller representative basins may be necessary. Requirements
and methods of soil-moisturesampling are discussed in section 4.4.1.1.
Soil moisture is generally studied in relation to soil physical characteristics.Opinions
vary as to which characteristics are most relevant in any one study. Reference to
sections 4.1, 6.1.1.3and 6.1.4.3will aid in deciding which characteristics should be
measured.In the U.S.S.R., bulk density, specific weight, mechanical composition,field
capacity and wilting point are determined at every soil-moisturesite. Below depths
of 3 m only bulk density and specific weight are determined. The determinations are
made once only, except where irrigation is applied. In such cases these soil physical
characteristics are determined before and after irrigation [6].
3.2.1.2 Water in the saturated zone

Water in the saturated zone is studied for a determination of ground-water storage
and its temporalvariationsfor water-balancecalculations;for evaluationof ground-water
resources; for interflow and base-flow studies; and for the development of models
for forecasting the ground-waterregimen. These objectives require the determination
of relations between surface and subsurfaceflow and variations in ground-waterstorage
for different times of the year.
Normally,the following long-termobservations are required:(a) stage and discharge
observations;(b) temperature and water quality of subsurfacewater by means of special
observation wells and by overflow artesian springs.
Observationsfora hydrologicalsurvey depend on thelithologicaland geomorphological
structureof basins and the character and distributionof the aquifers.Details are discussed
in section 4.4.2.Reference should be made also to sections 4.11,6.1.1.4 and 6.1.4.4.
A n example of a survey of this type in the U.S.S.R. is given below.
A minimum number of wells is installed in basins with a homogeneous lithology and
with one aquifer; a maximum number where aquifers are located in layers in basins
with an uneven topography. The total number of observation wells characterizes the
heads and ground-waterstorage for the entire basin. On individual sites,investigations
are made of the formation of surface run-offand subsurface flow and of ground-water
recharge and evaporation from soil. Such observations are made also on run-offplots
used to determine the effect of agricultural practices on the hydrological regimen.
When observation wells are sited the following principles are followed.
(a) Observation wells are located along a line, cutting the basin approximately in a
direction perpendicular to the direction of the principal flow of ground water or to the
direction of the main channel of the basin, draining subsurface water. As a minimum,
three observation wells are usually located along the slope in the upper part, in the
middle and at the lower part of the slope. In intermittent channels, an additional well
is bored for the determination of subsurface channel flow. In places where recharge
conditions of subsurface water change abruptly,e.g.,in polders,forest belts, etc., one
or two additional observation wells are installed.
(b) Where aquifers occur in layers,several wells are installed on the same site, each
well serving only one aquifer. Such a group of wells provides data for determining
relations between aquifers.
(c) The deeper the ground water occurs, the smaller is the number of observation
wells required for detailed characterization of the ground-water regimen.
67
Representafive and experimental basins
(d) Observation wells are sited in such a way that they are representative of the
hydrological region if possible. This will aid in the translation of results.
(e) A network of observation wells on representativebasins is required for the study
of the regional ground-water regimen. This network is not generally as dense as one
required for experimental basins (where research is directed towards an understanding
of the ground-water regimen and the relation of hydrogeological characteristics to
other hydrological variables).
(f) Observation wells are usually established over a period of several years;the first
to be installedare the most importantones,which reflecttheregimenof the major aquifers.
Observations of ground-water levels are carried out once in 3-5 days in all wells;
during high-intensityrainfalls and during snowmelt every day, and during dry periods
once in 5-10 days.
Some observation wells are equipped with automatic recorders. Where measurement
of the ground-waterenergy potential is required,piezometers are installed.Observations
of ground-watertemperature are carried out simultaneouslywith water-levelobservations,
but only on a selected network of wells, this network being selected to be typical of
the major aquifers.
Observations of ground-waterdischarge are carried out on streams fed by ground-
water springs by means of regular discharge measurements.
Water sampling for water quality is carried out on a selected network of observation
wells at the end of every season and about 3-5 times during occurrences of a high
ground-watertable at the beginning of the rise, in the middle, at the crest and during
the recession.
3.2.7.2.1 DETERMINATION OF PRINCIPAL CHARACTERISTICS OF AQUIFERS
For the calculation of ground-water storage and yield the following characteristics of
aquifers are determined :
1. The dimensions (geometry) of water-bearing geological formations, the thickness
of aquifers and hydrological formation constants such as: porosity,pore volume,
effective porosity, specific retention, and specific yield.
2. The followingaquifer constantsand the hydraulicpropertiesof aquifers:the coefficient
of permeability (hydraulic conductivity); the coefficientof transmissibility;the coeffi-
cientof storage;the piezometrichead;and the hydraulicgradient or piezometricslope.
The aquifer constants,by which hydraulic properties of the aquifer are characterized,
are determined by pumping tests. For evaluating the results of pumping tests in Europe
the equilibrium methods (Dupuit, Thiem) are used, while in the United States the
non-equilibrium ones (Theis, Jacob, etc.) are mostly used.
Ground-water storage is generally estimated by multiplying the thickness of the
aquifer by the specific yield. The hydraulic gradient is calculated by means of
ground-watercontour maps (hydroisohypsal maps) or by direct measurement of water
levels in wells. The thickness of aquifers is determined with the aid of hydrogeological
survey data and hydroisohypsal maps.If such data are not available,data from water-table
observations are used [4].
For details of an analysis of hydrogeological characteristics, see sections 6.1.1.4
and 6.1.4.4.
3.2.8 Infiltration
Infiltrationis one of the most important processes in hydrology and one of the least
satisfactory as regards accuracy of determination. It is recommended, nevertheless,
that on all experimental basins infiltration be included in the programme.
68
Plaiitiing of observations according to the research objectives
Infiltration measurements are made by laboratory methods such as those carried out
by the U.S.S.R. in large soil monoliths with undisturbed profiles at various rates of
moisture and frost, by infiltrometer in the field (see section 4.5) and by infiltration
analysis (see section 6.1.3.3).
The programme should be directed to obtain infiltration data or indices for basins.
This may require an approach by more than one method on any one basin.In addition,
theoretical studies of subsurface flow measurement (see section 6.1.4.3) will aid in
understanding the infiltration processes and will consequently aid the interpretation
of results.
3.2.9 Glaciers
Glaciers,as accumulators of abundant water resources in the solid phase, are important
for water-resources investigation and nations with glacial regions should include a
glacial basin in the representative basin network.
Observations should ultimately be directed towards determination of a water balance
and an energy balance, together with detailed photogrammetric surveys carried out
at regular intervals (e.g.,every five years).
In the past, glacier observations have frequently been restricted to only one or two
of the above objectives or only a part of one, and it has been found generally that
results have not been sufficiently informative.
Initial observations on glaciers should include items 1 to 6 below. Observations 7to 9
may be made later in the study.
1. Continuous flow measurement on the stream emerging from the glacier.
2. Regular measurement of the ablation of the glacier and surveys at the tongue of
the glacier.
3. Climatological observations at one or more points to indicate typical conditions
for the glacier (wind, temperature, humidity, solar radiation, etc.).
4. Precipitation measurements.
5. Snow surveys.
6. Photogrammetric surveys.
7. Glacier volume measurements by seismic methods.
8. Special observations for energy-balance calculations (albedo measurements, glacial
air circulation, etc.).
9. Temperaturesurvey of glaciers (ice temperature measurement at the depth of constant
temperature and deeper).
In some countries run-offplots are also installed in glacial basins for specific,detailed
studies.
Glacial studies are frequently made difficultby lack of access and require careful
consideration of the programme of observation required before they are classified as
representative basins.
Since representative basins with glaciers are likely to be few in number, no details
are given in the guide except for flow and climatological observationsand snow surveys
(see Chapter 4).
3.2.10 Erosion and sedimentation

Observations of erosion should be made in all basins. Where soil erosion processes
are very active and present great problems, observations must be intensive. Suspended
sediment discharge measurements are made on streams where concentrations are high,
but should also be taken wherever possible to supplement the research data.
69
Investigations of erosion, sediment discharge and siltation of reservoirs on

representative and experimental basins are aimed at: (a) determination of relationships
between erosion and sediment movement and physiographical features of the basin;
(b) determination of areas of critical erosion and high sedimentdischarge;(c) prediction
‘ of erosion changes as influenced by geomorphological changes or by man’s activity
on the basin; (d) prediction of silting rate of ponds and reservoirs. In basins where
erosion is severe, the programme should provide, where possible, for micro-levelling
of run-off plots and volumetric measurements of eroded materials, as well as for
sediment-dischargemeasurements. In such cases it is useful to make surveys and
measurements of rills and gullies. The study of siltation of ponds and reservoirs is
carried out mainly by volumetric surveys.
The following data should be obtained :
1. On run-off plots: (a) total amount of slope erosion; (b) mechanical composition
of the eroded material.
2. On stream channels: (a) annual suspended-sedimentdischarge;(b) data on bed-load
discharge; (c) mechanical composition of bed load and bed materials; (d) rate of
reservoir silting (if applicable).
For details of observational methods, see section 5.6.For analysis techniques see
section 6.1.5.
3.2.11 Quality of water

Information on the chemical, physical and bacteriological quality of water is of such
importance in water management that observations on the natural quality on many
representative basins and on the effect of cultural changes on the natural quality on
experimental basins are essential. In some representative basins, a natural change in
the quality of water may be important but may require very long-term observations.
Water-quality research on representative basins is aimed at determining the typical
water quality of streams in various hydrologicalregions,as related to the type of stream
flow and to climatic and basin characteristics.
Such research is also importantin relations between water quality and the hydrological
regimen. If observations of subsurface water quality are taken simultaneously,
relationships may be established between surface and subsurface water to determine
the role of subsurface water in stream recharge.
Observations on the chemical composition of water are of help in relating stream
flow to soil types and this may be of help in characterizing hydrological regions.
The study of the quality of surface water on experimental basins may be necessary
to solve specific problems. The objectives may, for instance, include an investigation
of the loss of fertilizers,a study of the procedure of soil formation,or particular research
to establish relations between surface and subsurface water.
The minimum programme of water-quality observations on representative basins
should include the determination of the content of the principal ions in natural water
(CI’,sod’,HC03’Na++K+,Mg”, Ca..), total dissolved solids and water temperature.
The programme of water quality on experimental basins depends on the objectives
of the research, and may be as for representative-basins or considerably larger. For
details, see section 4.10.
3.2.12 Ice phenomena in streams

Observations on ice phenomena on small streams are frequently useless.In some streams
particular problems occur. The stream may, for instance, freeze up completely to the
70
bottom and, owing to the continual increase of the thickness of the ice cover, some
water m a y infiltrate into the channel bed so that, if the stream is in the permafrost
zone, there is no flow at all.
Where large ice formations or naleds1 occur, a considerable amount of water m a y
I be temporarily stored in the channel and observations of formation,melt, ice thickness,
ice conditions, ice dams and ice jams may be required. Such observations should be
taken not only at the gauging station, but along the entire stream length.
3.2.13 Climatological data and energy balance

Climatological observations on representative and experimental basins are required
for a detailed analysis of variables affecting hydrological processes and water-balance
components.‘Theprogramme of climatological and energy-balancecomponents should
be extensive and should provide for the collection of such data as precipitation,
evaporation from the water surface, evapotranspiration, evaporation from snow,
snowmelt, soil frost and thaw, etc. The programme should allow for the collection
of all data affecting any one hydrological process, in order to establish relationships
between climatological data and hydrological processes.
Observations for energy-balance studies on representative and experimental basins
should be made according to definite objectives; such research deals with the energy
aspects of various hydrological processes, such as evapotranspiration, condensation,
freezing and thawing, etc.
In general, the programme for climatological observations should provide for :
(a) all the characteristicsof the climatologicalconditions which affect hydrological proces-
ses (run-off,evaporation, heat and moisture exchange in the unsaturated zone, etc.);
(b) calculation of values for energy and moisture fluxes, which affect hydrological
processes when no observations for the determination of the energy balance are made.
The programme for observations for the determination of the energy-balance should
include reliable data on aí1 components such as the energy exchange of the soil with
the atmosphere, soil energy exchange, energy yield with condensation and freezing,
energy spent in evapotranspiration, evaporation, ice and snowmelt, and soil thawing.
3.2.13.1 Climatological observations
Considering the above principles, the following climatological elements should be
observed as a minimum: precipitation; snow depth and density; temperature of the
air (maximal, minimal and at given intervals) ; air humidity (absolute and relative) ;
cloudiness ; soil temperature (minimal and maximal on soil surface), temperature of
soil frost and thaw, and if snow occurs, the temperature of the snow cover (at the
surface); wind direction and velocity.
The frequency of observations should be such as to obtain both a reliable daily mean
value of the variables and an idea (at least approximate) of the daily variations of these
variables.
Times of observation should, if possible, be similar to those observed on the standard
climatological network of the given country, so as to aid in the interpretation and
translation of the results.
I
3.2.13.2 Measurement of energy-balance components
Considering the discussion of section 3.2.13, the following items should be observed
I to determine the energy balance: downward short-wave and downward all-wave
l 1. Caused by freezing of water discharging through cracks in the ice cover on the surface,
or by discharge of ground water on to the surface with subsequent freezing.
71
radiation; upward short-wave radiation; net exchange of all-wave radiation; energy

flux into the soil; convective energy exchange with the atmosphere; and energy losses
by evaporation. The last two values are usually determined by means of combined
measurements of temperature gradients and air humidity at two or more elevations.
Reduction of the above items may be possible in some cases (see section 4.2.5.3).
Since observations for energy-balance studies are very difficult and require the use
of special instruments and qualified personnel, they should be planned only on selected
representative and experimental basins. For details of climatological measurements,
see section 4.2.
3.3 Planning observations for the study of the effect,

on the hydrological regimen, of a natural and/or
cultural change
Planning observations for experimental basins is complex and depends on the problem
to be studied, the physiographical conditions encountered and the resources available.
Section 2.4 discusses selection of experimental basins and section 6.3 analysis
techniques for the study of the effect, on the hydrological regimen, of a natural and/or
cultural change; these sections should give sufficient guidance for the establishment
of suitable programmes. Some examples are given below of the planning of experimental
basin studies in the U.S.S.R.,where variations of the hydrological regimen and water-
balance components under the influence of different factors are studied.
Experiments are run in order to estimate by direct measurement the values and
variations of all possible water-balance components and their relations under the
influence of the factors studied. In practice the condition is extremely difficult to fulfil,
owing to insufficient accuracy of measurement of most hydrological data. For large
basins it is extremely difficult to obtain reliable mean basin data for precipitation,
evaporation from land, soil-moisture content, ground-water fluctuations and deep
percolation. For this reason the programme of experimental work provides for
comparative treatment on two or more basins.
Comparative treatment may be used also on small basins; they are grouped to obtain
maximum variability of the factor studied, keeping other conditions approximately
equal. Factors studied are soil type (varying infiltration), vegetation, slope and aspect,
drainage density, erosion-channel depth, various methods of soil cultivation, irrigation,
drainage, etc.
In some cases the application of a cultural change is very difficult. This applies
particularly to phenomena which have a low frequency of occurrence. In such cases,
the experiment might be very important and artificial aids are used, e.g., infiltrometers
to study moisture loss by infiltration at different values of initial moisture content;
artificial drainage or wetting of soil cover to study the occurrence of overland flow on
slopes; the construction of an artificial ice crust on the soil surface under the snow cover
to study the effect of an ice crust on spring run-off: artificial soil frost and thaw at
various rates to study infiltration conditions in spring.
While the study of any phenomenon in an experimental or representative basin is
proceeding, the basin is also used, for economic reasons, for the study of other concrete
phenomena. Thoroughly equipped observation stations and trained staff enable
simultaneous investigations of a number of different phenomena.
Programmes of experimental investigation of some phenomena are given below.
The programmes are schematic and contain only necessary observations of certain
phenomena under various physiographical conditions; they do not deal with all the
variety of experimental work.
72
Plaririing of observations accorditig to the research objectives
3.3.1 The study of the influence of forest on the hydrological

regimen
I
The problem of the effect of forest on run-offhas been discussed for over 100 years
and still lacks a reliable solution.The problem is of especial interestforarid and semi-arid
regions in connexion with the role of forests in water and soil conservation.
Numerous investigations have proved that forest greatly affects the hydrological
regimen, water balance, run-off,infiltration, ground water and evapotranspiration.
Considerable changes of the forest percentage of river basins may cause changes of
hydrological characteristics and their interrelations.
In opposing physiographical zones, e.g.,in arid and humid ones, the influence of
forest may be quite different and even opposite.For this reason,research on the effect
of forest and forest cultivation on the hydrological regimen might be usefully included
in the programme of research.
The main purpose of the study of the hydrological role of forests in the U.S.S.R.
is to solve the following problems: (a) whether discharges (normal annual, maximum
and minimum) increase or decrease under the influence of forest, forest cutting and
afforestation;and whether the run-offdistribution varies over annual periods; (b) why
the run-offregimen changes under the influence offorest,forest cutting and afforestation
(or, in other words, which hydrological factors are affected by forests, forest cutting
and afforestationand in what way). Studies are by comparativetreatment and the water-
balance technique.Observations of flow and other water-balancecomponents are made
in two basins (wooded and woodless) simultaneously; if the two basins have similar
characteristics of relief, soil, vegetation, etc.,then the difference in flow characteristics
of the basins may be explained by the influence of forest.
Investigation of the hydrological role of forests by the comparative treatment method
may be made in two ways: (a) without running an experiment, i.e.,by observing the
two basins (wooded and woodless) in their natural state;(b) by running an experiment,
i.e., by cutting or planting forests in one of the two basins under study, leaving the
other in its natural state and organizing parallel observations in the basins before and
after the experiment.
The first variant gives lessconvincing results,but permits a relatively shortinvestigation
period in the U.S.S.R.(10-15 years). The second gives more convincing results, but
the observation period is longer (1 5-20 years, including basin calibration).
The first method consists of the following three stages: selection of basins; parallel
observations during 10-1 5 years;and analysis and interpretation of results. The second
experimental method requires the following three steps: selection of basins; calibration
of basins,i.e.,parallel observations of the basins in their natural state during 3-5 years
before carrying out the cultural change; and running the experiment (forest cutting
l or afforestation). The parallel observations during the experiment require a period of
10-15 years in the U.S.S.R.to let the soil acquire new properties.
Experiments by the first method are carried out in the U.S.S.R.not only on
experimental but also on small representative basins, especially where the problem is
to define only quantitativevariations between dischargevalues of forested and unforested
basins without a thorough analysis of the reasons causing the variations. In this case
the programme of observations is rather simple and includes only observations of
precipitation and flow. It should be noted that no statistical check is possible with this
method and any results should be interpreted with extreme caution.
The second method has been tried in the U.S.S.R. on experimental basins. Such
investigations are somewhat more fundamental in that they give certain reasons as to
why the hydrological regimen changes under the influence of forest. For this purpose
observations in the basins are carried out as given in the example in section 3.3.1.2.
Some comments on the various stages of the investigation are given below.
73
3.3.1.1 Selection of basins
Correct selection of the basins is important for the success of the experiment and for
the reliability of the results. Details are given in section 2.4.
3.3.1.2 Programme of observations
(a) M i n i m u m programme. For the problem of qualitative variations of flow

characteristics in forested and unforested basins, it is considered sufficient in the U.S.S.R.
to organize parallel observations in forested and unforested basins according to the
following minimum programme ; (i) daily measurements of amount of precipitation
on a network of gauges; (ii) streamflow measurements with automatic recording
equipment.
W h e n snow occurs in winter it is necessary to carry out snow surveys at the end of
the winter to estimate the maximum water equivalent of the snow pack before spring
snowmelt starts (see sections 3.2.3.1 and 4.2.2).
(b) Full programme. For the problem of determining why variations of flow
characteristics occur on forested and unforested basins, a full programme of observations
of all possible water-balancecomponents is carried out on plots and basins.Observations
in the U.S.S.R. include:
1. Measurements of rainfall intensity on basins and run-off plots by recording
raingauges.
2. Measurements of interception.
3. Streamflow measurements with automatic recording equipment.
4. Observations of evaporation by evaporation pans and lysimeters at evaporation
stations both in the forest and in the field on bare soil and on vegetated areas
(evapotranspiration) and by forest soil evaporimeters.
5. Measurements of soil moisture on basins and run-off plots.
6. Observations of ground-water level to determine ground-water resources.
7. Determination of soil infiltration capacity in basins and run-off plots.
8. Climatological observations on forested and unforested basins (air temperature,
humidity, wind velocity).
9. Snow surveys during the winter in basins and on run-off plots.
10. Observations of snowmelt in basins in spring.
11. Observations of soil frost and thaw in basins and run-off plots.
12. Measurements of evaporation from snow surface in basins.
Sometimes the full programme of observations includes observations for energy-balance
calculations which are essential to the study of evaporation and snowmelt.
The above programme is adapted for determining the effect on the hydrological
regimen of forest management techniques such as partial cutting of forests, forest
thinning, logging, etc.
References
1. KHARCHENKO, S. I. 1964. Zadachi i metodika issledovaniy i raschetov vodnogo balansa
oroshaemykh zemel [Tasks and methods of water balance research and computation for
irrigated lands]. Materialy mezhduvedomstvennogo soveschania po problerne izuchenia i
regulirovania isparenia s vodnoy povorkhnosti i pochvy (30 July-3 August 1963), p. 285-300.
Valdai, GGI.
2. ~ . 1966. Osnovnye polozhenia programmy kompiexnykh vodnobaiansovykh i
agrometeorologicheskikh nabludeniy, metodika rascheta vodnogo balansa i kharakteristik
74
vlagoobespechennosti selskokhozjaistvennykhpolei [Principlesof the programme of complex

water balance and agrometeorological observations, methods of computation of water
balance and characteristics to provide agricultural fields with water]. Materialy seminava
po raschetam vodnogo balansa reehnykh basseinov i organizatsii komplexnykh vodnobalanso-
vykh i agrometeorologicheskykh nabludeniy, p. 233-92. Valdai, GGI.
3. KUZIN, P. S. 1954. O b organizatsii i postanovke issledovaniy vliania agrotekhnicheskikh
meropriatiy na rechnoy stok [On the organization of investigations on the effect of agro-
technical arrangements on river run-off]. Meteorologia i gidrologia, no. 1 :23-5.
4. LEBEDEV, A.V. 1966. O zadachakh issledovaniy balansa podzenmykh vod na experimental-
nykh basseinakh gidrogeologicheskikh stantsiy [Onthe problems of water balance investi-
gations of subsurface water on experimental basins of geohydrological stations].Materialy
seminara PO raschotam vodnogo balansa rechnykh basseinov i organizatsii komplexnykh
vodnobalansovykh i agrometeorologicheskikh nabludeniy, p. 196201. Valdai, GGI.
5. OURYVAEV, V. A. 1953. Experimentahye gidrologicheskie isdedovania na Valdae [Experi-
mental hydrological investigations in Valdai]. Leningrad, Gidrometeoizdat.
6. RODE, A. A. 1965. Osnovy uchenia o pochvennoy vlage. Tom 1, Vodnye svoisfvapochv i
peredvizhenie pochvennoi vlagi [Principles of soil-moisture study. Vol. 1, Soil physical
properties and soil-moisture transfer]. Moscow, Academy of Sciences of the U.S.S.R.
7. SOKOLOVSKY,D.L. 1933. K proektu organizatsii seti stokovykh/balansovykh/stantsiy[On
the project of the organization of experimental-water balance stations network]. Zzvestia
GGZ,no. 59 p. 23-6.
8. VELIKANOV,
M. A.;LVOVICH,
M. I. 1932. Tipovaja programma dla stokovykh stantsiy
[Standard programme for experimental catchments]. Zzvestia GGI,no. 49, p. 10-18.
9. YUNUSOV, G.P.1965. K metodike rascheta vodnogo balansa v sviazi s khoziajstvennoy
dejatelnostju na vodosbore [On methods of computation of water balance in connexion
with economic activities in the catchment]. Trudy GGZ, 127 : 101-27.
10. ANON.1954. Rukovodstvo stokovya stantsiam [Guidefor experimental catchments]. Lenin-
grad, Gidrometeoizdat.
75
4 Methods of observation
and instrurnentation
4.1 General requirements

Instrumentation and observation of phenomena in representative and experimental
basins is extremely costly and it is very difficult to obtain accurate data on any one
characteristic. For this reason it is essential that as many methods as possible are
standardized;that instrumentsare of a similar nature,ifpossible (to simplify maintenance
problems); that observations are made by standard methods; and that instruments
are as used by national hydrological and meteorological services or as recommended
by W M O [249].
General requirementsfor such instruments are that they should be designed to ensure
accurate measurements, they should be reliable in operation,and they should be simple
and sturdy and be constructed in such a way that readings are easily made.
Instruments should be supplied with factory certificates and, from time to time,
tests should be made on them in special calibration laboratories or else their readings
should be comparedwith those of standard instruments(since allhydrologicalinstruments
eventually vary from their original calibration) [233,245, 2461.
Special attention should be paid to the method of installation,because the accuracy
of almost all instruments depends to a large extent on their proper installation [169].
In some sparsely populated regions the possibility of damage to instruments by
wild animals should be taken into consideration and spares should be available to
ensure continuous observations.Recording gauges should be used to make continuous
observations where possible. It should be noted that recording gauges, as a rule, do
not provide absolute readings and that regular and accurate readings of non-recording
gauges are therefore required in additionto obtain absolutevalues;for example,readings
of a staff gauge are associated with those of automatic water-level recorders,etc.
The officer in charge of the basin should continuously supervise the observational
quality bycregular inspection of all measuring devices and should check the skill of
the observers.
Observers should be urged to conform to the following major rules while making
observations:
1. In the field notebook only eye-witnessreports should be recorded,without substitution
of assumed or calculated values.
2. Observations should be made at the required time and in the planned order.
3. Gaps in records should be avoided and observers should be instructed that any
missing observation lessens the value of the whole observational series.
4. The instruments should be handled with care and protected from damage.
76
Methods OJ observation ancl Nistrtiaieiiiation
5. All the rules and directions given in the instruction should be followed.
Observers should be provided with a brief written instructionfor every type of observation
which is not included in the standard written procedures used by the hydrological
and meteorological services of the country.
Such an instruction should contain the following:detailed description of instruments
with diagrams;recommendationson routine care and action to be taken in the event of
breakage or malfunctioning; procedure for taking readings; times of observation;
guidance and direction on making non-routine observations (e.g., during floods,
rainstorms,etc.); procedure for checking charts and tapes;completion of field notebook;
procedure of primary processing of observational data.
The observer should not be limited to regular observations only. Visual observations
(without instruments) on atmospheric phenomena and weather changes which may
affect any element of the hydrological regimen should be taken when necessary. Such
visual observations are of great importance during rainstorms and floods; for example,
the water level during a flood should be marked at several points along the stream if
the gauging structures are submerged by flood waters.
Sampling techniques are discussed in section 4.1.1.These cannot always be followed
and simpler methods may be necessary in some circumstances. A description of the
minimum equipment required for research on representative and experimental basins
is given in section 4.12.A discussion of this equipment and any special equipment
that might be required is given below.
4.1.1 Sampling technigues

Hydrological methods necessarily involve sampling in space and time from infinite
populations. Uncertainty as to the space and time distribution of the various inflows
and outflows of the hydrological cycle has so far prevented the accumulation of sufficient
accurate data to establish firm physical laws governing this cycle.
Apart from streamflow, which is a priori an integrated observation for an entire
basin,hydrological observations are mostly point samples. Since observations at a point
are often required as well as averages for a basin, it is scientifically and economically
wise to establish,within any one basin, a number of so called ‘mastersites’where point
observations are made of a number of hydrological characteristics (e.g., precipitation,
soil moisture, infiltration,etc.). Such master sites are preferably installed on typical
soil-vegetationcomplexes (see section 2.2.3) and they need not be very numerous.
To obtain a statistically sound sample for average basin values, subsites can be operated
for shorter periods and the results obtained at such sites correlated with data obtained
at the master sites.
In representative and experimental basins, the particular need to sample basin
characteristics and to deal with the non-stationarityof basins due to natural evolution
(erosion,climatic changes,etc.), makes sound statistical sampling essential to improve
hydrological methods in general.
4.1.1.1 Sampling in time
Hydrological sampling in time is perforce non-random. For instance, daily mean

discharges observed over a period of a few years may be regarded as a non-random
sample from an infinite population of daily mean discharges. In representative basins
which are used for prediction purposes, samples which are non-random in time may,
nevertheless,be used with some confidence to generate non-historicsequences,provided
that time invariance or stationarity of the population is presumed.
Methods used are correlation [65, 1411 and stochastic treatment [li, 701. In some
cases,testing the representativenessof the sample by testing the data from an equivalent
time period of another hydrological variable for which a longer record is available
may increase the reliability of the data for use.
4.1.1.2 Sampling in space
Random sampling in space is frequently impossible in hydrological methods. The high

cost of hydrological installations, the inadequacy of instrumentation and the difficulty
of terrain make it necessary to resort to purposive or stratified sampling.
Random sampling has as a criterion that every individual has an equal chance of
being chosen and its greatest advantage is that the results obtained can be assessed
in terms of probability. Random samples may be conveniently drawn by the use of
playing cards or random sampling numbers [237].
For sampling on an areal basis it is convenient to draw a grid, the detail of which
must be related to the effort and expenditure involved in carrying out the sampling.
For instance,a much more detailed grid could be used for sampling by ring infiltrometer
than by using a sprinkling one.
Purposive sampling involves the selection of sampling points by some purposive
principle. A raingauge network, in which a gauge is installed in a basin at a location
that is considered representative, is frequently established by this method. Purposive
sampling is often useful if the sample is small, since the observer attempts to choose
a value near the mean and the sample may be representative. As the sample becomes
larger, however, the random sample becomes more and more representative of the
population, whereas the purposive sample, because of bias, does not. Moreover, while
the purposive sample might give more idea of the mean,it possibly does not give a good
idea of the variance, because values are chosen near the mean.
Stratified sampling is a combination of random and purposive sampling. The
population is divided into strata and a random sample is taken from each. For instance,
the stratification of a basin into geomorphological, vegetational,pedological or climatic
classes, or a combination of some or all of those with subsequent random sampling
in each class, is possibly the most suitable method for many of the observations
required in representative and experimental basins.
The sampling problem can be simplified still further by considering that the object
of observations in representative and experimental basins is generally twofold,namely
to obtain a mean value for the basin, and to obtain a value for the site for subsequent
correlation with other site characteristics.
For the first objective, accuracy need not be of a high order, and sometimes it is
not even necessary to know the actual values-the change of the variable with time
is all that is required. The latter objective, if it is the only one, does not require a
sampling technique.
When both objectives are pursued simultaneously, a number of sampling points
can be nominated as master sites where more accurate observations are taken. If,
for example,soil moisture is measured by a neutron scatterer,longer counting intervals
at the master sites will provide accurate data for a study of the moisture behaviour
at a single point, while at the same time the data can be used as part of the network
of moisture-sampling points with shorter counting intervals for the estimation of the
mean basin values.
Studies have been made of the sampling of soil moisture and of discharge by utilizing
a large number of samples in an analysis of variance [39,971. The following four errors
were considered: (a) instrument error; (b) timing error (random error associated with
the time required for a single observation); (c) position error (positioning the neutron
probe at the exact depth on successive measurements;positioning of current meter at
78
Methods of observation and instrumentation
correct points in the vertical to obtain the velocity in the vertical); (d) location error
(reflecting difference in soil moisture content from place to place; number of stations
in a gauging cross-section).
In both cases it was found that the location error was the most significantand relations
were developed between the expected standard error and the number of sampling points.
Investigations of a similar nature are invaluable for studying the sampling problem
for most hydrological observations.For further detail on sampling methods reference
I
should be made to standard textbooks on statistics [47,2051.
4.2 Climate
Hydrological research cannot exist without a knowledge of the macro-climate of a
basin. The precipitation pattern, for instance, is associated with lower atmospheric
conditions. For this reason,the climatic demands for research on representative and
experimental basins must extend beyond the basin divide,although detailed observations
should be confined to the basin to measure aspects of the meso- and micro-climates.
Most countries possess a standard meteorological network of climate-recording
stations with varying lengths and intensities of observation. When representative and
experimental basins are selected, the national meteorological network should be
consulted since if,in particular,researchinto a climatic change is projected,such research
will be assisted if a long-establishedstation is within the basin divide. If a new station
is set up, this should become part of the national network. Note that the WMO [234]
has issued a guide which deals specifically with climatic observations.
4.2.1 Precipitation
4.2.1.1 Generai
The general purpose of measuring precipitation in representative and experimental

basins is to obtain point values for correlation with hydrologicaland basin characteristics
at a site and to obtain mean basin precipitation.
The latter requirement is more difficult to meet than the former because precipitation
may have a high natural variability over the basin; both requirements are influenced
by effects of the measuring device and the immediate environment of this device. The
effects of natural variability of precipitation over an area may be reduced by using,
where possible, sound sampling practices; the effects of the measuring device and the
immediate environment may be minimized by followingrecommendedexposure practices
and by selecting the most appropriate instrument.
It must be realized that completely accurate mean basin precipitation values are
unobtainable and,in general,only an index of the mean basin precipitation is obtained.
This is not a serious problem provided there is some idea as to the order and variability
of the error involved (see section 6.2), since it is likely that errors are damped to a
certain extent in most relations involving precipitation [210].
4.2.1.2 Networks
For representative and experimental basins it is essential that attempts are made to
estimate the mean basin rainfall with a high degree of accuracy [7].This will involve
dense networks on any basin where the natural climatic variability is high. General
1 sampling principles are given in section 4.1.1 but, for access or economic reasons,such
methods are not always possible. In such cases sampling might be complemented by
79
elevation-precipitationcorrelations to arrive at indices of mean basin rainfall. Estimation

procedures have been developed, using data from fifteen widely separated American
basins, for average spacings between gauges required to obtain a correlation r of 0.90
between gauges for storm rainfall events [94].Figure 4.1 shows an estimating diagram
which may be used for conditions similar to those under which it was developed [56].
In the U.S.S.R.investigations on the accuracy of the measurement of mean areal
precipitation in relation to raingauge network density, areal extent, time interval of
precipitation totals and character of its falling (frontal,convective or orographic) have
been carried out. These investigationswere based on a great number of data, obtained
on specially organized areas (the density of the precipitation gauge network was up to
750 gauges on an area of about 600 km2).
The investigations were carried out in two regions in the northwestern part of the
U.S.S.R. European territory (region of excessive moisture), where frontal precipitation
is observed, and in the south-western part (region of insufficient moisture), where
convectiveprecipitation prevails. Some of the quantitative results of these investigations
are given in Table 4.1, which shows that the error of the average precipitation over
an area with a given network density decreases abruptly when the time interval for
averaging the precipitation over the area increases. In the region of prevailing frontal
precipitation, for example, with a network density of 1 gauge per 50 km2and an area
of 500 kmz,the average precipitation error equals 19 per cent for 24 hours, 8 per cent
for ten days, 4 per cent for a month and 2 per cent for a season.
For the region of prevailing cyclonic precipitation the corresponding figures are more
than twice as great (46, 17, 10 and 4 per cent). With the same network density and
averaging period, the error in mean precipitation over the basin decreases if the area
under investigation increases.
The data from this table may be used for planning precipitation gauge networks
on representative and experimental basins [781.
The minimum standards for networks in representative and experimental basins
are given in section 4.12.Note that greater densities are required in areas of steep
topography and/or areas under maritime influences.
FIG.4.1. Diagram for

estimating the distance
between gauges as a
functionof the two-year-
24-hour and two-year-l-
hour rainfalls [93].
T w o yeor 24-liour rainfall (mm)
80
-iWinbrn
8
O v - N T
3
O 3 6 0 6 0
2: v --*
O
O
3-I-2
V
s w-
mow 3P-m
inIA
3 4 m
min w N
o 3
3
d v
N. -
IA .n
3 N i n m b -3minw ---Nu
8
4
V
m
V
3
vv
O -I-ITtPO
N --wmm
2: V vv
O
3
-wr-m
3 N
-imo=
v --
mo
N o riNin
2: W
V
36
3
V
m IA N
N
-000
-in08
-IA
th
d k
73 73
3
81
Representative and experimental ‘basins
4.2.1.2.1 N E T W O R K REAPPRAISAL
Frequent review of the basin network is essential to permit determination of: (a) the
adequacy and possible refinement of the network and observation programme; (b) the
time required to achieve experimental significance; (c) the possible need for restating
objectives; (d) the presence of faulty gauges and poor instrument locations and exposures.
For some projects,reappraisal may permit a reduction in the network. Reappraisal
methods may include the study of the areal representativeness of point rainfall data.
Aspects treated in the past have been related either to the representativeness of total
storm gauge catch data (or daily or monthly rainfalls), with respect to the mean basin
rainfall (using correlation methods), or to methods of total storm catch by the isohyetal
method (see section 5.5.1).
Representativeness has often been defined by a comparison between absolute point
values at a gauge and mean basin values,or between the statisticalfrequency distribution
of time series values obtained at individual gauges and the time series represented
by the average of data from an entire network. In some of these studies [24,461, an
attempt has been made to establish the general shape of total storm gauge catch patterns
in special cases. In other studies the pattern of instantaneous gauge level precipitation
intensities during storms (see section 5.3.I) has been determined [5].
4.2.1.3 Precipitation gauges

Precipitation includes the water equivalents of snow, hail and other solid forms as
well as rain, dew and fog. Rain is normally the most common form of precipitation
and is certainly the most easily measured.
Precipitation gauges are either recording or non-recording.
4.2.1.3.1 RECORDING RAINGAUGES
Three types of recording raingauge are in general use, the weighing type,the float type
and the tipping bucket. The first type is the most satisfactory in cases where snow is
a significant proportion of the precipitation.Details of recording raingauges are given
in the literature [44,2311.
The movement of the float, bucket or weighing mechanism can be employed to
produce a digital record of rainfall.This can be done at the gauge,either mechanically
with the record being punched on paper tape,or electronically on magnetic tape where
a battery-operated tape deck is available. The alternative is to convert the movement
of the mechanism into an electric signal and transmit it by wire or radio to a distant
receiver, where a data logger collects records from a number of instruments.
4.2.1.3.2 N O N - R E C O R D I N G PRECIPITATION G A U G E S
Non-recording gauges are constructed either to be read daily (manual gauges) or at

longer regular or irregular intervals (storage gauges). The manual gauge is the basis
of most national networks. Its design and method of installation have usually become
established as a result of compromise [231].The gauge material, size of orifice and
capacity have often been determined by cost,while height has been decided as a result
of a compromisebetween the adverseeffectsof in-splashand wind.A further complication
is that,because of snow,most gauges are placed at a higher level than is best suited for
rainfall measurement [76,771.
Storage gauges are designed to accommodate expected rainfalls between visits.
82
l Methods of observation arid instrumentation
Accuracy of rainfall readings can be maintained by increasing the size of the orifice
~
relative to the diameter of the container. Inexpensive storage gauges are most suited
I to the dense network requirements of steep terrain (Fig. 4.2).
FIG.4.2. Storage
raingauge at 1,300m,
Kaweka Range, New
Zealand.
4.2.1.3.3 E R R O R S IN P R E C I P I T A T I O N G A U G E S
Wind affects the accuracy of precipitation measurement. Some errors may also occur
because of evaporation or wetting of the surface of the gauge [113].
Experimental evidence suggests that losses by wetting equal approximately 0.2 mm
per measurement, and losses by evaporation, depending on the design of collector,
may equal up to 6 per cent of the total fallen precipitation during the warm season
[19,58, 61, 133, 1641.
In some mountainous countries, in representative and experimental basin studies
the installation of one or more vecto-pluviometers[86]may be essential to study storm
direction and to correct vertical gauge catches.
Observational errors must also be taken into account, but these are generally of
a random nature. Regular maintenance can eradicate some sources of error, while
addition of antifreeze in cold countries and oil in hot ones can eliminate others. Splash
into the gauge can be minimized by employing gauge surrounds like short turf and
shingle, but if a specially constructed non-splash surface is employed, the gauge can
be installed with its rim at ground level. This gives the far greater benefit of lessening
wind effects on the gauge, thus increasing the catch compared with that of elevated
gauges [135,193, 2101. A well-tried variant of the gauge at ground level is one set in
a circular turf wall [lo31 built around the gauge at a diameter of 3 m. An alternative
way of modifying the surroundings of the gauge is to fit suitably-shapedwind shields
around the instrument.When properly designed,these enable much more representative
results to be obtained than do unshielded gauges fully exposed to the wind. An ideal
83
shield should: (a) ensure a parallel flow of air over the orifice of the gauge;(b) avoid
any local acceleration of the wind above the orifice; (c) reduce as far as possible the
speed of the wind striking the sides of the receiver; (d) not give rise to any splashing
towards the oriñce of the receiver.
Where shielding leads to complete icing-overin blizzard conditions [75], it should
be discontinued.
4.2.1.3.4 INSTALLATION M E T H O D S
Installation methods for precipitation gauges are described in the literature [231].In
representative and experimental basin research it is essential to use, at least nationally,
uniform standards and it is recommended that orifices are sited with a uniform height
above the air-vegetation interface. This might require elaborate structures in forested
basins [175].
It is recommended that where gauges are installed in any way that is different from
the usual practice, they should be tested alongside the national gauge to provide a
basis for comparison.
4.2.1.3.5 SELECTIbN OF PRECIPITATION G A U G E S
The selection of precipitation gauges is mostly dictated by such considerations as

research requirements,capacity required,precipitationtype (solid or liquid), frost and/or
evaporation problems, and observational frequency.
With manual and storage gauges, the sensitivity is related to the relative diameters
of orifice and container and to whether a measuring glass or a dipstick is used. The
sensitivity ofrecording precipitationgauges depends to a certainextent on type (weighing,
tipping bucket, float (syphon or storage)) and manufacture. The choice must depend
on the purpose, although for experimental basins and for the smaller representative
basins a chart scale of 20 mm per hour will be found essential.For run-offplots a more
detailed chart is required while for the larger representative basins less detail will be
satisfactory, depending on what types of analysis are proposed.
Table 4.2 gives a guide to the choice of precipitation gauges.
4.2.1.3.6 M E A S U R E M E N T O F D E W A N D FOG
For details of dew and fog measurement, see section 4.2.3.3 and the literature [231].
4.2.2 Snow cover

Measurements of snow are made to evaluate the water equivalent and its importance in
thewater balance of a basin.Determinationof the maximum water equivalent of the snow
coverat the beginning of the snowmeltperiod is of special interestfor studyingstreamflow
characteristics influenced by snowmelt. Also of interest are the variations of the water
equivalent with time and the distributionof snow over differentkinds oflandscape(forest,
arable land,ravines,etc.) or over the basin as a whole. The water equivalent of snow is
determined by means of snow surveys,either areal or lansdcape [136,138].
The areal snow-surveymethod is fairly accurate and is the most desirable. The method
is based on selecting,within the basin,snow courses which represent all types of relief.
For this reason the method is restricted to comparatively small basins with a relatively
homogeneousvegetation.Snow courses on such basins are located in the shape of straight
or broken lines,stretching from one side of the basin to another in a direction approxi-
mately perpendicular to the direction of the main stream,thus crossing various slopes.
84
-dea
O 0 0
888
-dea
O 0 0
888
c
o
a
b
85
This will ensure an approximate agreement between the lengths of the courses crossing
various slopes,and the total length of all courses with areas with similar slope conditions
and the area of the entire basin.
The number of courses depends on the configurationof the basin:two or three courses
are considered to be quite sufficient for wide basins of rounded form,and four or five for
narrow basins. Courses are located approximately parallel to each other in the direction
of the main stream, with equal distances between them. The mean value of the water
equivalent of the snow pack is calculated arithmetically averaging the areal snow-survey
data.
The second method-landscape snow survey-is a more economical one and is used
on larger basins, which have various types of vegetative cover and more complex topo-
graphical conditions.Non-homogeneityof vegetation and relief affects in particular the
distributionof snow storage over a basin.
For the use of this method the entire basin is grouped into principal landscape elements
(grasslands,forests,scrubland,valleys, slopes,river channels,etc.) and the total area of
each element is determined.
Surveys are made on several courses 0.5-2km long,each course being located within
a defined landscape. Such a complexity of courses may characterize the mean conditions
of snow accumulationon basins up to 100-200k m 2(mountainous regions excluded). O n
larger basins several landscape complexes should be selected and each course should
characterizethe mean conditionsof the snow accumulationin differentparts of the basin.
Therefore, the number of courses on a basin of up to 100-200 km2or on part of a
bigger basin is determined by the number of principal types of landscape. The mean
water equivalent of the snow pack is estimated as the mean weighted value of the snow
survey data on individuallandscape complexes,taking into accounttheir respectiveareas.
For a large basin,the mean water equivalent of the snow pack is estimated as a mean
weighted value on the basis of mean snow-storagedata of individual parts of the basin
(taking their areas into account).
The snow course in an open (woodless) area may be from 1 to 2 km long,depending
on the roughness of the landscape.The snow course should cross not only even areas,
but slopes of different exposure,valleys,hills and other areas characteristic of the given
landscape.The snow course in forests is normally 0.5-1.0km long and should cross all
parts which are characteristic of the given forest area. Snow surveys in ravines should
be made along 2-5 cross-sectionswith a total length of not less than 0.2 k m and not
more than 0.5 km.
With areal and landscape snow surveys,the measurement of snow depth is made at
equal distances every 10-20 m and measurements of snow density at each fifth point of
the snow-depthmeasurement.
Closer distances between the measurement points are not desirable,since snow depths
in adjacent points have a comparatively high correlation coefficient and snow density in
adjacent points varies less than the depth [250].
Itshould be noted that,in steepbasins at high altitudes,snow depths are highly irregular
over the basin and snow accumulations in sheltered areas are not significant unless these
areas contribute substantiallyto run-off.Experience indicates that most snow areas are
frequently bare before run-offstarts and only the gully accumulations contribute sub-
stantially to run-off.
In very steep basins, snow accumulates in gullies only and the objective of obtaining
snow depth and density information for the whole basin for water-balancestudies is, for
practical reasons,unobtainable.In such cases the only practical approach is to relate an
area of accumulationto run-off[159].
In many basins,because of topography and lack of access,only basin indices can be
obtained.To provide basin indices.snow surveys should be made at index courses estab-
lished in each of the hydrologically distinct accumulation areas which contribute
86
Methods of observation and instriimentation
substantially to run-off.Surveys should be carried out at least once during the accumu-
lation period,preferably on the date which corresponds to the average date of maximum
accumulation. More frequent surveys will be required in climatic areas where significant
winter melt occurs.
Areas prone to avalanches,steep icy slopes,corniced areas and other dangerous sites
should obviously be avoided.Particular care should be exercised in the selection of snow
coursesto avoid unintentional samplingbias. Carefulobservations,includingphotograph-
ic records of snow conditionsin the basin during at least one snow season should precede
the selection of snow courses.For dates of snow surveys see section 3.2.3.1.
Because of the time and effort generally required to make snow measurements and
associated observations,and with the object of lessening the inherent risk of sudden
adverse weather changes at high altitudes in certain latitudes, reasonable access is of
primary importance in the selection of index courses.
Precipitation observations are usually made simultaneously with snow surveys on a
basin. In regions with an unstable snow cover,where estimation of the maximum water
equivalent of the snow cover is often difficult because of thaws,the data on the amount
by precipitation gauges may be used for correction of snow-surveydata.
Noting that the correlation coefficient of the snow-coverdepth in adjacent measuring
points within a distance of about 20 m is small,the errors in estimation of the mean
snow depth or water equivalent of snow may be evaluated approximately by means of
the following equation:
Px = -CT
ZVN
CV
x 100 = -
VN
x 100
where:
Px = relativevalue of the error expressed in percentage(67 per cent level);
u = standard deviation;
5 = mean value of snow depth or water equivalentof snow cover;
N = number of measuring points;
Cv = coefficient of variation.
The number ofmeasuring points N,required for the estimation of the mean water equiv-
alent of the snow cover over the basin with a given accuracy Px,is calculated by the ratio
In the U.S.S.R. a movable snow stake and a weighing snow sampler (see Fig. 4.3)are
used for snow surveys [250].The Mount Rose sampler (Fig. 4.4)which is used in the
U.S.A. is described in the literature [231].
The complete set ofa weighing snow sampler consists of: (a) a metal cylinder,60 c m
high with a cross-sectionalarea of 50 cm2,open on one side;on the outside the cylinder
is graduated to determine the depth of the snow core;(b) a shovel for cutting the core;
(c) a steelyard,each graduation of the scale corresponding to a weight of 5 g.
The volume of the core equals 50 h (where h = the core height), and the weight of the
core equals 5 m (where m = the reading on the balance scale). These are used to estimate
the snow density by the ratio:
87
ia)
Legend
1. Snow sampling tube.
2. Cutter.
3. Cover.
4. Movable ring with a hook
(5) for suspension.
6. Balance beam.
7. Balance prisms.
8. Suspension ring.
9. Prism with a hook. 2
10. Movable balance weight.
11. Balance weight stopping
screw.
12. Slot for pointer. 13
13. Scoop.
FIG.4.3. Movable
snow stake (a) and
weighing snow sampler
(b), U.S.S.R.
G F
Legend
A. Snow-sampling tube. D. Driving wrench. G. Screw couplings.
B. Tubular spring balance. E. Spanner wrenches. H. Scale.
C. Cradle. F. Cutter.
FIG.4.4. The Mount Rose sampler, U.S.A.
88
Methods of’observation and instrumentation
Thus to obtain the snow density the number of graduations of the weight scale should
be divided by the reading of the cylinder scale multiplied by ten.
The instrumental error in measuring snow density by the weighing snow sampler is
obtained by summing the error in measuring the snow depth (which may be estimated
as 0.5 cm) and the error in taking readings by the balance scale (which may be assumed
to be half a graduation or approximately 2.5g). The greatest relative errors occur when
the depth and density of the snow cover are small,and the smallest errors when the snow
depth is equal to the height of the snow sampler and the snow density is above normal.
When the snow depth is 50 cm and the density 0.40,for example,the relative error in
measuring the density will not be more than f 2 per cent.
In mountainous basins snow surveys are very difficultand sometimes dangerous for
a snow surveyor. In areas with difficult access and considerable snow depth where long-
distance methods of measuring are possible one of the simplest methods used is the in-
stallation of permanent snow scales with readings taken by means of optical devices.
Measurement of the water equivalent of snow in separate points can be made by an
instrument based on the principle of counting gamma rays absorbed by snow.Cobalt-60
is used as a source of gamma radiation and a long-distanceGeiger-Müllercounter may
serve as a detector (see section 4.11.3).
A device to measure the water equivalent of the snow cover,based on registration of
the diminution of the natural radioactive radiation of the earth under the influence of
a snow cover,is being installed in the U.S.S.R.The impulses of a medium radioactive
background of the earth on a selected course before and after snowfall are measured by
terrestrial or aerial survey [235].
4.2.2.1 Snowmelt
Data characterizing the intensity of snowmelt may be required in studying processes of
spring floods caused by snowmelt in order to develop methods of calculation and fore-
casting of spring run-off.
The simplest methods are based on the relation between snowmelt and positive air
temperatures. One of these methods consists of determining the dependence between a
reduction of the water equivalent of the snow cover,obtained by snow surveys,and the
sum total of the daily mean positive air temperatures.
Where researchinto snowmelt processes is to be done all factorsaffecting the snowmelt
should be taken into account.
In the U.S.S.R.a special method, which is based on an inventory of all the energy-
balance components of the snow pack has been developed. All the components of the
radiation balance and of the turbulent heat and moisture exchange of the snow pack with
the atmosphere are considered.With these investigations the water-balance method is
usually applied using data from snow surveys,measurements of evaporation from the
snow cover, precipitation observations and calorimetric estimation of the water-liquid
phase in the snow pack.
Simplified versions of snowmelt-intensitycalculationshave also been used,Such meth-
ods are generally based on the use of climatological data such as wind velocity,total and
lower cloudiness,vapour pressure and air temperature.
4.2.3 Interception of precipitation by vegetation

4.2.3.1 Interception of rain
4.2.3.1.1 FOREST VEGETATION
4.2.3.1.1.1 General
A large part of the rain falling upon forests is evaporated from the aerial parts of trees
[87] and from the litter beneath them. Hydrologists disagree on the importance of
89
interception as a water loss in excess of normal transpiration.Some reason that energy

dissipated in evaporatingintercepted water cannot be used for transpiration134,147,1511.
Thus, they believe that evaporation of intercepted water results in lower transpiration
rates. O n the other hand, some [79,1991 suggest that sources of energy for transpiration
and interception loss may not be the same. In fact,they have shown that interception
loss during the dormant season greatly exceeds maximum potential evapotranspiration
calculated by empirical formulas.Although the net effect of interception on the water
balance of a forested catchment is not clearly understood,substantial interception dif-
ferences between forest stands and specieshave been demonstrated. Because interception
loss may affect total water yield from forest lands,hydrologists need estimates of these
losses.
The objective of this sectionis to discuss the whole interceptionprocess,to recommend
proven sampling techniques and instrumentation,and to present factors which will be
usefulin planning interception studies. For terminology,see section 1.5.
Gross rainfall,throughfall,and stem flow can be measured in the field.Litter intercep-
tion presents a special sampling problem because it cannot be measured directly.After
individual components of the interception process have been determined, total intercep-
tion loss and net rainfall can be calculated by solving algebraic equations. Thus, total
interception loss is calculated as
Is = Pg-(Tf+ Sf) + Litter Is (4)
and net rainfall is
Pn = P,-Is (5)
where :
Is = interception loss;
Pu = gross precipitation;
Tf = throughfall;
Sf = stem flow;
Litter I, = litter interceptionloss;
P, = net precipitation (see section 1.5).
4.2.3.1.1.2 Vuriubles
A complete treatmentof allvariables affecting interceptionby forestvegetation is beyond
the scope of this guide. The objective here is to review the variables known to affect
interception and to discuss their relative importance in the interception process. For
convenience,interception variables are classed as climatic factors or as stand character-
istics.
Total rainfall and storm frequency are the two most important climatic variables.
Analysing gross rainfall and interception factors by covariance techniques usually re-
moves 95 per cent or more of the variation between individual measurements. Other cli-
matic variables (i.e.,rainfall intensity, wind speed and air temperature) are sometimes
statisticallyrelated to interceptionfactors,but their net effect is small [91].
With the exception of gross rainfall,stand characteristics such as type of stand (coni-
ferous or deciduous)are usually more importantsources of variation between stands than
climatic variables. Interceptionloss in deciduous species is greater in the growing than
in the dormant season,but the seasonal effect is less important in conifers [125,1581.
Canopy density (an expression of species,stand age,stocking,etc.) is directly related to
interceptionloss [125,158].Stem flow varies with bark roughness and branching charac-
teristics,averaging 10 per cent of gross rainfall in beech [144],but stem flow is insignificant
in mature Douglas fir 11961.
90
Methods of observation and instrunientation
4.2.3.1.1.3 Methods and instrumentation
4.2.3.1.1.3.1 Gross rainfall. Conventionalrainfall-samplingtechniques are discussed in

section4.2.1.Theseare applicableto the samplingof gross rainfallfor interceptionstudies.
In some cases a sealed-surfacetechnique is used instead of precipitationgauges.
4.2.3.1.1.3.2 Throirghfall. Water filtering through the forest canopy varies from point
to point by 100 per cent or more [98].This large spatial variation has encouraged many
investigatorsto use samplers with large receiving areas in attempts to reduce throughfall
variation. Small cylindrical gauges are favoured because they are easily obtained and
positioned in the field.Furthermore,because gross rainfall is sampled with round gauges,
the use of round throughfall gauges avoids the problem of comparing data from different
types.
Point throughfall amounts are directly related to distance from tree trunks [98,1881,
but the correlation in closed forests is too weak to justify a stratified sampling scheme.
If throughfall is measured in very open stands,stratification may be desirable.Sampling
plots can be divided into homogeneous zones [126].
4.2.3.1.1.3.3 Stem flow. Stem flowis usually less than 10 per cent of gross rainfall and
it is often omitted in interception studies. This omission leads to overestimatesof total
interception loss and stem flow must thereforebe measured in any complete interception
study.
Stem flow is sampled by sealing collars of copper or tin sheetingto trees to divert down-
flowing water into containers for measurement.The collar should project about 2.5 c m
from the tree bole. Wider collars are sometimes used on rough-barkedspecies,but these
probably catch some throughfall in addition to stem flow [144].
The best method is to locate smallplots randomly within the study area and to measure
stem flow from all trees within these plots. Measured water volume is readily expressed
in conventional depth units by dividing by plot area. Plots should be at least 1.5 times
the crown area of the largest plot tree or 20 m2for very small trees [92].Measured in this
manner, the coefficient of stem-flowvariation is a tenth to a twentieth of that of single-
tree samples,or only slightly greater than that for throughfall.
4.2.3.1.1.3.4 Litter interception. During storms,some water is retained by the litter

layer,where it is unavailableto plants but is subject to evaporation.Although throughfall
and stem flow have been extensivelymeasured,litter interceptionhas rarely been studied.
It may, however,account for as much as 10 per cent of the total interception loss,and
no study should be considered complete unless litter interception is included.
The recommended method is based on the knowledge that litter interceptionis a func-
tion of the amount of litter on the forest floor,its moisture-holdingcapacity,and the local
climate.Litter samples are collected from the undisturbed forest floor to establish wetting
and drying curves and litter weight. It is then possible to estimate moisture fluctuations
through time and to express litter interception losses in conventional depth units. Sam-
pling under natural conditions is preferred because it ensures natural moisture drainage
and natural drying [gol.
4.2.3.1.1.3.5 Sampling intensity. Below is given an equation which is useful for deter-
mining the number of gross rainfall,throughfall and stem-flowsamplesneeded to achieve
a predetermined sampleaccuracy [208].In the equation
n is the number of samplesneeded,t is the tabulated value for the desired confidencelevel
91
and thedegreesof freedom,02 is the population variance,and dis the maximum permissible
difference between sample and population mean. Experience of other investigatorsis the
best source of estimates on population variance ; study objectives dictate sample accuracy.
Figure 4.5 shows the average coefficient of variation data of interception factors for
some stand and climatic conditionsin the eastern parts of the U.S.A. [91]. It shows that:
(a) each factor in the interception process is a different sampling population; (b) the
coefficientof variation is inverselyrelated to the gauge catch until the latter reaches about
10mm,but is independentfor larger gauge catches;(c) throughfall variation is related to
timber type and,for deciduous trees,to the season of the year;(d) stem flow measured on
40 m2plots is slightly more variable than throughfall.Data in the figure can be used as
a first approximation of sampling needs for other studies.The standard deviation (u) for
a given storm size can be computed from Figure 4.5 by multiplying the coefficient of
variation by the average gross rainfall,throughfall, or stem-flowgauge catch. With an
estimate of oz from Figure 4.5 and the proper t and d values necessary to meet study
objectives,the number of samples needed can easily be calculated by solving equation (6).
A= STEM FLûWPINE,ANNUAL
E= THROUGHFALL.HARDWOODS,SUMMER
C: THROUGHFALL.PINE,ANNUAL
D=THROUGHFALL,HARDWOODS.WINTER
E: GROSS RAINFALL. ANNUAL
FIG.4.5. The relation

of coefficientof variation
\
to selected interception-
loss factors. o I O 20 30 40 50 60 70 eo 90 100
AVERAGE GAUGE CATCH, MILLIMETRES
4.2.3.1.2 HERBACEOUS VEGETATION
4.2.3.1.2.1 General
Whereas forest and shrub communities have been widely studied,little research has been
done into the measurement of interception loss in herbaceous communities [33]. Most
work dates from 1940,although pasture grasses were considered earlier [loo]. The lack
of activity is attributablemainly to the difficulty of obtaining reliable measurements.The
mechanical and spatial restrictions imposed by the reduced forms of herbs make the use
of macro-sampling methods impossible. Micro-variations require refined and indirect
measuring techniques.
4.2.3.1.2.2 Sampling methods

The variety of methods availablemakes it impossibleto standardizesamplingprocedures.
Because of difficulties in separatingstem flow,throughfall and drip in the laboratory and
the field,these parameters are normally measured as one.
The techniques considered are divided into two sections: gross interception loss and
net interception loss.Early efforts generally measured gross interception loss while later
techniques have attempted both. The more primitive methods are worth consideration
92
as they may provide quick results for checking against more complicated but less
mobile techniques.
A note of warning must be sounded here. The investigatormust not rely too heavily
on variance information from Figure 4.5or from other studies. Such data provide a first
approximation of sampling needs, but calculations of actual variance must be made as
soon as enough data are available. For example, studies have been reported [115]in
which throughfall variation was higher than expected and more gauges were required to
meet study objectives.Too often variance is calculated after the study is closed and when
it is too late to correct gauging intensity.
4.2.3.1.2.3 Gross interception loss. Some of the earliest attempts,where the height
and density of the sward permitted,used standard precipitationgauges,placed 3 c m above
ground level,amongst the growing vegetation.The throughfall thus collected was compar-
ed with precipitation catches on open ground [162,1811. Others used test tubes flush
with the ground to avoid disturbing the vegetation. [74]Results skewed by overland flow
are a possibility in this method.
Others placed troughs in rows amongst the vegetation [lo, 431. Troughs integrate
throughfall to a greater extent than gauges, although they do tend to overestimate the
amount. They record the less dense outer portions of the canopy and do not measure
stem flow.
Stem flow forms a significant proportion of the gross precipitation reaching the soil.
Failure to record this value gives overestimation of the interception loss. Stem collars
cannot be used and therefore the surface may be waxed under a sward in an effort to
catch throughfall and stem flow [89].This method has been refined by others [50],but
even so its success in dense swards is doubtful.
The best method appears to be the cutting of specimensat ground level.They are then
arranged on a screen,irrigated with a known quantity ofprecipitationand the throughfall
collected 143, 134, 1981.
These methods measure stem flow but introduce artificial arrangements in vegetative
form. A test using a modified Northfork infiltrometer (Fig. 4.6)has been tried in dense
rye dairy pasture. This method measures leaf-interceptionloss but does not accurately
record the large quantity of moisture retained at the base of the stem. This moisture
could be considered as surface detention. The method is satisfactory for low-growing
and clustered plants.
4.2.3.1.2.4 Net interception loss. In some laboratory studies [34]samples grown in

nutrient solutionswere used. Paired samples on scales were irrigated to prove conclusiv-
ely that leaf-surfaceevaporationreduces transpiration loss and increasesnet interception
loss by the presence of dead matter. The system separated throughfall and stem flow,
avoided the problem of taking soil-moisturemeasurements [207]and utilized plants grow-
ing in their natural position.
Other workers confirmed these results by extending this work and using floating lysi-
meters in a grassed field [151].
4.2.3.2 Interception of snow

The basic method for determining throughfall is a comparison of precipitation gauges
and/orsnow boards installed beneath the control gauges in adjoining clearings.If rainfall
is absent,snow-packmeasurements can be used. It may be necessary to measure stem
flow,particularly during the thaw period and conventionalcollars are adequate.
Automation of sampling plots is complicated by the immobility of snow and by low
temperatures.Sampling,phytomorphological and climatic requirements are the same as
for rainfall.
93
Represenlutive and experimental busins
FIG.4.6. Herbaceous
interception measure-
ment using a modified
Northfork infiltrometer,
Ministry of Works,
N e w Zealand.
4.2.3.3 Interception of dew and fog

Although different in origin,dew and fog present similar interception problems. Of
chief concern is the accurate measurement of the gross precipitation,as the canopy con-
centrates the moisture which can then be collected by conventionalthroughfall and stem-
iiow techniques.
Various methods are available for measuring dew. Weighing was the first to be used,
but has generally proved cumbersome [116].The Kessler-Feuss recorder is a refined
weighing type. Thermoelectric,absorption and optical methods [6,60, 971, lack repeat-
ability,accuracy and sensitivity respectively.Depositionon artificial surfaces is not easily
related to deposition on plants. Other researchers have developed an integrating device
which involves the extension of a plastic element by wetting [i 191.This method has ob-
vious advantages over previous ones.
Several of the devices available for dew measurement could be used to record fog. Of
these,the most common equipment consistsof a series of wire or fibre strands in a frame
sited normally to the direction of fog drift. The size of particle caught depends on the
diameter of the strand.Artificial resins,because of their uniform hydrophobic properties,
are rapidly supersedingglass fibre and spider’sweb for collecting fine droplets [121].
94
4.2.3.4 Future research
In evaluating net interception loss the whole plant-water-soilecosystem must be investi-

gated. Three points requiring study are listed [79]:
1. Transpiratory regulationby plants under the stress of limited water availability.
2. The probability that energy availablefor evapotranspirationfrom wet leaves is greater
than that from dry leaves.
3. The possibilities that under severe winter conditions transpiration is limited more by
the availability of water than of energy and that interception increases the quantity
of water favourably exposed to the energy supply.
It appears that,despite boundary effects,the lysimeter (see section 4.2.4.2.2)can make a
significant contribution to the study of interception processes. Detailed studies of the
water system in plants must be combined with phytomorphological studies (see section
4.6)to aid the extension of data for large-scalewater-balancework. This will mean new
measuring techniques (e.g., isotopic, conductivity, photographical) to strengthen the
relationship between evaporation and transpiration. With sufficient knowledge it should
be possible to obtain indirectly net interception loss from such parameters as gross inter-
ception loss and soil moisture.
4.2.4 Evaporation
Water is transferred to the atmosphere from open water,bare soil and wet vegetation by
evaporation and through leaf stomata by transpiration. For these two processes acting
together the term evapotranspiration is used by some workers,while others employ the
term evaporation.
The concept of potential evapotranspiration(or potential transpiration)has been devel-
oped to overcome the problems that arise when the effects of soil-moisturestresses on
the water loss of a plant have to be taken into account,as it is still not completely clear
what happens to transpiration rates as soil moisture becomes limited. Potential evapo-
transpiration implies that the vegetation has a copious water supply with climate as the
only control of transpiration.
As for precipitation,evaporation and evapotranspiration are defined in terms of depth
of water. There is a great variety in the means employed for measuring these factors,
with certain methods requiring very complex instrumentation and otherslittle more than
a bucket [114].The cost of the installation does not,however,indicate the precision of
the estimate.
There are several methods of estimating evaporation:
1. Water-balance methods : these employ the basic water-balance equation or modifi-
cations of it, the amount of evaporation or evapotranspirationbeing deduced by dif-
ference (see section 5.3.4.1).
2.Energy-balancemethods:for details, see section 4.2.5.4and [231].
3. Aerodynamic approach [231].
4. Empirical formulas;these have usually a limited regional application [173,213,2171.
5. Lysimeters and evaporimeters (see section 5.3.5.5).
6. Evaporation pans (see section 4.2.4.1).
4.2.4.1 Evaporation pans

Pans form an arbitrary but consistent method of estimating evaporation from a basin.
Many differenttypes areinuse;some being square and otherscircular;some being mount-
ed above ground and others sunk in it so that the level of the water is approximately
that of the ground. The 20 m2 and the GGI-3000(with an area of 0.3 m2) evaporation
95
I
FIG.4.7. Plan of 20 m2
pan.
Legend
i. Cylindrical reservoir 2 m
deep.
2. Reservou with measuring
glass (3).
4. Tube.
5. Stilling chamber.
6. Connecting tube.
7. Benchmark tube.
8. Water level indicator.
9. Volumetric burette.
FIG.4.8. Volumetric
burette.
FIG.4.9. Measuring
tubes.
96
Methods of observation and Nutrumeritation
pans are used on the basic network of the U.S.S.R. for measuring evaporation from water
surfaces.In the U.S.A. and some other countries the Class A pan (with an area of 1.14m2)
is used [30].
The U.S.S.R. 20 m2 evaporation pan is of a cylindrical design with a flat bottom,made
of welded sheet steel 4.4mm thick.The pan is 5.04m in diameter and has a depth of 2 m .
(Fig.4.7).A measuring reservoir is used for adding water to the evaporation pan up to
the normal level.The water level in the pan is measured in a stilling chamber which has
a benchmark tube with a hole for a volumetric burette.The latter is of a cylindrical de-
sign with a height of 60 mm and a cross-sectionof 20 cm2 (Fig. 4.8). For measuring the
water volume taken by the burette from the evaporation pan,measuring tubes in the form
of graduated glass cylinders are used (Fig.4.9).
The GGI-3000evaporation pan (Fig.4.10)consists of a tank, a raingauge (Fig. 4.11)
a volumetricburette (Fig.4.8)and measuring tubes (Fig.4.9).The pan is a cylinder 60 c m
high with a cone-shapedbottom,made of metal and protected from corrosion or specially
painted.The pan is 618 mm in diameter. A metal benchmark tube is placed in the centre
of the pan. During observations a volumetric burette is placed on the tube. A special
needle fixed on the tube indicatesthe level at which the water in the pan should be kept.
The raingauge has an orifice with an area of 0.3m2.
FIG.4.10. GGI-3000
evaporation pan.
1 FIG.4.11. General view of GGI-3000pan and a raingauge.
97
The American Class A pan is of cylindrical design.The pan is made of galvanized iron
or some other metal protected from corrosion.The diameter of the pan is 120.6c m and
its height is 25.4 c m (Fig. 4.12). T w o white lines at heights of 5 and 7.5c m below the rim
of the pan are painted on its inner side to facilitate maintenance of the water level. A
. stilling chamber with a water-levelmeasuring device is enclosed in the pan.
The 20m2evaporation tank and the GGI-3000pan are sunk into the ground or mount-
ed on anchored floating platforms on lakes or reservoirs.The United States Class A pan
is mounted above the ground on a framework made of 5 x 10 cmz timbers.
Measurements of precipitation, water-surface temperature and other climatological
factors such as wind velocity,temperature and humidity of the air are made simulta-
neously with evaporation observations.Evaporation and precipitation observations are
made once or twice a day,precipitation being measured in the morning and evening at
hours nearest to the fixed time of climatological observations.
4.2.4.2 Evaporimeters and lysimeters

Evaporimeters are tanks filled with soil which are used mainly in the measurement of
evaporation from bare soil or of evapotranspiration from various types of vegetation.
Lysimeters are sunken tanks (weighable or not weighable) filled with soil (which is
selected with special attention being given to the monolithic character of the soil prism)
and placed on characteristic locations.Lysimeters are used for more detailed studies of
the water balance and subsurface-watermovement.
4.2.4.2.1 EVAPORIMETERS
4.2.4.2.1.1 General
For measurement of evaporation from the soil (particularlyfrom agricultural fields) stan-
dard gravimetric or 500 cm2 weighing-type evaporimeters are used in some countries.
They are 100 or 50 c m deep,depending on the depth of the soil layer which has the most
intensivewater exchange between the soil and the atmosphere.These evaporimeters may
be used to measure evaporation from grassland,arable land,and waste land and to meas-
ure evapotranspiration from crops and evaporation from soil under the vegetative cover.
Gravimetric or weighing-typeevaporimeters with an area of 0.3m2and 1 .Om2are used
to measure evapotranspirationfrom crops.
Soil evaporimeters of the weighing type are considered by some to be reliablein meas-
uring evaporation values for a five- or ten-day period with an error not greater than
10-15per cent.
98
Hydraulic soil evaporimeters with a surface area of 0.2m2and a depth of 1.5 m are
used to study daily variations of evapotranspiration.The big hydraulic evaporimeter
(BGI)with an area of 5 m2and a depth of 2 m is considered by Soviet scientists to be the
most reliable reference instrument for the study of daily variations of evapotranspiration
for any period.
Measurement of evaporation from swamps and marsh-riddenareas is made by means
of evaporimeters for swamps [lo,35, 44, 68, 183, 247, 2481.
4.2.4.2.1.2 Soil evaporimeter GGI-500-100

This evaporimeter consists ofinner and outer cylinders and a collector.Inner and outer
cylindersare made of sheet steel 2 mm and 1 mm thick respectively;the collectoris made
of galvanized iron, 0.8mm thick.
~
A soil monolith is placed in the inner cylinder,which has an inner diameter of 252.3
mm and a height of 1,000mm.Necks for hooks to raise and transfer the evaporimeter
are welded to the upper rim of the cylinder,which is also provided with a shelter to
protect the clearancebetween the walls of inner and outer cylinders.
The inner cylinder has a removable bottom with a hole forthe water that leaks through
the monolith. The inner cylinder is inserted in the outer one which has a diameter and
height of respectively 30 mm and 35 mm more than those of the inner cylinder.The bot-
tom of the outer cylinder is solid and waterproof.
The collector is of cylindrical shape with an inner diameter of 253 mm,height 30 mm
and a funnel-shapedtop.The funnel has two holes;the first,in the centre,is for collecting
water leakingthrough the hole in the bottom of the evaporimeter;the second,in the upper
part of the funnel near the rim,is for pouring the water into the measuring glass.
4.2.4.2.1.3 Soil evaporimeter GGI-500-50

This evaporimeter has a design similar to the evaporimeter GGI-500-100and differs in
height only.The height of its inner cylinder is 500 mm and that of the outer one 535 mm
(Fig. 4.13).
4.2.4.2.1.4 The small-type hydraulic soil evaporimeter

This evaporimeter has an inner cylinder, outer case cylinder,annular float,tank and
measuring devices (Fig. 4.14).The monolith is placed in the inner cylinder,which has an
inside diameter of 505 mm and a height of 1,500mm.This cylinder has a removable bot-
tom with holes for water leaking through the monolith and is inserted into the outer case
cylinder so that its outer ring rests on the rim of the case cylinder.The outside diameter
of the case cylinder is 571 mm and its height 1,305mm.The collector for water leaking
through the holes in the inner cylinder is mounted on the bottom of the case cylinder
which has on its outer surface a ring,which rests on the upper surface of the float.
A n annular float is used to keep the evaporimeter with the monolith afloat.The inside
and outside diameters of the float are 601 mm and 1,620mm respectively.
The floating part of the evaporimeter is inserted into the tank, sunk into the ground
and filled with water. Three micrometers controlling the vertical position of the floating
system and one for determining the water level in the tank are the measuring devices of
the evaporimeter.
4.2.4.2.1.5 The big hydraulic evaporimeter (BGI)
This evaporimeterhas an evaporation surface of 5 m2and a depth of 2 m.It is a standard

reference in the U.S.S.R.(Fig. 4.15)and consists of three principal parts: a cylindrical
99
Legend
1. Inner cylinder. 4. Base plate catch. 7. Outer cylinder 10. Spring.
2. Base plate. 5. Latch. 8. Hook.
3. Water collector. 6.Handles. 9. Bracket.
FIG.4.13. GGI-500-50soil evaporimeter, U.S.S.R.
Legend
1. Tunnel shaft for weights.
2. Weights.
3. Electric micrometer screw
of the water-level gauge.
4. Float of the water-level
gauge.
5. Cup with mercury (on the
float).
6. Float weight.
7. Rubber rings.
8, 10. Reservoir.
9. Evaporimeter casing.
FIG.4.14. Small-type hydraulic soil evaporimeter.
100
Methods oj. observation and instrumentatton
Fisa
Legend
1. Ventilation tube. 10. Emergency outflow pipe. 16. Water-collectingpipe.
2. Reinforced concrete. 11. Outflow pipe from the 17. Supporting beams for the
3. Roof-supportbeam. annular reservoir. console.
4. Beam support. 12. Weights. 18. Monolith.
5. Annular reinforced con- 13. Indicator. 19. Axle.
crete reservoir. 14. Reservoirs for collection of 20. Stilling chamber.
6. Annular float. surface and subsurface 21. Recorder.
7. Float necks. flow. 22. Flume.
8. Outflow pipe. 15. Socket of the electro-
9. Inflow pipe. thermometers.
FIG.4.15. Vertical cross-sectionof the big hydraulic evaporimeter installation.
evaporimeterproper,containing the soil monolith; an annular outer water tank made of

reinforced concrete and containing an annular float to balance the evaporimeter; and
mechanisms for measuring and registering changes in depth position of the evaporimeter
with respect to fixed benchmarks and water level in the annular water tank. Regularly
spaced around the float are twelve metal calibrated necks which jut out above the water
surface in the tank.Suitable radial cantilever beams are fixed to the supporting ring of
the evaporimeter to hold it suspended in the air.These cantilever beams rest on the twelve
' necks of the float.Three cylindersare placed in the water tank at regularintervalsto house
small floats which operate recording units for registering the level of immersion of the
floating system [169,1821.
4.2.4.2.1.6 Forest hydraulic evaporimeter (large type)
This type is used in the U.S.S.R.and some other countries for the study of total soil
evaporation values, total transpiration of various kinds of tree and daily variations in
each.
The foresthydraulicevaporimeter consistsof an hydraulicbalance and an evaporimeter
weighed on this balance; the pan has a soil monolith with a tree growing in it. The area
of the pan is 3 m2 and its depth is 1.5 m.A complete set includesfive evaporation pans,
each pan being installed on a special car which can be transferred along rails by means
of an electric winch.
101
The major elements of the hydraulic balance are a platform supported on four floats
by a system of beams;four reservoirs connected with one another;four floats,placed in
the reservoirsto support the evaporimeter;and the platform.
The hydraulic evaporimeter is supplied with a special device for measuring and record-
ing the weight of the monolith [183).
4.2.4.1.2.1 Soil-weighing evaporinieter
This instrument has an area of 0.3m2 and consists of an evaporimeter,an outer cylinder
and a water-collectingwell. The evaporimeter (60 c m high) is made of sheet steel 3 mm
thick,with a movable perforated bottom, bolted in three points. In the upper part, the
evaporimeter has a shelter 7 c m wide which protects the clearance between the side of
the evaporimeter and the outer cylinder (Fig. 4.16).
FIG.4.16. Soil.
weighing evapoIimeter
with a surface iirea of
0.3 m2 installed in a
maize field.
The outer cylinder is made of sheet steel 3 mm thick and with the bottom inclined to
the observation well. Under the shelter the evaporimeter has a rectangular opening for
water flowing from the monolith surface after heavy precipitation.A trough is welded
under the opening to direct flowing water into the outlet tube.
The water leakingthrough the monolith is collected on the bottom of the outer cylinder
and flowsinto the special tube.Both tubes protrude out of the cylinder border and convey
the water falling from the monolith surface plus any leaking water into a special well,
situated 4 m from the evaporimeter.T w o collectors are placed in the well to accumulate
this water.
102
4.2.4.2.1.8 Soil-weighing evaporinieter (large)
This evaporimeter has an area of 1.0m2and a depth of 80 cm. The design is similar to
that of the 0.3m2evaporimeter,differing only in dimensions.
4.2.4.2.1.9 Evaporimeter for swanps, GGI-B-1000
This equipment consists of two concentriccylinders and a movable shelter to protect the
clearance between the cylinders from precipitation.
The inner cylinder,with a diameter of 356.5 mm and a height of 520 or 720 mm has
a solid bottom,above which is placed a wire bottom.A glass measuring tube is fixed on
the outer side of the cylinder and has a cover with a hole for the air. In the lower part,
the measuring tube is connected by means of a branch pipe with the space between the
wire and solid bottoms and with a valve for discharge of any excess water. The outer
cylinder has a lug where the measuring tube of the evaporimeter is inserted.
4.2.4.2.1.10 Lysimeter-compensatingevaporimeter
This lysimeter,used by the Valdai Scientific Research Hydrological Laboratory of the
U.S.S.R. comprises a set of four evaporimeters with a cross-sectionof 0.3 m2and a depth
of 2.0 m each;they are placed on concrete cylinders sunk into the ground (Fig. 4.17).
In the underground chamber joining the outer cylinders there are four compensating
vessels fixed on supports. Each vessel may be displaced vertically and readings of its
position are taken on a scale graduated on the vertical support. Each vessel is linked to
the evaporimeter by a rubber hose which is connected with the centre of the evaporimeter
bottom by means of a tube through the side of the chamber. These evaporimeters are
weighed on a standard platform balance [35].
FIG.4.17. General view of lysimeter-compensating evaporimeter installation.
4.2.4.2.1.11 Soil raingauge
A soil raingauge consists of a bucket and an outer cylinder for installation of the bucket.
The cylindricalraingauge bucket has an orifice of 500cm2and its height is 40 cm.
103
The outer cylinder has a height of 28 c m and a diameter of 35 cm.Three spring supports
for the raingauge installationare welded to the inner side ofthe cylinder bottom.
4.2.4.2.1.12 Methods of observation of evaporation from the soil surface

Observations made on soil evaporation sites include those of evapotranspiration;evapo-
ration from soil covered with vegetation ; and precipitationand water infiltration through
the soil monoíith in evaporation pans.These are made together with other observations
such as soil moisture (when monoliths are changed); observations on the growth and
state of plants;and observations on wind velocity, temperature and humidity of the air.
Changingthe soilmonoliths intimeis oneof the major requirementsforusing the evapo-
rimeters correctly. In the U.S.S.R. soil monoliths in GGI-500-50evaporimeters are
changed at least once a month in the zone of excess and sufficient moisture,and twice a
month in the zone of insufficient moisture.
With GGI-500-100 evaporimeters small hydraulic evaporimeters, weighing evapori-
meters (with areas of 0.3 m2and 1 .Om2, lysimeters and evaporimeters for swamps,soil
monoliths are changed once a year.
Evaporimeters and lysimeters are normally weighed once in five days, or five to ten
days for those with areas of 0.3and 1.0m2.
Hydraulic evaporimeters should be read morning, afternoon and evening, at hours
nearest to the fixed time of climatologicalobservations.
4.2.4.2.1.13 Evaporation from snow

The GGI-500-6evaporimeter is widely used in the U.S.S.R. to study evaporation from
snow. A complete set of equipment consists of the evaporimeter and the outer cylinder,
a balance with a capacity of 5 kg and an accuracy of 1 g,various smallweights and a shed
for the balance and thermometers.The GGI-500-6evaporimeter (Fig.4.18)is made of
duraluminium and consists of three parts: a cylinder with an area of 500 cm2and a height
of 6 cm,a removable bottom,and a cover.Each evaporimeter has an outer cylinder with
Legend
I. Cylinder.
2. Bottom plate.
3. Cover. 1
4. Shovel.
5. Outer cylinder.
FIG.4.18. GGI-500-6
snow evaporimeter
104
I Methods of observation and instvrrrnentation
a solid waterproof bottom, and a shovel for digging up and cutting a snow monolith.
During observations the evaporimeter is placed in the outer cylinder.
Observations on evaporation from snow and on the temperature of the snow-cover
surfaceare made morning and evening.This applies only when the depth of the snow cover
on the plot is 6c m or more.
In dry weather at temperatures below O" C and in the absence of snow,monoliths are
changed once every five days. In addition they are changed when snow drift or precipita-
tion has occurred during the period between the given and previous time of observation
or when the level of the compacted snow in the evaporimeter is lower than the upper rim
of the cylinder and structure of the snowfall has been disturbed [248].
4.2.4.2.2 LYSIMETERS
1 4.2.4.2.2.1 Purpose
1 The purpose of the lysimeter in representative and experimental basin research is to
establish a number of subsurfacewater characteristicswhich are measures of the capacity
of the soilto storewater,to convey it up or down inthe unsaturated zone and to determine
the variations in moisture conditions from the potential level (see section 1.5) to the leve-
where soil-moisturetension restricts the availability of soil moisture for evapotranspiral
tion.The amount of water drained off or the amount of evaporated moisture as determin-
ed by the lysimeter normally serves only as an indication of the order of magnitude
of these quantities under field conditions and has no value where more accurate
indications are needed.
The three major uses of lysimeters are:
1. Determination of the variation in soil-moisturein a given soil, and velocity of infil-
trating water and recharge due to precipitation (this enables the correlation of plant
growth and moisture content in successive soil horizons).
2. Prediction of the rate of evapotranspiration for a given soil-moisturecondition and
potential evapotranspiration and potential of the moisture uptake of the plant roots
in successive zones of the soil profile.
3. Assessment of the capacity of moisture storage of the soil and evapotranspiration
(this enables correlationof precipitationwith storage data which might lead to a better
understanding of subsurface flow).
Lysimeters for more detailed scientific investigations require a good deal of auxiliary
instrumentation.Only a few lysimetersshould be used and they must be of a high accuracy
and,often, of a large size. A n undisturbed soil is only seldom needed. These lysimeters
are generally constructed for permanency. When research of short duration and with
' restricted instrumentationis done to obtain resultsfor practical application,a larger num-
I
ber of lysimeters with different soil profiles and ground-watertables is used. These will
generally require undisturbed soil,be of small size,be used for a short time only and will
1 have to be cheap.
The use of a large number of small lysimeters has often to be regarded as a transitional
solutionwhere an investigationin the field without use of a lysimeteris too time-consuming
or is,for other reasons,less desirable.It should be borne in mind that methods are availa-
ble to determine in the field without a lysimeter as much as or even more than a lysimeter
can give.
4.2.4.2.2.2 Principle of construction
The lysimeter for scientific purposes will often be built to give an accuracy of 0.1 mm of
water. The lysimeter for practical problems can be considered sufficiently accurate with
an error of up to 5 mm of water.Because a lysimetercontainer of volume 1 m3and height
1 m weighs some two tons,the accuracies of 0.1 and 5 mm respectivelymean accuracies
I 105
of 1/20,000and 1/400of the total weight,or 1/2,500and 1/50of the weight of the soil-
moisture content.These accuracies ensure a good balance for the scientificlysimeter and
a good moisture determination for the practical lysimeter.Different methods of weighing
have been tried out for the accuratelysimeter,includingroman balances,platform scales,
the determination of the volume of replaced water of floating containersor of the hydrau-
lic pressure in bolsters (which was pioneered by the U.S.S.R.). The last-mentionedmethod
is recently drawing more and more attention,partly bacause of the low cost of the weigh-
ing instrument [72,205, 2061.
The instrumentationof practical lysimetersis less generally accepted.Gypsum or nylon
blocks often give erratic results.The use of ceramic plugs or other moisture-absorbing
substances such as filter paper (which are brought into close contact with the moist soil
and are then taken out,weighed and dried after attaining moisture equilibrium) appears
to be promising but is as yet insufficiently tested. The neutron probe and gamma-ray
moisture-determination apparatus use costly instruments and still suffer difficulties of
calibration.These nuclear devices, however, get more and more preference. Where low
costs are important,the filter-paperplug as a technique of determination merits closer
attention (see section 4.4.1.1.4).
Different types of lysimeter are used in various countries. A description of lysimeters
used in the Netherlands is given in sections 4.2.4.2.2.3and 4.2.4.2.2.4.A description of
a lysimeterused in the U.S.S.R. is given in section 4.2.4.2.2.5.
--i25 cm------,
F
Q Q
Legend
1 r
a Container,inner dimensions 120 x 140 x 180 cm. e Access tubes to soil, also suitable for drainage or
b Iron frames on either side of pressure cells. supply of ground water.
c Pressure cells. f Calibrated reservoir for drained water.
d Posts for support and high enough to give access g Suspended soil-filled roof of passageway between
under container. containers.
h Self-recordingevaporation pan.
FIG.4.19. Weighing lysimeter: details of container.
106
Methods of observation mid imtriitt~eiitution
4.2.4.2.2.3 Construction of the weighable lysimeter
Lysimeter containers of a fair size,say 1.40 x 1.40 x 1.70 m are arranged at both sides
of a passageway in a pit of, say,4 x 3 x 3 m,deep enough below the bottom of the
containers to allow the inspection of the measuring implements for leakage and wide
enough to allow measurements in the passageway.Each container rests on three pressure
cells,placed on masonry supports (see Fig. 4.19).A pressure cell is a case of some 40 c m
diameter and 4 c m height,closed at the top with a membrane and filled with oil.O n the
membrane a tapered plate, serving as a piston, conveys the weight of the container to
the oil.By bringing the oil under sufficient pressure,the membrane lifts the piston until it
is just free from the tapered rim of the cell. This is shown by a light.In this position the
membrane is flat,so that no vertical forces are exerted by it. The container,weighing 6
to 8 tons,is lifted by the pressure of 1.75 to 2.15atmospheresof the oil in the three pressure
cells.
The pressure in each set of three pressure cells is measured in a manometer stand.Each
manometer consists of a leg of 2 c m diameter and 1.80 m long,concentrically placed in
a closed wide leg of 10c m diameter and 10c m height.A rise in pressure of 104 c m mercury
Il k
m
T= J
Mercury Legend
a Three pressure cells carrying one container. g Tapered ring.
b Manometers. h Tapered plate.
c Reading tube with scale. i Lid.
d Body of pressure cell. j Tubing to pressure chamber
e Rubber ring. k Container with soil.
Membrane.
FIG.4.20. Weighing lysimeter, showing pressure cells and manometer stand.
107
Representafìve and experimental basins
results in a drop of the meniscus in the wide leg of 4 c m and a rise in the narrow leg of
100 cm. The desired accuracy of 0.1 mm water requires an accuracy of reading of each
manometer of 0.04m m . This can be done with screw gauges and delivery of data in print
on punch tape.Simpler and cheaper,however,are a hydraulicmagnifier and adding device
(see Fig.4.20).
The three manometer legs, belonging to the same container,are connected to a thin
glass tube in a slanting position,Oil is poured on the mercury. A change in height of
1 c m in the three mercury columns will produce a change in height of the meniscus of
3 x 1 x 22/0.42= 75 c m in the slanting glass tube which is assumed to have an inner
diameter of 4 m m . The magnification is therefore 75. If the moisture content increases
by 0.1mm then three readings of 0.04 mm increase are substituted by one reading of
3 c m increase of the length of the oil column.The size of the thin glass tube has to be
selected in such a way that the oil meniscus does not deform. Changing the size of
the manometer tube enables choice of desired magnification.
The hydraulic transmission is temperature-sensitive.It is therefore necessary to install
the manometer and magnifier in a thermostat which keeps the surrounding air at a con-
stant temperature. Because it needs a rather complicated construction to extend the
thermostatto the pressure boxes,part of the hydraulic installation is better left out.Then,
however,it is a good precaution to install under at least one of the containers an identical
dummy pressure box connected to a manometer and magnifier in the thermostat,where
the result of inadequate temperature corrections can be measured separately.For further
details refer to the literature [17].
4.2.4.2.2.4 T h e non-weighoble lysimeter
The lysinieters for practical problems are used mainly with undisturbed soil cylinders.
The principal problem is obtaining these large,undisturbed soil cores.Plastic tubes of
0.50m diameter and 1.30m length are pressed into the soil with hydraulicrams and soil
anchors.At the same time the soiloutside the cylinderis dug away to facilitatepenetration.
When the upper side of the plastic cylinder is level with the soil surface,a hole is made
next to the cylinder,as access for a bottom plate to be pushed under the cylinder (see
Fig. 4.21).
The soil coresare placed on both sides of a pit and are dug into the soilat some distance
from the passageway if gamma-ray moisture determination is applied. Where nylon
blocks or devices with a short duration of uninterrupted use are applied,the soil cores
are better placed against the outside of the wail of the pit for better access to the measur-
ing units.
The soil cores are prepared for measuring the drain discharge by filling the lower 5 c m
of the plastic cylinder with coarsesand or gravel.A dischargetube for drainageis attached
and a bottom plate is fixed.Then the cylinder is placed in an oil barrel and the space
between the cylinder and the barrel is filled with bituminous material to ensure that it is
completely watertight. The cylinders are then lowered into the soil and the bottles to
catch the drain dischargeand to supply the water lost by evapotranspirationare connected.
4.2.4.2.2.5 Lysimeters of the U.S.S.R.

Hydrometeorological Service
These consist of cylinders with an area of 0.2 mzwith depths of 1.0,lS, 2.0,and 2.5 m,
and are common in some countries.Every cylinder,filled with a monolith,has an outer
cylinder and a water-controllingdevice. The outer cylinder and the cylinder with the
monolith play the same role as those in weighing soil evaporimeters (see section 4.2.4.2).
The water-controllingdevice creates and automatically keeps the ground-waterlevel
inthe lysimetermonolith at a required depth.Measurement of the amount of water added
to or removed from the lysimeter in the process of returning the water level to the fixed
108
I Methods of observation and instrumentation
FIG.4.21. Use of lysimeter for general survey (undisturbed soil cores for study of the relation
between past nuclear fall-outand the water balance,taken in differentcountries,are brought
to a central place for investigation in order to avoid the cost of scattered research).
point correspondingly shows ground-wateryield to the unsaturated zone and ground-

water recharge by infiltration.
The lysimeteris weighed on a standard platform balance.For this purpose the lysimeter
with monolith is lifted by an autocrane and put on a balance.
4.2.4.2.2.6 Major disadvantage of the lysimeter
The aim of applying,in the field,results obtained with the lysimeter installationis limited
by the influences of the distribution of the root density over soil layers and the uptake
of moisture by the root. The plant probably has a high adaptabiIity to adverse moisture
conditions and is able to increase its capacity to extract water. Root density and uptake
activity are not,however,measurable in a non-destructiveway. Moreover,it is uncertain
that the conditionsin the field resemble those in the lysimeter to such an extent that they
may be assumed to be identical. Therefore the techniques to determine the moisture
balance of the soil covered by vegetation and under field conditions have to be applied
simultaneously with the lysimeter technique to obtain fully reliable results [218].
4.2.5 Other climatic observations, including energy balance

4.2.5.1 Climate stations
Climatic observationsare normally made at a site within a basin,termed a climatestation.
The records obtained are point records and indicate the macro-climateof the area or,
at best,the meso-climateof the basin or the valley in which the basin is situated.
All representative basins require at least one climate station (termed a base station)
109
Representative and experimenial basins
to record long-termmeso-changesin the basin climate and to provide reasonably typical

climatic data for the entire basin. For this reason it is important to select for the base
stationa sitewhich is typical of the meso-climateof the area.The milieu of the site should
remain as constant as possible for the duration of the experiment and it is not absolutely
necessary that the site be within the basin. Established stations which are part of the
national network should be utilized where possible.
Climate stations for representative basins can be manually operated, although for
isolated sites partly automated stations should be considered (see section 4.2.5.2).
For all basins where the effect,on the hydrologicalregimen,of a natura1 (representative
basin) or a cultural(experimentalbasin) changeis studied,it is essentialto study the micro-
climate in addition to the meso-climatesince,especially where vegetation changes are
concerned,the micro-climateis likely to change considerably with subsequenthydrologi-
cal effects.
This requires one base station with requirements similar to those of representative
basins in general, plus auxiliary climate stations at master sites (see section 4.1.1).
Where possible, the base station should be fully automated to provide a continuous
record of climatic variables.
4.2.5.2 Climate station requirements

4.2.5.2.1 INSTRUMENTATION
The maximum instrumentations required for various types of climate station are given
below.These requirements can be altered to suit local needs.
4.2.5.2.1.1 Fully equipped base station

Instrumentation should comprise the following: evaporation tank; evaporimeter; lysi-
meter; automatic raingauge; vecto pluviometer; cup anemometer (wind run only);
thermograph (atmospheric); thermocouples or thermometers to measure soil tempera-
ture at typical depths,e.g.,5 cm,10cm,20 cm,30cm, 1.20m and others;grass minimum
thermometer; hydrograph (hair type)-several wet-and-drybulbs could be established
at different levels; sunshine recorder; actinometers; pyrheliometer.
Automation can be achieved by chart recording or by recording signals from sensors
on magnetic or punched tape. Potentiometricrecording is expensive but is the most suit-
able method.
4.2.5.2.1.2 Auxiliary climate station

The instrumentation of an auxiliary climate station depends on the objective of the re-
search; if, for instance,detailed energy-balancestudies are undertaken in experimental
basins,auxiliary climate stations should be equipped with at least a flat-plateradiometer,
thermometers (soil and air), a raingauge,an anemometer and a hygrometer.Observations
at auxiliary climate stations can be limited to master sites in typical soil-vegetationcom-
plexes and to, say,one or two days for typical moisture conditions. Moisture conditions
(as determined, for instance, by soil-moisturemeasurements-see section 4.4.1.1)can
be classified into,say,three classes;climate observations for two days on each moisture
class are likely to provide sufficient data for correlation with the base station and, by
moving the equipment around,the micro-climateof a basin can be rapidly established.
4.2.5.2.2 LOCATION
Certain criteria for the location of climate stations are essential to ensure representative-
ness of the site and for comparing data between basins.
110
Methods of observation and iiistrnmentation
(a) 'The base station should preferably be near the centre of the basin at approximately
mean basin altitude. This requirement is not extremely important, especially if auxiliary
stations are used and the base station is used only for correlation.
(b) The dimensions of the plot for the station will depend on the number and type of
instruments required.The instruments must be arranged in such a way that each instru-
ment has adequate exposure. The plot should have a level surface and be uniformly
vegetated and away from high trees and other obstructions,slopes and hollows.Adequate
security measures should be taken to protect the instruments.
(c) There should be room for expansion of the climate station and any agent which
may affect future site exposure should be eliminated.
(d) It is important that precipitationmeasured in the basin is representative of aclual
rainfall received by the air-vegetationinterface. For this purpose, instruments should
have their recording surface level with the air-vegetationinterfaceor be a uniformly small
distance above this face.The installation of other climatic equipment at main climatic
stations should be according to the national standards.
In forested basins it may be desirable,when possible,to cut an area of the forest so
that the precipitation gauges are installed relatively close to the ground.
In each case, the angle of the gauges relative to the height of the nearest obstruction
should not exceed 45".
4.2.5.3 Instrumentation and observational techniques

Detailed descriptions of instruments and methods of observation are given in the litera-
ture [29,30,2?0,2311;limitationsin measurement and errors associated with observation
are also given [170].In particular,radiation measurements may not always have been
made for the purpose of hydrological studies.There are four fluxes of radiation near the
earth's surface: downward short-wave,upward short-wave,downward long-wave,and
upward long-wave.
In detailed hydrological research on experimental basins three or four fluxes of radia-
tion should be measured.The four fluxesusually measured are: (a) downward short-wave
(insolation,incident solar radiation); (b) upward short-wave(reflected from the surface);
(c) downward all-wave(hemispherical radiation of both short and long wavelengths);
(d) net exchangeof all-waveradiation(differencebetween upward all-waveand downward
all-wave). From these four measurements,upward and downward long-waveradiation
can,at least in theory,be determined by subtraction.The downward short-waveis meas-
ured with a pyrheliometer and the downward all-waveby a flat-plateradiometer.The
upward short-wavecan be dispensed with if albedo can be measured at intervals. As an
alternative,a pyrheliometer can be used facing upwards. Horizon obstruction is likely
to exist in the places to which radiation data are being applied (for example, in studies
of snowmelt) and this should be checked before records are analysed.At a mountain
station,radiation may be both diminished by loss of short-waveradiation at the end of
the day and increased by reflection from slopes that face the instrument, especially if
those slopes are snow-covered.Reflections from clouds cause short,abrupt rises in the
radiation record,which might not entirely cancel out with time if clouds form regularly
at the same place.
Apart from measurements of the solar beam at normal incidencefor research purposes,
most measurements of downward radiation,both short-waveand all-wave,are made on
horizontal surfaces. In hydrological studies of snowmelt or transpiration on slopes,
observation on horizontal surfaces is inappropriate.Most methods for converting them
for slopes of various degrees of aspect and steepness are valid only for the direct solar
beam. Conversion of diffuse short-waveradiation and of long-wavefrom the sky must
be done separately. It must be remembered that radiation stations are under the open
sky and measurement of radiation beneath a forest canopy, necessary in some studies,
111
has been only experimental and in types of forest that are described too poorly for the
measurements to be transferred easily to other forest types [I 32,2441.
4.3 Surface water

4.3.1 Objectives
Streamflow recording and measurements are normally problem-orientated and analysis
is concerned with quantitiesrather than with the particular qualities of the records.
For representative and experimental basins, the quality of the records is of primary
interest and analysis is undertaken with a view to a better understanding of the interac-
tion of water on the basin characteristics.L o w to zero flows represent important phases
of the hydrologicalcycle and correct recording and interpretationof such phases presents
problems not entirely overcome by present-day instrumentation and procedures.The
following sections deal briefly with present accepted approaches to stream-flowmeasure-
ment and discuss site selection, controls. precalibrated devices, instrumentation and
accuracy requirements for gauging stations on representative and experimental basins
[18, 239, 240,241,2421.
4.3.2 Streams and gauging-siteselection

The choice of a gauging station is broadly governed by the following factors. It should
be situated where measurements taken will best serve the purpose for which they are
required; be readily accessible by available transport;be such that measurements made
are sufficientlyaccurate for the purpose for which they are required;be the cheapest to
construct of all available sites.
4.3.2.1 The stage-measurement cross-section

Considerations peculiar to experimental and representative basin research indicate that
a thorough knowledge of the geological situation at and adjacent to the proposed site
is necessary (see section 5.1.2.2.1.2).
Apart from evapotranspiration,water losses in a basin may occur by deep percolation
(see section 1.5), by subsurfaceflow in the channel and by interflow which does not reach
the channel above the proposed measuring site. Where losses must be accepted,ground-
water flow studies based on permeability and pressure systems will help to assess the
relativequantities of unmeasured lossesfrom the basin.In gauging-siteselection,therefore,
advantage should be taken of favourable geological formations which ensure that all
outflow from the basin is, at the selected sections,carried on or easily accessible from
the channel surface.
Perennialflow is desirable as a basis for storm run-offcalculations,and is almost essen-
tial as a guarantee of minimal channel losses and recorder maintenance and reliability.
This may not be applicable to arid or semi-aridregions or small basins. In such cases
advantage must be taken of channel-flowperiods to test for transmission losses. Visual
inspection or the use of portable weirs will help to localize suspect loss areas which may
then be sealed or grassed over if possible.
4.3.2.2 The flow-measurement cross-section

When the station is to be fully field-rated by the use of a current meter, site selection
entails consideration of a suitable section for a cableway (manned or slackline), foot-
bridge or powered boat.
112
I Methods of observation und instrumentution
Discharge measurements of the highest accuracy can best be carried out in channel
reaches and at cross-sectionswhich conform to certain specifications [231].Water-surface
slopes can best be measured on a straight reach,but ñood rating can be carried to higher
stages when velocities are not excessive. Depth measurements are facilitated by the ab-
sence of excessive depth, line drag and surface disturbance. In flood periods, shallow
depths are associated with excessive velocities and surface disturbance.
The general requirements indicate a straight reach where the velocity range is 0.33-5
m/sec and the depth range is 0.5-6m.
The measuring section should have gently sloping banks,preferably grassed,and the
stream bed should be free of obstructions (boulders,tree stumps,etc.).
l
A gauging stationmay,therefore,consistof a stream section controlled either naturally
or artificially so as to give gauge-heightstability and sensitivity,and a section,not neces-
sarily with a low flow control,at which discharge measurements by current meter may
1 be carried out. Stage and flow ahould be measured on the same cross-section where
possible.
4.3.3 Natural controls (highflow-low flow)

The natural control at a gauging station comprises those existing physical characteristics
of a stream, at and downstream of the gauge,which determine and stabilize the stage-
discharge relation. These include the natural configurationsand vegetative cover of the
stream-bedbanks and overflow area.
A n effective control gives stability and sensitivity to the stage-dischargerelation in the
low- to medium-flow region and should be of such a nature that the stage-discharge
relation at high flows is predictable.
A gauging station with a variable control gives at best a record of low and flood-flow
frequencies. For quantitative assessments of flow,repetitive field ratings together with
laborious office manipulations are necessary.
4.3.4 Precalibrated devices-weirs, flumes and orijïces

The employmentof precalibrated flow-measuringdevices is mandatory for experimental
basin research and justifiable for the smallerrepresentativebasins in the steeper terrains.
4.3.4.1 Experimental basins

Variations in storm discharge might be so rapid and so erratic as to prevent use of the
current meter. The presence of foreign matter,floating grasses and general discoloration
prevents the use of the more rapid chemical-gauging methods. Weather conditions will
also be unfavourable because storm run-offin small basins will occur only during rainfall
periods.
A method which has proved effective is to create a temporary storage area so that
regulated flows can be maintained during the time necessary to carry out a flow measure-
ment.
4.3.4.2 Representative basins

Factors similar to those in experimental basins will be present to prevent good current-
meter work. High velocities coupled with bed roughness and passage of bed material
will provide additional hazards.
113
4.3.4.3 Desiderata for measuring structures
The desirable features of structures are conflicting and the final installationis a compro-
mise in which the followingfeatures must be considered:
The shape of the structure should permit the passage of water Rithout the creation of
undue afflux in the vicinity;
The head loss must be sufficient to maintain free flow throughout the stage range;
The structure should have as large a capacity as possible to permit the measurement of
low-frequencyfloods;
The storage area above the measuring structureshould be as smallas possible to minimize
modification of the hydrograph;
The structure should be winter-proofand should be constructed in such a way that flow
measurements can be made under winter conditions;
The intake pipe(s) should be arranged in such a way that fluctuations in the well are
minimal, that well level corresponds to river level at the gauging station and that
flushing of the well can be carried out;
The structure should permit easy calibration with the accuracy required ;
The profile of the crest should be designed to give reasonable sensitivity at ail flows
(this is particularly difficult to achieve at low flows) and this may require an additional
special structure to measure the low flows;
The structure should be leak-proofwith cut-offwalls sufficientlydeep to bring to the
surface deep subsurfacechannel flow;
The velocity of approach should suit the calibrationof the weir and the approach-channel
geometry should remain sensibly constant for long periods ;
The structuremust be self-cleaningand capable of maintaining its rating when operating
in rivers carrying silt and bed load.
4.3.4.4 Permanency of rating

The precalibration of a weir does not remove the necessity for field checks,but since the
rating of a regular geometricalstructure conformsto set laws,a deviation from the model
rating,as verified by a field measurement,is generally applicable on a percentage basis
throughout the stage range.It should be noted that back water, particularly if variable,
will upset the model rating.
Check rating of precalibrated structures,particularly at medium stage, is advisable
and volumetric ratings at low stages should be carried out at periodic intervals as a mat-
ter of routine.
Volumetric rating also serves to indicate stage-settingerrors.
4.3.4.5 Stage inaccuracies

The accuracy of flow as derived from a known stage-dischargerelation is only as accurate
as the measurement of stage. Random recording errors ensure that monthly or annual
mean discharges will tend to be more accurate than the determination of a discharge at
a point in time.Non-randomerrors arising from incorrect zero setting,insensitiveinstm-
mentation and careless maintenance tend to be cumulative and flow-measuringerrors,
particularly in the low-flowregimen,can be in error by hundreds per cent.
General information on the well-documented types of structure is to be found in
section 4.3.4.5.1below but it should be recognized that there is also a definite place for
the model rating of natural controls,modified if necessary for increased low-flow sensi-
tivity. A detailed geological engineering survey should be made before a structure is
designed. For details,see section 5.1.2.2.1.2.
114
4.3.4.5.1 PRECALIBRATED ARTIFICIAL S T R U C T U R E S
V-notch and sharp-crested weirs.l 120", 90",3(90"), a(90'). See Figures 4.224.26.
Capacity. Approximately 700 l/secwith a head of 60 cm for the 120" V-notch,and 500 l/sec
with a head of 62 c m for the 90"V-notch.Others have a proportionately smaller capacity
with a lower limit of 0.6 l/sec.
Accuracy. Percentage error= 250 dH/Hwhere dH is the allowable error in the recording of
FIG.4.22. 120" V-notch

sharp-crested weir,
Forest Service, U.S.A.
I
I
FIG.4.23. g(90') V-notch sharp-crestedweir,Ministry of Works, New Zealand.
i 1. The use of V-notches in low-lyingareas may adversely affect the ground-waterlevel. In

such areas the structural arrangement of the V-notch weir should be such that damaging
water levels may be lowered by dismantling the weir plates.
115
FIG.4.24. Sharp-crested weir of Valdai Hydrological Laboratory (U.S.S.R.);

concrete (120");
wooden (90").
116
head.Given d H = 2 mm the desirable 2 per cent accuracy is reached at H = 250 mm.For

lower heads d H must be reduced to achieve the required 2 per cent accuracy. See section
4.3.7.9.1.
Siting criteria or advantages.The apex of the notch is located at the weir height above the stream
bed. Width of approach section is six to eight times the height of the weir. The V-notch
provides a simple means of increasing the head for a given flow with consequent reduction
in the percentage discharge error arising from the allowable recording error.
Limitations. Not recommended for streams with heavy sediment load. The V-notch must be
protected against floating debris by a vertical screen about 15 c m high, measuring from
2 cm below the V-notchapex,and at a radius of approximately 120 c m from the weir centre.
Pondage and seepage problems may occur.
FIG.4.25. 120" V-notch

sharp-crestedweir with
removable plate, the
Netherlands.
Cross-section A-A 6
10
'14
IO
Legend 6. Bench mark.

1. Shelter. I. Marks. 12. 13. Rubber sheeting.
2. Well. 8. Groove wall. 14. Edge of the groove.
3. Connecting tube. 9. Counterfort. 15. Edge of the clay wall
4. Ladder. IO. Shield of the weir. 16. Floor.
5. H o o k gauge. 11. Outflow cock. 17. Crib.
FIG. 4.26. Plan of wooden weir (U.S.S.R.).
117
Compound V-notches (i). $(90") + 90"where i(90") is 20 cm high. See Figure 4.27.
Capacity. As for the 90", but with a small increase for the á(90").
Accuracy. Needs to be model rated and then checked by field calibrations.Low-stagevolumetric
gaugings are facilitated if a collector vessel is attached to the weir plate.
Siting criteria or advantages. Combines the capacity of the 90" V-notch with the low-flow
sensitivity of the 2(90") V-notchand is used where the important flows are within the capacity
of the a(90") weir (6 lisec).
Limitations. Where flow is intermittent, pondage volumes must be limited to the minimum
necessary to conform with the pondage specifications of the k(90") V-notch.Figure 4.27
shows a proposed arrangement.At high flows the intake-pipeposition will not conform to
standard requirements.
Compound V-notches (ii). $(90")+ 90"as above, but with sloping concrete side walls.
Capacity. The slope and width of the side walls are planned to suit flow requirements.
Accuracy. As for compound V-notch.Requires field or model calibration.
Sifìng criteria or advantages.Enables the V-notchweir to operate at fullcapacitywithout overflow
and provides a rateable section for higher stages.
Limitations. As above.
Compound V-notches (iii). 90" V-notch with sloping concrete side walls 25 cm thick. See
Figure 4.28 and [166].
Capacity. Five types.Rated by models up to 22,500 l/sec.(i) 2 hor.,1 vert.;)¡i( 3 hor.,1 vert.;
(iii) 5 hor., 1 vert.;(iv) 10 hor.,1 vert.; (v) horizontal (14.6 m wide).
, ,
FIG.4.27. Compound V-notch (a(90")+ 90")sharp-crested weir, Ministry of Works, New

Zealand.
118
FIG.4.28. Compound V-notch sharp-crestedweir with sloping concrete side walls, Ministry
of Works,New Zealand.
Accuracy. Given by the corresponding equations as follows where H> 0.7 m :

(i) 2 hor.,1 vert.:
(ii) 3 hor., 1 vert.:
(iii) 5 hor., 1 vert.:
2.765H4.447= Q < 8.5 m3/sec.;
(iv) 10 hor., 1 vert.:
1 3.718H3.2g4
13.523H7.308
= Q > 8.5 m3/sec.;
= Q < 2.49 m3/sec.;
(v) Horizontal (14.6m wide): 7.313H4.655= 7.87> Q > 2.49 m3/sec.;
7.497H3.927= Q > 7.87m3/sec.
Sitting criteria or advantages. Given rating equations are for a 14.6m wide rectangularchannel
with the apex of the V-notch equal to, or greater than 46 c m above the stream bed.
Limitations. For 3 :1 side slopes,a reduction of the approach channel to a trapezoidalshape
with a base width of 10.42m gives a 12 per cent increase in discharge at heads of 122 c m
and 154 cm. A curved approach channel increases the discharge by a further 2 per cent.
Cipoìietti. A trapezoidal,sharp-crestedweir with 1 :4 side slopes.
Capacity. (i) 50 c m wide: discharge range 10-85 I/sec;(ii) 150 c m wide: 30-1,000 l/sec;
(iii) 250 c m wide: 50-4,500 l/sec.
Accuracy. Low at low heads.
Siting criteria or advantages. Less susceptible to blockage than the V-notchweir.Siting criteria
are rigid to allow the use of standard rating tables. See bibliography [217,2241.
Limitations. Not recommended for use where extreme low flows occur. Heating in winter is
difficult compared with that for V-notch weirs.Fairly limited capacity range.
Rectangular sharp-crested weir.l See Figure 4.29.
Capacity. Unlimited.
~
1. The use ofsharp-crestedweirs is not recommendedfor streamscarryinga significantsediment
and debrisload,where submergedflow conditionsmay exist,or where ponding is impractical.
119
FIG.4.29. Rectangular sharp-crested weir, Ministry of Works, N e w Zealand.
Accuracy. Low at low heads [123].

Siting criteria or advantages. Construction easier than that of the Cipolletti but available dis-
charge formulas not so reliable [124].
Limitations.As for the Cipolletti.
Broad-crested triangular weir. Various side slopes. See Figure 4.30.
Capacity. U p to 53,200 l/sec with a head of 183 crn and 5 :1 side slopes.
Accuracy. Less accurate than sharp-crested types.
Siting criteria or advantages. Easily constructed. Less likely to collect debris than sharp-crested
weirs.
Limitations. The rating is dependent on the cross-sectionof the approach.This is taken 305 c m
upstream of the weir [222].
Broad-crested rectangular weir (Romijn type). See Figure 4.31.
Capacity. 10,000 ]/sec for a head of 200 cm.
Accuracy. Low at low flows, and high at high flows. Less sensitive to approach conditions than
are the sharp-crested weirs.
Siting criteria or advantages. Insensitive to submergence and no ponding necessary.
Limitations. Difficult to construct [21,22].
V-shaped broad-crested weir (V-Romijn type). See Figures 4.31 and 4.32.
Capacity. Maximum capacity 500 I/sec for a head of 60 cm.
Accuracy. High at high flows, low at low flows.
120
FIG.4.30. Triangular weir (crest 2:1 slope), Department of Agriculture, U.S.A.
Siting criteria or advantages. Used where H-flumesor V-notches are unacceptable because of
pondage or submergence.
Limitations. Difficult to construct.
Crump weir. For construction methods and photographs of installations,see bibliography 1391.
Accuracy. Dependent on accuracy of head measurement.
Siting criteria or advantages.For operation requires a minimum head loss.Operates satisfactorily
with submergence ratios of up to 0.96.Self-scouring.Construction cheap.
Limitations. General aggrading of the stream bed can easily bury this low weir.A two-pen
recorder employing two stilling wells is used. Head-measuring accuracy is difficult and
questionable because silt may block the pressure holes in the weir.
Flumes. HS, H,HL. See Figures 4.33 and 4.34.
Capacity. Range of discharges:
HS-flume:0.005-23.0 l/sec.
H-flume:0.01-2,390I/sec.
HL-flume:0.15-3,320l/sec.
Accuracy. As for 90" V-notchbut the heads corresponding to given flows are much lower with
the accompanying necessity for reduction of the allowable error in recording of head. With
submergence ratios > 0.35 the up- and downstream levels should be recorded [18].
Siting criteria or advantages. Ideal for ephemeral or intermittent flow because no pondage is
required.Self-clearingis aided by a side-slopedfloor in the larger flumes. Screening against
surface debris or bed load is recommended.
Limitations. Not recommended for perennial streams where there are sustained low flows in
the order of 0.6 I/sec because of head-measurementdifficulties.It is difficult to maintain
the level of the zero-flowsetting because of floor irregularities and blockage of intake by
silt, earthworms,algae, etc. [222].
121
Parshall flume. Throat widths range from 7.6 cm to 1,524 cm. See Figure 4.35.
Capacity. 3,960 l/sec for 244 cm throat and 16.2 cm head.
Accuracy. Good for near-capacity flows, but not good for low flows. Operates best under
uniform flow-regimen conditions,e.g.,irrigation.
Siting criteria or advantages. Does not require a pond and when used with two recorders can
operate when submerged. It is self-clearing and tolerant to variations in the velocity of
approach.
Limitations. Construction is costly because the dimensions must be maintained within fine
limits.Normally two recorders are required,except under free-flowconditions when a certain
loss of head is necessiry [223].
H-flume combined with V-notch. Preferably a i(90") or a f(90")V-notch.See Figure 4.36.
Capacity. As for H-flume.
I
Long section Cross-section.
0.1 < h/lr0.75
Meosure h upstream from the weir
at o distance of 4 h (max.)
Plan
Long section Cross-section

0.1 < h/l< 0.75
Measure h upstream from the weir
at a distance of 4 h (max.)
2 5 10' 2 5 102 2
Heod'h (cm)
FIG.4.31. Broad-crested and V-shaped broad-crestedweirs.
122
I Methods of observation aid instviimeritatioiz
FIG.4.32. V-shaped rectangular weir (V-Romijntype), the Netherlands.
FIG.4.33. H-flume
(76.2 c m deep) with
Coshocton silt-samplin
wheel in foreground,
Department of
Agriculture, U.S.A.
123
1o2
5
IO’5
5
a-
Plan
b = I.OSd(HS)
b = 1.9d (H)
b = 3.20d (HL)
;IO’
-1 5
P
$
E l
2
m k T l
Long section Cross-section
2 5 IO’ 2 5 10’ 2
Head h (cm)
H-flumos
FIG.4.34.
Accuracy. Low-stage accuracy as for V-notch;high-stageaccuracy as for H-flume.

Siring criteria or advantages.Eliminates the inaccuraciesof the H-flumeat low-flows.Normally
two recorders are used with a head drop to the separate V-notch box. Otherwise a single
recorder is used with the lower crest level of the V-notch level with the floor of the flume.
Limitations. The separate V-notch box is protected with mesh. Over-capacityflows spill over.
Standard flume ratings at high flows will be affected by a slight back-watereffect when a
single recorder is used [57].
San Dimas rectangular flume. Has a rounded converging entrance section and a sloping floor.
Capacity. 28,300 I/sec for a 305 cm width.
Accuracy. Reasonably good at near-capacityflows but accuracy decreases at low flows.
Siting criteria or advantages. Useful when extreme variations in flow are expected but where
accurate flow measurements at low stages are not important.
Limitations. Requires a large slope to the floor,and free fall to the outlet.It is therefore suitable
only on steep gradients or where there are head drops.
124
IO’
5
IO’
5
IO’
4
5
û IO‘
-2
\
- 5
m
D
P l
7
.
P I Long section Crosr.seciion
1 5 IO’ 1 5 10’ 2
Head h (cm)
FIG.4.35. Parshall
flume, the Netherlands.
N.Z. steep-floored self-cleaning flumes. See Figures 4.37 and 4.38.

Accuracy. Good at high flows. Inadequate at low flows because seasonal algae growth w ill
cling or grow on the concrete control section.A suggested improvement is to replace the
low-flow control with a (stainless) steel plate, say 6 c m high.
Siting criteria or advantages. Head measured in pool upstream of structure.The arrangements
for a field-checkrating should include a means of volumetric rating at low flows.
Limitatiotu. Field rating by current meter on the structure is difficult because of turbulence
and velocity.Model rating is advisableif there is no suitablefield-ratingsectionnearthestation.
Not suited to heavy bed-load conditions.
Swiss self-cleaning concrete flumes. See Figures 4.39 and 4.40.
Capacity. Restricted for economic reasons.
Accuracy. Difficult to assess.
Siting criteria or advantages. Suitable for mountain streams with heavy bed load.
Limitations. Construction expense limits use to small basins. Field rating is difficult.Head is
measured in an area subject to extreme afflux [186].
125
FIG.4.36. 61 c m H-Rume (capacity 312 l/sec) combined with 20 c m S(90') V-notch weir
(capacity 12.5 ]/sec), Ministry of Works, N e w Zealand.
FIG.4.37. Steep-floored,self-cleaning flume, Ministry of Works, N e w Zealand.
I26
Methods of observation and instriiineiztaiion
FIG.4.38. Steep-floored,self-cleaningflume, Ministry of Works, New Zealand.
Orifices.
Capacity. Limited to small flows where the sensitivity of other measuring devices is inadequate.
Accuracy. Five times that of a V-notch weir.It is particularly useful for flows below 8 c m of
head where the V-notch calibration is unreliable.
Siting criteria or advantages.When used with a tank containing several orifices and where
unlimited head is available, this method is the most accurate for measuring low flows.
Limitations. Head requirements for this device limit its use to sites where considerable head
(approximately 120 cm) is available.
Controls. See Figure 4.41.
Acciiracy. Insensitive,but a steeply angled section maintainsthe low-flowchannel at the recorder
intake pipe.
Siting criteria or advantages. Actas a bed-level control, but the lack of an upstream apron
permits the deposition of coarse bed material around the intake pipe. This causes erratic
low-flowrecording.
Limitations. For optimum performance requires an upstream control to produce scouring
velocities and a pool at the recording section. The upstream control is sited initially on a
temporary basis to find the optimum position for maintaining the pool. Heavy material will
be deposited at the upstream control and periodic removal will be necessary.
4.3.5 Flow measurement-current meter,

chemical,miscellaneous
Current meters provide the most convenient, accurate and economical means of measur-
ing discharge. Comparative gaugings using different types of current meter (cup and
propeller) prove that results are sensibly identical [39].Others have reported appreciable
discrepancies in concurrent gaugings done with the same type of meter, but by different
field procedures [131, 2031.The theory and practice of current-meter measurements is
given in the literature [214,2311.
127
SECTION A-A SECTION W B
25
20
Ul
L
01
L
al 15
E
I
U
O
al
r
FIG.4.39. Design,
5 f
photograph and head-
discharge relationship O
of a Borgne water- O IO 20 30 40 50 60
measuring flume,
Di schar g e mhc.
Luette, Switzerland.
128
FIG.4.40. Water-
gauging station at
Rappengraben,
Emmental, Canton of
Berne, Switzerland.
FIG.4.41. Uncalibrated bed-level control, Ministry of Works, N e w Zealand.
129
4.3.5.1 Current-meter limitations

The following limitations are observed:
1. Streams with extreme turbulence are not accurately measured.
2. Operation is impossible in velocities greater than 7 m/sec.
3. Heavy debris and sediment or weed loads preclude use.
4.Accuracy,dependent on the number of verticals,is, at best,in the order of 3 per cent.
5. Rating is necessary when cup or propeller is damaged (cup damage is frequent).
6.Unsteady stages as encountered in small basins cannot be satisfactorily measured
except when the continuous gauging method is used [192].
7. Auxiliary equipment is expensive.
4.3.5.2 Chemical-gauging methods

Measurement of the rate of flow of streams is carried out by the addition of a substance
such as salt,dye or radioisotopehereafterknown as a tracer.
Three general tracer methods are available,the procedure being much the same what-
ever tracer is used.
4.3.5.2.1 SALT-VELOCITY M E T H O D
C o m m o n salt is used and the time of passage between two measured points taken. A n
electrical detection method is used.
4.3.5.2.2 SALT-DILUTION M E T H O D
A solution of tracer is added at a constant rate and its concentration measured at the
minimum distance downstream at which mixing is complete. A turbulent stream is best.
This method, unlike the salt-velocityprocess,does not require knowledge of the cross-
sectional area. Sodium dichromate and radioisotopes(see section 4.11) have been used.
4.3.5.2.2.1 Fluorescent dye-dilution method
This method is being developed and shows promise for use in turbulent streams. The
principle involved is similar to that of the radioisotopemethod but has no corresponding
dangers and tracer suppliescan be stocked ready for emergency use.Dyesused are sodium
fluorescein (found unsatisfactory because of adsorption) and rhodamine-B (used by the
United States Geological Survey over distances up to 200 km for discharges of over
600 m3/sec). Another dye,called Pontacyl,is at present in use by the United StatesAgri-
cultural Research Service in Arizona and appears to be the most satisfactory.
4.3.5.2.3 . I N T E G R A T I O NO R ‘GULP’ M E T H O D
In this method the tracer is added rapidly as a single dose of a known quantity. At the
sampling point its concentration is determined at frequent intervals,and the mean con-
centration over the time taken for the whole ofthe tracer to pass is obtained.The advan-
tage of this method, commonly used with radioactive isotopes and dyes,is in avoiding
the use of constant-rateinjection equipment.
4.3.5.2.3.1 Use of dyes
A method known as the colour-velocitymethod consistsoftiming a cloud of dye observed

visually over a measured distance.Spillway chute measurements ofvelocities up to 24 m
have been accomplished.
130
The accuracy is low and the method is used only where velocities are too high for
other methods.
4.3.5.3 Miscellaneous methods

4.3.5.3.1 HEAD-ROD METHOD
This provides a simple way ofmeasuring dischargein shallow,comparatively swiftstreams

or on the crest of a broad-crestedweir [41].
4.3.5.3.2 SLOPE-AREA M E T H O D
This is a method used for the estimation of peak flows,based on water-surfaceslopes as

defined by flood marks (grass,debris) and the channel geometry as surveyed after the-
passage of the flood [117].
4.3.5.3.3 VOLUMETRIC G A U G I N G
This method,used for low-flowmeasurements,is particularly suitable for the calibration

of structures at stages where, owing to surface tension and fluid viscosity, model ratings
are unreliable.
The discharge measurement is made by diverting the flow into a container of known
volume and timing the filling process. The average of a series of calibrations is taken to
minimize timing errors at start and fìnish.
4.3.5.3.4 OPTICAL C U R R E N T - M E T E R G A U G I N G
This is used by the Geological Surveyformeasuring velocities of 10 to 20 m/secas encount-

ered in lined channels. The meter is a stroboscopic device measuring surface velocities
without immersion and the results require the use of a coefficientto adjust for mean velo-
city. Use is obviously limited to the smaller channels.
4.3.5.3.5 .MEASUREMENT O F F L O W T H R O U G H CULVERTS
Culverts can be used as peak-flood measuring devices in small catchments. Correlation

with rainfall is generally carried out as a means of extension of flood records.Peak levels
are recorded using a crest gauge and the flow is estimated by slope and channel-geometry
methods or by current-metergauging extensions.
For more refined measurements the culvert is, by use of sloping weirs,made sensitive
for low flows [187,226], and rated by model studies.Stage height is recorded continuously
by use of an automatic recorder.
4.3.5.3.6 M E A S U R E M E N T OF F L O W T H R O U G H C O N T R A C T E D OPENINGS
This method,based on the draw-downthrough a contracted opening such as that between

bridge piers, is used for flood estimation. Results are possibly more reliable than those
found by use of slope-areamethods,but the method is not as widely used [45].
4.3.5.3.7 FLOAT M E A S U R E M E N T S
In the U.S.S.R. one commonly used flow-measuringmethod is the use of floats. This
requires measurement of time taken for a float to pass between successive cross-sections
to obtain a measure of the surface velocities.Coefficients are used to convert the surface
velocity to the mean velocity. Such coefficients are obtained from analysis of detailed
131
current-metermeasurements and by analysis of hydraulic channel characteristics.In the

U.S.S.R.aeroplanes are widely used for such float measurements and special devices
have been constructed to drop the floats at rapid regular intervals [238].
Such methods are very useful for regular measurement and for flash floods when it is
impossible to use a current meter.
4.3.6 Stage measurement

Stage or gauge height is the level ofthe water surface above the zero of a gauge.
For artificial structures the gauge height zero corresponds to the known level of zero
flow,but for natural gauging sections the staff zero may be taken as 1 m lower than anti-
'cipated low-flow levels. This procedure guards against possible lowering of the stream
level by flood damage to the existing control.
The staff zero should be fixed by reference to several locally positioned benchmarks so
that staff replacement can be carried out accurately.
4.3.6.1 Staff gauges

Staff gauges are used in conjunction with automatic recorders to provide an external
check on the recorded water level. In section 4.3.7.7,intake pipes are discussed,Vertical
staffs should preferably consist of 2 m sections stepped back from the water's edge to
avoid flood damage.
Portable,stepped staff gauges (see Fig. 4.42)are also useful. Enamelled steel or plastic
staffplates in 1 m lengths and graduated in decimetresand centimetres are fixed to durable
wooden staff boards.
Staff gauges sloped to suit the river banks cause less afFiux and obstruction to swiftly
flowing streams.Graduations of the gauge board on site are necessary.
Ifa 30"angle to the horizontalis suitable,it is advantageousto use standard staffplates.
4.3.6.2 Automatic stage recording

Means of conveying the stream level to the recording unit include:(a) a float suspended
by wire or steel tape and operating in a vertical float well; (b) a pressure bulb operating
through copper tubing; (c) a servo-manometeroperated from a mercury manometer,
measuring variations in gas pressure with depth of water.
4.3.6.3 Float-recorder stations

The recording media used to date are of two types;graphical charts and punched tape.
The qualities sought are stage and time recording to a scale suitable for ease in interpre-
tation of the chart to the required accuracy. The stage-reversing strip-chartrecorder,
though initially expensive,becomes economical because daily or weekly chart-changing
routines can be eliminated.
The advent of units for the conversion of chart information directly to computer tape
has reduced the advantage of the punched-taperecorder.
4.3.6.4 Punched-tape recorders

These have the advantage that the tape can be mechanically translated to cards and pro-
cessed by an electronic computer. The time of unattended operation is generally three
months,but may be longer if necessary.Stage is recorded to one-hundredthof a foot.The
power source for operatingthe punch-outmechanism may be A.C.or D.C.(see Fig.4.43).
i 32
Methods of observation and iiutrumentalion
FIG.4.42. Water stage measurement by port-

~ able staff o n a stepped gauge, U.S.S.R.:
(a) step-
ped pile gauge (i, bench mark; 2, highest high
water; 3, lowest low water); (b) portable meas-
uring staff with stilling device, for stepped pile
gauge.
133
FIG.4.43. Digitaliwater-levelrecorder,Geological Survey, U.S.A.
A disadvantage of these recordersis the absence of a visual record of stage,calling for

an elaborate system for detecting and restoring periods of either missing or incorrect
records (see section 5.4).
4.3.6.5 Pressure-bulbrecorders
These have the advantage that a small-diametercopper tube connects the pressure bulb
to the recording instrument,thus avoiding the expense of a stilling-wellinstallation.
A disadvantage is that scale limitationsmake the instrument essentially a utility instal-
lation to be used where highly accurate stage records are not required.A weekly-chart
unit is the standard equipment,but recent developments include a strip chart and range
adjusters.The CSIRO of Australia has developed a pressure-operatedrecorder which can
operate for a year or more without attention.The power supply is a 1.5volt battery [40].
4.3.6.6 Servo-manometer
This is specifically for use where the installation of standard stilling wells ia difficult or
impossible because of unstable terrain, swamp and/or slumping river banks. The main
componentsare a gas purge system,a servo-manometerassembly and a servo-controlunit.
134
Polyethylene tubing leads the purging gas (nitrogen) to the operating position located
in the river. The pressure from the gas-purge system corresponds to the water pressure
and through a servo-metercan operate a standard water-levelrecorder.The power unit
is a dry battery and field calibration of the instrument is required.
The advantages are portability and hence easy removal to alternative sites. The dis-
advantages are the initialexpense and the possibility of temperamental behaviour.
4.3.7 Operation, maintenance and accuracy requirements

of gauging stations
The operation and maintenance of a gauging station begins with the completion of the
stationconstruction.Thereafter,the quality and continuity ofthe recordsis dependent on
the training and conscientiousobservations of the observer,and on periodic station in-
spections by the engineer in charge.
Good-qualityrecords require that water levels as indicated by the river staff be read
accurately,that unusual phenomena which may give rise to or explain discrepancies be-
tween recorded level and staff level are noted and correctly interpreted and that missed
recordsare kept to a minimum and are replacedby syntheticrecordswhich are as accurate
as possible.
Continuity of record requires that the observer be familiar with the simpler sources of
instrumental failure and, because of the increasing complexity of instrumentation,the
observer should be aware of the extent to which he may tamper with the instrument.
4.3.7.1 Winter operation

The operation ofa flow-measuringstation,particularly where artificialstructuresare used,
will be seriously affected by freezing temperatures.Structures requiring a ponding area
will be most affected and ice-removal measures include electric-cableheating and the
constructionof a shelter heated by a propane radiant heater.
In countries with severe winters, the only method of preventing frost is to arrange a
frost cover over the reach upstream and downstream of the gauge and over the station
itself. In the design of a gauging station this factor has to be taken into account.
The cover usually consists of several units and is removed in warmer periods.Snow
which may accumulatein the channelprevents a free passage offlow during floods (snow-
melt periods); it is essential,therefore,to remove the snow or to build a special guard to
prevent snow accumulation in the channel.
4.3.7.2 Effect of ice on the stage-dischargerelation

The presence of ice in a river or stream will affect the stage record. If the ice is between
the station and the control the stage will be increased and if it is upstream of the station
the stage will be decreased [45,2151.
4.3.7.3 Freezing of stilling well

The freezing of the floatin the well may be recognized by the abnormally steady stage
record. Prevention methods include the use of salt or oil on the basis that little inter-
change will take place between the salt-or oil-treated,stilled well water and the flowing
water outside. A more positive approach is to lower an open pipe, of diameter slightly
greater than the float,to a depth sufficient to prevent the escape of oil at the lowest water
level.Approximately 50 litres of kerosene is poured into the pipe and the float is lowered
into the kerosene.Where the choice is between electric power and propane gas,electricity
is more effective, but it is not necessarily the most reliable ice preventative method. Such
methods are, moreover, useful only where light frosts occur.
I35
4.3.7.4 Summer or tropical operation

Humidity and insects will adversely affect the operationof stream recordersin the tropics.
[176].
For recorderconstructionzinc-treatedsteelis used.Recorder charts should be protected
against excessive humidity by the use of specially treated charts and drying agents (silica
gel, etc.). Recorder parts will rust unless protected. Closely fitting doors and windows
should be used to prevent insects from devouring charts and fouling or drinking recorder
ink,and the recorder should be sealed against moisture by using sponge rubber where the
float wire enters the well.
4.3.1.5 Blockage of control

The blockage of controls, particularly important and frequent in artificial structures,
occurs because of freezing (see section 4.3.7.1) or because of the lodgement of boulders
and trees (and in the smaller structures,grasses,leavesand feathers). Preventativemethods
include the screeningby mesh (partial or complete) of the structure (see section4.3.4.5.1).
4.3.7.6 Float-recorderstructures
The recorder structureis made up of two parts :(a) the recorder housing of size and mate-
rial to suit individual requirements and local conditions of humidity and temperature;
(b) the recorder well which may range in diameter from 40 c m to 125 c m and may be
made of concrete,iron or even of wood.
The concretestructureis used for the permanent key stationswhere ancillary recording
attachments,electric plum bob, telemetering devices, and occasionally both a graphic
and digital recorder may be required.
In general,the structure has to be adapted to suit the gauging site (bridge piers,etc.)
and the flow conditions (sediment concentration,etc.).
4.3.7.7 Intake pipes

T w o intake pipes are normally used, but for controlled sites che second intake pipe is
unnecessary because of the absence ofexcessivebed-levelchanges.Furthermore,twointake
pipes give rise to circulatory currents in the well, so that where suspended sediment con-
centrations are high,complete sedimentation of the well and intake pipes may occur even
in the course of a single flood.Where a high silting rate is expected the distance of intake
pipe invertfrom well invert should be 1 m. The relation of the intake pipe area to the area
of the recorder well is important.Excessiveintakepipe area resultsin diminished damping
of surface waves, while an intake pipe area proportionallysmall will cause damping-out
of rapid water-levelchanges and underestimation of peak-flood levels.
For small basins, and in general where rate of stage rise is in the order of 10 cm/min,
a ratio of 1 per cent is recommended.
For lake recorders where stage changes are slow and waves and short-term seiches
confuse the record,an intake-pipearea of less than 1 per cent is recommended.In both
cases it should be noted that silting of the pipe must be considered and regular flushing,
particularly of small basin stations,is necessary.
4.3.7.8 Flushing
Standard flushing equipment for permanent recorders should include a flushing tank,
a hand pump and an intake pipe with valve. Local meteorological or water-supply
conditionsmay permit the replacement of the hand pump by a reservoir of size sufficient
136
I to permit repeated flushings. Flushing of the small-structureintake pipe and well will,
because of access problems,require the use of a portable pump for desilting.The recorder
float must be removed.
4.3.7.9 Accuracy requirements

I
Streamflow measurement should be undertaken with the knowledge that the accuracy
requirements will exceed those obtainable in the measurement of related characteristics
in the basin. Such characteristics are point samples and are dependent on replicationfor
spatial accuracy.It would be absurd and economically impossible to measure each point
sample with the accuracy which is necessary and possible for an integrated sample such
as streamflow.
4.3.7.9.1 A C C U R A C Y OF CALIBRATED STRUCTURES
A n accuracy of f 2 per cent is claimed for model-ratedstructureswhen flow occurs at the

least sensitive section of the characteristiccurve.This section ofthe curve may be located
by reference to the general dischargeequation q = CHn,by means of which the percent-
age sensitivity may be defined as (lOOn/H) dH where dH is the allowable or probable error
in stage.
For a given value of dH, the greatest percentage error in discharge occurs when H i s
least (assuming that the exponent n is a constant). Therefore, to maintain a constant
sensitivity,the proportion dH/Hmust be preserved and with falling stage this becomes
difficult.
It may be shown by means of the general equation q = CHn that,for n = 2 (an average
value), &must be read to within 1 per cent of H to achieve an accuracy of 2 per cent in
discharge.
A value of dH = 2 mm is a practical limit to recording accuracy; therefore H = 200
mm represents the lowest stage at which it is theoretically possible to achieve a 2 per cent
discharge accuracy.
The head H may,for a given discharge rate,be increased by constricting the structure
outlet. Practicalmaintenance problems limit the use of constricted openings to the a(90')
V-notch.A n alternative method of achieving sensitivity is the use of a device with a
value of the exponent n considerably less than 2.For comments on orifice usage (see
section4.3.4.5.1). The special advantages of the H-flume,for which the exponent increases
with stage from II = 1.5to n = 2.5,should be noted.
4.3.7.9.2 A C C U R A C Y O F FIELD-RATED CONTROLS
When the recommended field practices are used,the accuracy of a current-metergauging

may be within 4 per cent of true discharge but, as a rating table represents the mean
values of many gaugings,the final accuracy will approach 2 per cent.
At heads of 1 m or more the stage heights should be read and recorded to within 1 cm.
At heads of less than 1 m the measurements should be more accurate.In general,readings
should be within 1-2mm.
I
4.3.8 Run-of plots
:1 A run-offplot is a part ofa slopeisolated from the surroundingarea and used for measure-
ments of overland flow.Run-offplot dimensions,depending on the programme,may be
from 30 to 40m2up to severalhectares (see section 2.4.1).
I A run-offplot is surrounded by a water dividing wall and on the lower part of the plot
137
a flume is constructed (Fig.4.44).Along the walls,on the outside of the plot,small ditches
are made to prevent water flowing from the adjacent slopes into the plot. The gauges
are normally placed in a shelter and from them the water is removed either by means of
an underground pipe or by a collecting channel.When pairs of plots are used, the gauges
can be placed in one shelter house. Water dividing walls may be removable or stable.
In the U.S.S.R. stablewalls are made oframmed clay or reinforced concrete.and remov-
able ones are made of timber or sheet iron.Removable walls are demounted for the period
of mechanized soil clutivation of the plot. The walls are dug in at a depth of 20 c m and
raised above the land surface by 15-20c m (earthen walls by 20-30cm).
2
\ Cross-section A -A
Legend 4. Shelter for measuring 3'

I. Ditch. devices.
2. Concrete wall. 5. Protecting cap.
3. Water-collectingfiume. 6. Protecting metal sheet.
FIG.4.44. Plan of a run-off plot used in the U.S.S.R.
Collecting channels are made of concrete,or tarred wooden planks puttied at joints
and covered inside with roofing iron,or of metal welded strips,or of metal pipes with a
cut-off sector equal to one third of the pipe circumference.For ease of maintenance,
flumes with a rectangular cross-sectionof not less than 10 c m width are usually used.
The outer wall of a flume is made 20 c m higher than the inner one and is raised 10 c m
above the land surface (Fig. 4.45).The inner wall of a flume should be lower than the
land surface by at least 10 c n but not more than 20 cm. T o prevent water leakage
through the joint between the inner flume wall and the plot, a metal sheet is placed in
such a way as to cut nearly horizontally into the ground by 15-20c m (with a slight slope
in the direction of the flume) on one edge and to rest with the other edge upon the flume
wall [163, 1691.
To prevent precipitationfailing directly into the collecting channel the latter is covered
with a metal cap. The flume slope is made as high as possible.
A shelter housing the gauge is either dug into the ground or placed outside the plot at
a distance of 5-10m.The shelterroofis made to rise not more than 0.5m above the ground
surface in regions which have snow cover in winter.The inner shelter dimensions depend
on the size of the measuring tank and the means of water removal.
Inthe shelter,a metal measuring tank and a weir are installed(Fig.4.45).For water-level
measurements a hook gauge and a water-levelrecorder are placed above the tank.
The recommended dimensionsof the measuring tank are:length,1.6m;width,0.8m;
height,up to the weir. The dimensions of the weir (angle and maximum head) are deter-
mined by the value of the estimated peak discharge.
138
Methods of observation and instrumerztatioon
Legend
I. Hook-gauge. 4. Measuring tank, with weir 7. Sewer.
2. Netting. cut (6). 8. Metallic gauge for the
3. Inflow tube. 5. Outflow tap. hook-gauge.
IIG. 4.45. Measuring devices in a run-off plot shelter.
Low-flowmeasurements are made volumetrically,allowing water to flow into the tank

for at least 15-20min. For high discharges the flow is measured by the weir,which is
calibrated up to a head of 50-60mm.
The accuracy of flow measurement with the volumetric method is 1-2per cent, with
a weir 2-4per cent for a head above 50 mm,and 5-10per cent for a head below 50 m m .
In cases where high discharges or intensive soil losses from the plot are expected and
where water removal from the measuring tanks by sewer-pipesis difficult because of
minimal slope,no shelter is erected and the gauge is placed in an open flume.A triangular
weir with thin walls or a combinationof a flume and a small weir placed below the flume
is used as a gauge. A continuous record of the head is obtained by a water-levelrecorder
located in the stilling well linked to the flume.Submergence of flumes at high head is
prevented.
Some run-offplots are used to measure overland flow and interflow.These are called
water-balanceplots in the U.S.S.R. and they are used in regions where relatively impervi-
ous strata occur no deeper than 3 m (Fig.4.46).
Such run-offplots have the following additional features:(a) an impervious wall along
the boundary down to the first relatively impervious stratum and dug 20-30c m into the
imperviousformation;(b) a separate receiving device;(c) a tank for measuring interflow.
The interflow measuring device is a perforated pipe with two thirds of its circumference
provided with wire and gravel filters. The pipe is located at the depth of the impervious
stratumunder the flumewhich serves as the collector for overland flow (Fig.4.47).Where
the imperviouslayer is relatively shallow,open channels are used.
Observations on run-offor water-balanceplots requiremeasurementsof the water level
in the measuring tank and of the head over the weir and a daily change of recorder charts
or tapes.
139
Representative und experimental busins
FIG.4.46. General
view of the forest waiter-
balance plot, Valdai
Hydrological
Laboratory, U.S.S.R
During flood periods (whether caused by rainstorms or by snowmelt) special attention

is paid to the gauges and a 24-hourwatch is usually organized,with an observerexamining
theflumes at regular time intervals,carrying out readingsand making corresponding notes
on recorder tapes and in a diary.
During dry periods,the run-offplots are always ready for sudden registration of storm
flow. If run-offplots are located at a considerabledistance from a populated area,then
stage recorders (with an automated device to set them into motion at the moment the
water level starts to rise in the measuring tank or in the stillingwell) are used.The use of
such recorders makes it easier to carry out routine maintenance of plots and their prepa-
ration for storm-flowregistration.
In winter,the water is removed from the tanks and,to measure flow during thaw,stil-
ling wells and weirs are winter-proofed in such a way that they are operative under all
conditions.If interflow occurs during winter,shelter houses of water-balanceplots should
be supplied with heaters.
For run-offplots, exact synchronization of rainfall and run-offis desirable.This can
be done by using a single clockwork mechanism to record both events.Twin pens are
used on a chart with a preferable speed of 4 cm/hr.The flow-measuringstage scale needs
to be at least 1 :i.
4.4 Subsurface water

4.4.1 Water in the unsaturated zone
4.4.1.1 Soil moisture
Soil moisture plays a significant role in the water balance of an area through its influence
on the infiltration rate,the run-off,the evapotranspirationrate and the storage available
140
Methods of observation and instrnmentaation
for infiltration water prior to ground-waterrecharge.Since soils vary significantly,both

horizontally and vertically,collection of areal moisture content information is extremely
difficult. Methods for predicting soil moisture or soil-moisturechange in a basin using
other climatic factors have been developed, but these procedures require further assess-
ment before recommendation for general use.
Legend
1. Perforated tube for inter-
flow.
2. Metallic netted filter.
3. Gravel filter.
4. Impervious wall (clay or
concrete.
5. Clay bed.
6. Polyethylene film.
7. Flume.
FIG.4.47. Water-
collecting devices for a
water-balance plot,
U.S.S.R.
4.4.1.1.1 NETWORKS
In soil-moisturemeasurement,it is likely that it will always be possible to determine the

change in average soil moisture with time more accurately than the average soil moisture
content at any specific time. The prediction of either factor to a high degree of accuracy
will,however,be possible only when the dynamics of the phenomena involved have been
successfully explained by means of mathematicaland physical models.
As a step towards this goal, it is necessary to relate soil-moisturedata accurately to
other climaticvariables. It is therefore recommended that,in order to minimize inaccura-
cies in the correlation of such data,access tubes for neutron equipment be placed near
each basin raingauge at master sites (see section 4.1.1),i.e.points where the vegetation,
topography and soil are representative of a specific part of the basin. Although access
tubes may be readily installed by hand,truck access to the sites facilitatesuse of the heavy
neutron probe and scaler.This advantage should be considered in planning the network.
When determining the mean basin soil-moisturecontent,it is necessary to determine
the following:whether or not the method used for averaging is scientifically sound;range
of soil moisture fluctuations within the plot or basin; and the error of the mean with a
given number of observation points (N).
The method of determining the mean basin value depends on the method of distrib-
uting observationpoints.The error associated with the mean will be lower if sound sam-
pling techniquesare followed (see section4.1.1).Inthecaseofevenlydistributed measuring
points, the data on soil-moisturestorage, calculated at every point, are subjected to
141
statistical,processing for the estimation of the mean basin soil-moisturecontent (W )

and the variation coefficient (Cu).In this case, the relative error of the mean (P)with a
probability of 0.67 may be approximately estimated by the equation :
Cu.100
P.lOO= --. (7)
1’N
Hence,it follows that the number of points necessary for obtaining the mean basin soil-
moisture content with a given accuracy P is given by the equation:
This equation may be applied for the approximateestimationof the number of measuring
points for soil moisture on a typical plot or on a whole basin.
The determination of the mean water content in the soil for the whole basin can be
carried out using data obtained from homogeneous sites;the averaging must be done
taking into account the percentage area of every plot.
It has been shown that, for the measurement of a change in soil moisture, a lower
sampling accuracy is required than for the determination of soil-moisturecontent at any
specific time, and also that soil-moisturedetermination at subsites does not need the
accuracy required at master sites [96].
4.4.1.1.2 MEASUREMENT METHODS A N D EQUIPMENT
Soil-moisturedata,depending upon the discipline interested in this information,is expres-

sed as (a) the average amount of water in a soil or soil layer at a specific time, given as
a volume for easy conversion to depth in centimetres;or (b) the change in average amount
of water in a soil or soil layer with time;or (c) as the availability of water in a soil or soil
layer,expressed as the percentage difference between the field capacity and the wilting
point.
Currently,the five most frequently used techniques in the field are; tensiometers for
controlled plot studies,electrical resistanceunits for obtaining a continuous recording of
soil moisture in top soil layers or relatively dry soils,thermogravimetric methods (which
are simple and economical and form the standards for calibration of other methods),
lysimeters for measuring the change in soil moisture with time,and neutron-scattering
methods (which, although involving high capital cost,are the most accurate and should
prove time-savingwhen employed on large projects). The neutron-scatteringmethod is
recommendedfor use on representative and experimental basins.
Lysimeters are discussed in section 4.2.4.2.2.Brief surveys of electrical resistanceunits
and the neutron-scatteringtechnique are given below. For further details of soil-moisture
measurement methods and equipment, see section 4.11.3[71,194, 225, 231, 2431.
4.4.1.1.2.1 Electrical resistance units
Calibrationof individualunits in individual soils and soil horizons is essential for quanti-
tative measurement of soil-moisturecontent.A calibrationcurve may be prepared by the
use of a pressure-membrane apparatus. The resistance and soil-moisturecontent are
determined for several values of soil-moisturetension and the number of calibration
points required on each block is determined by the uniformity of the blocks. The bulk
density of field samples must be known in order to calculate the percentage volume of
soil moisture.A n unknown hysteresis error is included in the calibrationof the resistance
unii.
142
Meihods OJ observation und instrumenturion
Where marked wet and dry seasons occur,simple resistance units can be very effective
in supplying useful information on soil-moistureconditions.This is not to suggest that
they should be used to compute actual soil-moisturecontents from precise calibrations,
but simply to be used as ‘on-off’switches indicating whether water is available or not
at specific depths in the soil [174].This method is useful for checking leaks in basins.
If run-offis an important facet of the hydrological problem,resistance units can also
be used to study the depth of penetration of rainwater under different land treatments or
covers [51].
Resistance units are particularly useful when reliance must be placed on observers
with very little training,
4.4.1.i .2.2 Neutron-scattering technique
The greatest advantage of the neutron method over the other methods outlined in section
4.4.1.1.2is achieved when the study requires knowledge of the change in average soil
moisture with time. Experiments conducted in the U.S.A.for example, showed that
thirty-eightsites were required to obtain the soil-moisturecontent to a standard error of
1 per cent, but only two or three sites to measure the change in moisture to the same
precision.It is fortunate that for most basin research programmes greater precision is
required for the change in moisture than the actual content at any one time [234].
4.4.1.1.2.2.1 Access tubing. Plastic, steel, stainless steel and aluminium tubing have
been used and all are satisfactory if thin walled, as long as the tubing has a close fit to
the depth probe and a calibrationhas been run with the tube in position.It may be possible
to installthe tube by driving it down at the same time as the soilis augered out from inside
the tube.A solid rod driven down inside the tubing may be needed to move small stones
aside in gravelly subsoils. A n alternative method is to drill the hole mechanically. It is
important that the tube should fit tightly with no air spaces around it and a small mound
of earth should be pressed around the top of the tube to stop water intake. If the soil
cracks badly around the tube,a new installation is required. The bottom and top of the
access tubing should be sealed with rubberstoppers.It is also desirable,when installingthe
access tubes, to locate the tops of the tubes a uniform height above the soil surface.
Thus a certain distance of lowering will then bring the depth probe to a common soil
depth for all tubes.
4.4.1.1.2.2.2 Menscirement site. It is essential that the site be representativeof its im-
mediate area and that the conditions be maintained constant.Care must be taken not to
trample the vegetation or to compactthe surface in the immediate area of the access tube.
To prevent a change in run-offand evapotranspiration occurring right at the tube, it is
advisable to use a small stand when taking measurements.A smooth,flat (not necessarily
level) area must be maintained for the surface-probemeasurements.
4.4.1.1.2.2.3 Calibration. The most accurate aid to calibration is a metal drum with
height and diameter of 1.25 m and filled with soil of known moisture content and bulk
density.Although the difference in calibration between soils is often small,it is advisable
to complete careful measurements on the full range of soils to be encountered in the
research.Fixed standardsshould be used for periodic checking of the probe.The applica-
tion of the neutron method to the determination of evapotranspiration is discussed in the
literature[23,711.
4.4.1.1.3 F R E Q U E N C Y OF M E A S U R E M E N T
Soil-moisturemeasurements should be made frequently,with the most importantrequire-

ment that observations reveal the soil-moisturepattern throughout the year. In periods
143
Representatbe and experimental basins
with slight soil-moisturechanges,infrequent observations are sufficient;in other periods

a more frequent sampling programme should be organized.This will permit an accurate
assessment to be made of the effect of the growing period and seasonalchanges.After one
year’s operation,however, this schedule should be reappraised,as experience may show
that a less frequent interval is adequate.
The soil-moisturecontent of the surface O to 15 c m layer should be measured using a
neutron surface-moistureprobe (not suitable for forest lands) or sampling for thermo-
gravimetric-moisturedeterminationsand bulk-densitymeasurements.The neutron depth
probe is recommended for soil-moisturemeasurement. in 15 c m increments,from the 20
c m depth.
4.4.1.1.4 LABORATORY METHODS
One method of measuring moisture stress existing in soil-moisturesamples obtained in

the field uses chemically treated ñlter paper as a passive gravimetric moisture stress sensor
[154]. The usefulnessof this method is illustrated in severalDapers [26,83,1551.
4.4.2 Water in the saturated zone

All representative and experimental basins should be instrumented to permit definition
of the geometry of the ground-waterflow systems and to permit quantitative evaluation
of their significance in the hydrologicalbalance of the basin.
For water-balancecalculations in representativeand experimental basins,the study of
the upper aquifers (which have a hydraulic relation with streamflow) is particularly im-
portant.
As ground-waterflow systems are regulated by the permeabilityofthe surficialmaterials
and a basin may be relatively ‘tight’or ‘leaky’,it is essential that the degree of ‘tightness’
or ‘leakiness’of a basin be determined and that if the hydrologicalmodel developed for
the basin is to have any validity a measure of the amount of water entering and/or leaving
the basin as ground water be known (see sections 6.2and 6.3).
Where possible,research basins, particularly experimental basins,should be relatively
tight. Studies of leaky basins are, however,required to enlarge our knowledge of the
significance of ground water in basin hydrologyand thereforeno basin should be abandon-
ed unless the degree of leakiness introduces problems too great to be overcome by pres-
ent methods of study or by available finances.
The use of radioisotopes in ground-water studies has been most successful and is
recommended wherever facilitiesare available.For details,see section 4.11.5.
4.4.2.1 Hydrogeological assessment of a basin

To evaluate the degree of tightness of a basin, the following procedure is recommended:
1. Using available data,develop a three dimensional map of the basin to its divides and
to a depth encompassing the first basin-widelow-permeabilitymaterials (i.e. the aqui-
clude); this model willpermit a first approximationto be made of the probable ground-
water flow systems-recharge, storage and transmission and discharge).
2. Complete and quantify the above model by including field information on well inven-
tory,drilling,samplingand E-logging,major water users,sample testing of permeabili-
ty, geophysical data obtained by seismic and resistivity techniques,and pump or slug
tests in test holes if possible.Note that the quality of water in springs and perennial
streams is a good indicator of the aquifer from which the water is derived.Water tem-
perature may also be related to the character and depth of the aquifer.
144
Methods of observation and instruinentalion
Ground-waterinstrumentationshould then be planned to fit the flow systemshighlighted

by the above studies.If facilitiesare available,model studies of the flow systems should
also be completed (see section 5.1.2.2.1.1)[64, 1391.
4.4.2.2 Instrumentation
4.4.2.2.1 GEOLOGICAL S U B S U R F A C E T E C H N I Q U E S
Ground-water equipment is designed to permit location of the ground-water surface

(water table) and determination of the ground-waterenergy potential (measured by pres-
sure) of a series of points. Pits,wells,springs and piezometers are used to observe ground-
water conditions.The depth of an observation well is determined by the ground-water
level position and the aquifer thickness.While drilling the wells, soil samples are taken
every 10-20cm. A well consists of a sectionalcolumn of pipes at least 50-75mm in dia-
meter, a filter,a settling tank and a well collar.The filter is the most important part and
is designed for the admittance of water into the well and for prevention of well silting.
The filter type is determined by the mechanical composition of the aquifer.If the aquifer
is 2-3m thick,the filter is installed for its entire length.In case of thick aquifers,the lower
portion of the filter should be out of the ground-waterlevel fluctuation zone. A loam
collar is arranged near the pipes to prevent surface-waterpenetration down the well.In
case of a stage-likearrangement of aquifers,they should be carefully isolated from one
another.
When ground water is not far (up to 2-3m) from the surface,the observations can be
made in wells arranged by pit digging. Structurallythese wells are similar to those above.
The followinginstrumentsare used to observe the ground-waterlevels:
Water-levelgauge-where the water table is within I m;
Electrical contact gauge-for wells of 3-5 m deep (Fig. 4.48);
Floating-typegauges and level recorderswith a long-periodclock-at depths ofsome 10 m;
Pops and water whistles ofvarious designs which are lowered down the wells suspended
on a marked tape or cord;
Various transmitting-typegauges;
Piezometers and pressure gauges (piezometersmeasure pressure at a point-they are open-
ended 5 c m pipes slotted at the bottom).
Water levels should be read with an accuracy of 1 c m from the well collar and corrected
to the level of the surface.The position of the well collars should be checked periodically
by levelling from the nearest benchmarks.
To measure the ground-water temperatures with an accuracy of O.l"C,portable or
stationary mercury or electrical thermometersare used.
Water samples are taken for chemical analysis with the help of cylinders of various
designs,bathometers and pumps. Prior to sampling,the wells are pumped to review the
water.The wells used for observations of water quality should be lined with inert materi-
als: ceramics,asbestos cement,duralumin,etc.
The number of ground-waterinvestigationwells can be kept at a minimum if electrical
resistivity equipment and/or radioactive tracers are used to define the flow patterns. A
nest of piezometers should be installed at each point where a well is located [73].Draw-
down tests evaluate the effectsin the differing aquifers and test wells should be sampled
and correlated with the major, deep,exploratory wells drilled in the initial phases of the
project. In-placepermeability tests with small vacuum-typedrill rigs are described in the
literature [i, 55, 137, 142, 1431.
4.4.2.2.2 GEOPHYSICAL SUBSURFACE TECHNIQUES
Electrical resistivity and spontaneous potential of earth materials in the vicinity of a bore
hole are measured. A sonde,which is lowered into the hole and recorded as it is pulled
145
out, has four electrodes-two for emitting current and two for recording potential. The
precise definition of formation boundaries is best accomplished by close spacing of the
electrodes. A resistivity log may reveal the lithology and sequence of rock formation and
formation boundaries ; the presence, approximate ionic concentration, and location of
fresh and salt-waterbodies present; the amount of casing in an old well; and an estimate
of porosity. The spontaneous potential log indicates permeable zones (in terms of relative
but not absolute permeability) and may be used to compute ground-waterresistivity [i 501.
Legend
1. Brass tube with scale.
2. Driving insulators.
3. Contacts.
4. Voltmeter.
4
\s + - - i/ I
l
I
5. Electric battery.
6. Shelter.
'3
FIG.4.48. Electrical
contact gauge.
ïhe two most commonly used radiation methods are the neutron and the gamma-ray
techniques.Both methods are used for determination of the porosity of any one formation
(see section 4.11.5).
A sonic log records the velocity of sound waves travelling through a formation.The
speed of propagation is dependent on the elastic properties, porosity, fluid content and
pressure in the strata. Reliable indications of low porosities and denser lithologies
can be derived by this method. The literature discusses well-logging methods in ground-
water investigations [118].
In conjunction with electrical logging programmes, geochemical subsurface methods
could be used. Temperature logs are easiest to acquire and are useful in delineation of
water from separate aquifers.
146
The results of pump, slug or recovery tests carried out on wells or piezometers can be
used to evaluate aquifer constants and the hydrological parameters of the subsurface
water. The data collected may be in the form of a continuous analogue, with float or
small-holeadapter,or for long-termrecords,a paper or magnetic tape. Either form can
be easily processed by electronic means and the accuracy of these measurementscan be
made to 0.3 cm.
I 4.4.2.3 Distribution of measuring points
1 Assuming a simple homogeneous and symmetrical basin,which requires instrumentation

of only one side,the followingstations(see Fig.4.49)constitutea minimum instrumented
1 network.
i. Recharge stations;one water-table well to record water-tablefluctuations and one
two-pointpiezometer set to measure downward potential.
2. Transmission and storagestations:two to three singlepiezometers or water-tablewells
along the flow direction;if multilayer, additional layouts for each permeable layer to
evaluate the horizontalpotential.
3. Discharge stations: one water-table well to record bank storage and two five-point
piezometer sets to evaluate upward potential or base flow.
4. Basin outiet stations;two 6ve-poiiit piezo~etersets te ev~!mte underflow or leakage
past artificial stream-controlsections.
This must,of course,be modified according to the research programme and local condi-
tions.
FIG.4.49. Ground-waterinstrumentation (idealized researchLbasin).
147
4.4.2.3.1 ACCURACY
At the present state of knowledge of ground-waterflow systems,no absolutely accurate

analysis technique has been devised. Field measurements must be checked against water-
balance studies for the basin. The agreement reached will depend to a large extent on
the uniformity of the surficialmaterials,and the potentialgradient and tightness of divides
or aquiclude.Increased accuracy can be achieved,of course,by increasing the coverage
of the measuring points. The required water-balanceperiod will depend on the time of
adjustment of the flow system.
4.4.2.3.2 ADDITIONAL R E Q U I R E M E N T S
Installations similar to those outlined will be required to investigate special geological

conditionssuch as buried valleys,fracture zones and leaky divides.
4.5 Infiltration
Net precipitation (grossprecipitationminus interceptionloss) arriving at the land surface
becomes divided into two principal components-one involving surface storage and trans-
lation elements and the other subsurface storage and translation elements. These two
systems are separated by a clear physical boundary,but interactin complexways.Because
this interaction involves not only downward motion of water, but also the return of
subsurface water to the surface,infiltration cannot necessarily be considered the single
connecting process between the two major systems.In addition,the infiltration capacity
of a given soil is highly variable, being influencedby texture and structure,moisture
content,protection of the surface from rainfall impact, temperature and other factors
more difficult to evaluate.
Lack of reliable data on infiltration has hindered progress for applications in basin
hydrology. Methods of obtaining such data are analysis of rainfall and run-offdata (see
section 6.1.3.5)by infiltrometer (see below); laboratorymethods;and analysisof relations
between precipitationand ground-waterrise [219].Analysis methods,generally applicable
to small basins only because of heterogeneity of large basins,require costly installations
and long periods of maintenance between informativeevents.
Infiltrometerobservationsmay give a fairly good insight into differences in permeability
of the soil. They do not, however,give exact values of the hydraulic conductivity (see
section 6.1.4).Furthermore,they do not give an evaluation of the infiltration but only an
order of magnitude of the infiltration capacity of the soil.
4.5.1 InJiltrometers
Inñltrometersare frequently used to speed up the collection of data. Various researches

have tested and described numerousdesigns [171].The gravestrisks in using infiltrometers
are probably lateral flow and installation disturbances such as fracturing or compacting
of soils.Lateral flow is generally reduced by using wetted border strips and installation
disturbances by using larger test areas. There are two general types of infiltrometer in
use: the flood type and the rainfall-simulatortype. The various sizes and arrangements
in these two types give results which seldom agree in specific amounts measured and
occasionally differ in their relative ranking of soils tested. The larger sprinkler-type
infiltrometers with wetted borders give results more closely related to those obtained by
analyses of hydrographs from natural rainfall.
148
Methods o/ Observation and insrriimentution
4.5.1.1 Flood-type inñitrometers

This type of infiltrometer is more popular in irrigation than it is in basin hydrology.
Rainfall impact and the ensuing turbidity of the influent are not simulated in these tests.
The equipment is usually highly portable,cheap and convenient to operate and for this
reason is sometimes used in preliminary testing for general relativeranking of basin soils.
It is not safe to assume that results would indicate the order of magnitudes during a storm
event.
4.5.1.2 Ring itúìitrometers

Rings are used in these tests as a means of confining the surface area of applied water as
opposed to the tube or cylinder used to confine a column of soil. Rings are driven into
the soilonly a few centimetresand a hydraulichead is maintained insidethe ring by adding
measured volumes of water. Lateral flow is restricted by driving a larger ring,concentric
with the first, to the same depth and maintaining in it a similar head of water. Fifteen-
centimetre lengths of 22.8 c m diameter and 35.6 c m diameter tubing give a convenient-
sized unit. The diameters used, however,range in size and the method of application
varies according to the user. A constant head in the inner ring, supplied by a burette
read at intervals during the run,has been used [160]. Some researchers eliminated the
outer ring in favour of a bufferpond wherein an unmeasured head was maintained to
satisfy lateral flow.[85]In all these tests, the accumulation of infiltration is plotted and
differentiated with respect to time,to estimate the diminishingrate of infiltration capacity
for the soil tested.
In the U.S.S.R. infiltrometers of the PVN type used for loamy soils consist of two metal
rings bounding areas of 400 and 1,600cmz and of two feeding tanks (Mariotte's vessels)
with a volume of6 litres (Fig.4.50).The experiment is run until an approximately ,constant
infiltration value (fc)is reached.
Legend
i. Feeding reservoirs.
2. Graduated glass tube.
3. Frame.
4. Tube with air inlet tap.
5. Water outlet tube.
6. Inner ring.
I. Outer ring.
4.5.1.3 Cylinder or tube innltrometers
Tubes 15.25 c m in diameter and 38 or 51 c m long are jacked into the ground to confine
a column of soil against lateralflow when subjected to a smallhead of water.The gradu-
ated burettes shown in Figure 4.51 have been used to maintain a headinsidethe tube [160].
149
These burettes are read at regular intervals, as with the rings,to determine the rate of
infiltration. A perforated disc is usually placed on the surface to hold turbidity of the
surface water at a minimum.
FIG.4.51. Burettes and

perforated discs used in
drop infiltrometers.
4.5.1.4 Rainfaü-simulator inûitrometers

Rainfall is generally simulated in these units by drip screens or by nozzles. In the units
described in the literature,water is sprayed upon drip screens designed to form the desired
raindrops on yarn threads spaced in the screen for a suitable distributionpattern [8,168,
1851. The screen is mounted on the top of a tower about 6 m above the plot surface,to
enable the raindrops to reach terminal velocities (see Fig.4.52).
The type-Finfiltrometer used in the U.S.A. incorporatesnozzles with the spray directed
upwards and allowed to fall back on the plot [202].Nozzles are spaced,tilted and operated
at a standard pressure, to give acceptable distribution patterns on the plot. The height
of the spray is usually only 2-3m and terminal velocity is therefore not reached by the
drops in falling to the plot surface.The type-Fspray-headarrangement can accommodate
seven nozzles on each side for high intensities on 1.85 x 3.69m2rectangularplots, plus
a wetted border of 0.77m width. Spray heads can be rotated inwards or outwards to
refine the setting to a standard intensity. Alternate nozzles can be removed for lower
intensities and sheet-metalhoods can be pulled or pushed to cover or uncover all nozzles
150
[G.4.52. Tower housing a nozzle and screen for drop formations approximately 6m above plot surface [185].
simultaneously.The type-Finfiltrometer is housed in a collapsiblesheet-metalenclosure

about 3.5 m wide by 5 m long and 2.5m high to the eaves.
The larger,more complicated rainfall simulator shown in Figure 4.53was designed for
use in soil-erosionstudies[94].This device overcomessome of the disadvantagesofformer
rainfall simulators,which generally covered only small areas of soil and often failed to
produce many of the characteristics of natural rainfall.
Large spray nozzles are mounted on movable racks about 2.5m above the soil and spray
when moving in one direction. Rainfall intensities of 3.18,6.35and 12.73 cm/hr can be
obtained to simulate storms ranging from a light rain to a downpour. Impact energies
and the number and size of drops closely simulate a natural rainstorm.
This portable device is built in units that can be joined at the sides and ends to vary the
number and size of plots included in a test run. One unit covers an effective length of
about 5 m and each additional unit adds another 6.15 m to the plot length.The width of
individual units can be as much as 4.3m.
Small troughs are placed diagonally across the slope to measure the amount of water
applied to the plot and run-offis measured with a flume and stage recorder at the foot
of the test slope. One per cent of the run-offis collected at the flume and this is used to
calculate soil loss from the plot.
151
FIG.4.53. Demountable rainulator assembled for simultaneous operation on three plots [94].
The rotating boom shown in Figure 4.54is ready to apply water to a pair of mulched
plots in a study of highway-bankstabilizationmeasures.Collapsible booms and pneumat-
ic tyres make this a highly portable unit with a minimum of reassembly problems.
A smaller unit,the Purdue sprinklinginfìitroineter [14],uses an overhead nozzle direct-
ing a spray downwardsfrom 2 or 3 m to a plot 1.175 m square.Three nozzles are provided
to give rainfall intensities approximating to 6.35,8.25,or 11.43cm/hr as desired. Run-off
is measured cumulatively by means of a stage recorder in a collecting tank.
Sprinkler-typeinfiltrometers rely mainly upon preliminary and post-testcalibrations
of rainfall intensity. The plot surface is covered by some impervious sheet or pan and
the unit is operated for a given period at the recommended standard pressure to ensure
equilibrium in run-offbefore calibration is started.
In using sprinkling infiltrometers,conditions close to rainfall should be created. Both
the infiltration capacity and the surface detention depend on the rainfall intensity and
duration and on the initial soil-moisturestatus. Experimental plots should be bounded
by walls to prevent lateral surfaceflow and flumes with recording gauges should be used.
O n small sites (0.1-10 mz)sprinkling infiltrometers assure an even sprinkling of the
area with a given rainfall intensity and raindrop energy.In this case the rainfall intensity
may be changed during the experimentif required.
152
Meihods of observuriori and insirunierirurioti
FIS.4.54. Rotating-boom rainfall simulator [2121 for highway-bank stabilization studies,

U.S.A.
O n larger plots (up to 500 ni3) industrial sprinkling installations,which however fail
to assure an even sprinkling of the area and a constant rainfall intensity and raindrop
energy,are used.
.
The use of sprinkling installations for sprinkling plots with areas of 1-5m2is assumed
to be most effective and economical.
Figure 4.55 gives,as an example,the design of a sprinkling installationfor field exper-
iments of the State Hydrological Institute of the U.S.S.R.The installation consists of
four demountablesections with sprinkling areas equal to 1.5 m2each.
Experiments on the estimation of the infiltration capacity of plots with an area of 1 m2
and over must include regular soil moisture sampling up to a depth of 1.5-2m in order
to establish the initialand final moisture content and the depth of the wetting front below
the surface [66,67, 1951.
4.5.1.5 Portable rainfall-simulator infiìtrometers

A portable rainfall-simulator infiltrometer has been developed by the United States
Geological Survey [1531. The instrument measures infiltration of simulated rainfall in
I small plots of undisturbed soil and can be hand-carried to sites inaccessibleto vehicles,
The instrument is shown in Figures 4.56and 4.57.
Disturbance of surface soil by driven plot frames or cylinders has been eliminated in
this design.Moreover,water requirements have been reduced to 4.5or 9litres per meas-
urement;distilled water or controlled chemical solutionsmay therefore be used.
153
4.5.1.6 Operating sprinkler inMtrorneters

Procedures of operation are generally dictated by the types of analysis intended. If interest
extends only to the rate of infiltration after prolonged wetting, the sprinkled application
Legend 5. Head regulating reservoir. 10. Pump.

1. Sprinkler section. 6. Frame. 11. Hoses.
2. Netring. 7. Winch. 12. Outflow pipe.
3. Frame. 8. Edge of the operating plot. 13. Tap of rhe sprinkler
4. Supply reservoir. 9. Flume. section.
FIG.4.55. Sprinkling infiltrometer, U.S.S.R.
FIG.4.56. Portable infiltrometer in field use, Geological Survey, U.S.A.
154
Methods of observation and ìmtrumentation
continues until run-offbecomes constant.Thereafter,the difference between the rate of

sprinkling and the rate of run-offis equal to the rate of infiltration capacity (since no
further changesin storageoccur). Some investigatorsmake a second wet run the following
day,in an attempt to standardizeantecedent moisture effects for comparisonsof infiltra-
tion between sites.
Infiltrometersusing the larger plot areas give results more indicative of those derived
from analyses of hydrographsfrom natural rainfall.Interest then extends to the depletion
curve ofinfiltrationcapacity throughouttheperiod ofsprinkler application.Initialabstrac-
tions,including canopy interception,depression storage and, subsequently,the volumes
Reservoir and
.
.
control unit
y111
RA I N U L A TOR
Scale 1:4
Alternate-base
(separate splash
Aand run-off’
FIG.4.57. Schematic diagram of portable inñltrometer,Geological Survey, U.S.A.
155
of surface detention causing overland flow. must however be evaluated and deducted
from the rainfall run-offdifference in estimating the curve of infiltration.For this purpose
an analytical run followingthe wet run on the second day is made [161,202].The analytical
run is started at a timewhen the depression storage from the wet run becomes negligible,
but before any appreciable portion of the soil profile has drained,and is continued until
the rate of run-offagain becomes constant. Recession flows are observed in detail for
all three runs: initial,wet and analytical. Since no portion of the profile is allowed to
drain between runs, analyses (see section 6.1.3.5)are based upon the premise that the
rate of infiltration is a constant throughout the analytical run.
The bibliography [2,1711,gives a more complete resum;. Selection of the proper in-
filtrometershould be based upon the needs and resources of the researcher but,if results
are to be applied to natural rainfall,a rainfall simulatoris preferable.Although the larger
plot adds to the applicability of results,it also detracts from the portability of the unit
and poses greater problems of water supply for operations. The type-F infiltrometer
using a 1.85 x 3.69 m rectangular plot has been used [161]to evaluate the infiltration
capacities of various soil-vegetationcomplexes and to prepare isodiabrexal maps (maps
showing lines of equal infiltration capacity) on two basins in Illinois.These maps proved
very useful in designing networks of small plots for sampling the hydrological perform
ance of the two basins under study.
4.6 Measurement of phytomorphologicalcharacteristics

4.6.1 General
Vegetation measurements and descriptions are essential for the extrapolation of results
of hydrological studies. Vegetation-measurementmethods that have been used may be
grouped according to the following criteria: frequency of occurrence per species per plot,
number of individuals per unit of area, area covered per species per unit of area, and
weight or volume per speciesper unit of area [31].Areal coverage and weight or volume
are the most meaningful for hydrological studies. In addition to live vegetation, soil-
surface features such as quantities of mulch or litter,bare soil and rock should be meas-
ured.Areal cover or foliage volume is related to precipitation,interceptionand raindrop
impact energies.Soil-surfacefeatures affect interception,erosion and sediment yields.
In countries with a seasonal vegetative cover,phytomorphologicalcharacteristics bear
a direct relation with certain hydrologicalprocesses and interrelations between the water-
balance components.
Interception and evapotranspiration are the most important characteristics that are
related to phytomorphological characteristics, but development of foliage,grass cover
and agricultural crops also influencethe intensity of such processes as erosion,sedimenta-
tion,surface flow and infiltration.
4.6.2 Measurement methods of surface vegetation

Hundredsof methods have been used to measure vegetation.Many of these are presented
in publications [3,25,3 1, 36,81, 1801.A suggested samplingprocedure,based on some of
the information in these publications and on internationalexperience,is presented below.
4.6.2.1 Aerial photographs and field-sample plots

Aerial photographs are used in the office for preliminary identificationofplant communi-
ties. Plant-communityboundaries can be checked in the field and revised if necessary.
i 56
Sample plots should be located at random within the identifiable plant communitiesby
the use of grids and a table of random numbers (see section 4.1.1).This will result in
stratified random sampling within basins. After field data have been tabulated and ana-
lysed,the number of samplesneeded for a desired level of accuracy can be determined by
the use of formulas contained in standard statisticstextbooks.If sampling is inadequate,
additional field-sampleplots should be established. Field plots should be permanently
marked for future sampling.
4.6.2.2 Sampling herbaceous species and low shrubs

The point-quadratmethod [146]can be used to measure foliage cover of herbs and low
shrubs,For hydrological studies it is suggested that only the first hit as pins are pushed
towards the soil surface be recorded.From tabulated data this method permits calcula-
tion of percentage cover per species and percentage of soil surface occupied by mulch
or litter,rock,moss and bare soil.For calculation of foliage volumes,the heights of each
plant unit (foliage branch,stem or bunch) should be measured and recorded.For measure-
ment of possible vegetation changes over time,it is desirableto have permanentlyinstalled
metal stakes 30 m apart.Ifpins are spaced at 5 c m intervals,the first 15 m or 300pins per
transect should be sufficient.The point frame (Figs.4.58and 4.59)is placed successively
along a tape stretched between the two metal stakes.Measurements of herbs and decid-
uous shrubs should be made at the end of the growing season but before leaf fall.
FIG.4.58. A point-quadrat frame used to measure herbaceous cover. Pins are here spaced
2.5 cm apart. Alternate pins can be removed to give 5 cm spacing.
I 157
FIG.4.59. A large point frame used to measure shrub vegetation.Pins are spaced 7.5c m apart.
4.6.2.3 Sampling trees and iarge shrubs

The permanent 30 m plot suggested above m a y be used as one edge of a rectangular plot
for sampling woody species. For each woody plant the length of the tape covered by live
foliage should be recorded. For trees and large shrubs the depth of crown or foliage
should also be recorded and, for both trees and shrubs, heights should be recorded [99].
From intercepts and depths of foliage one can use the formula for the volume of a cylin-
der (V = nr2h) to estimate volumes of foliage per unit of area or per basin. An example
of a field data-collectionsheet is given in Figure 4.60and formulas for calculating foliage
volumes are given below.
Woody vegetation
Volume of foliage = nr2h.
Znr2 x crown depth = foliage volume per 100 m plot.
Foliage volume per 100 m f 100 = cubic metres of foliage per metre of line intercept =
foliage per hectare.
Herbaceous vegetation
Volume of foliage = Zpercentage cover x height of species x nr2.
Volume of foliage +- length of line in metres = cubic metres of foliage per metre of in-
tercept line = foliage per hectare.
Zvolume of foliage of woody plants + volume of foliage of herbaceous plants = volume
of foliage per hectare.
Volume of foliage per hectare x hectares per basin = volume of foliage per basin.
To extrapolate data from intensively studied basins to other areas, it is often desirable
to have counts of woody plants per unit of area. It is suggested that a plot 5 m wide be
158
M e t h o h of'observation and instrumentution
established adjacent to each 30 m transect for counts of woody plant numbers. For in-
tensive studies,heights of trees and shrubs and stem diameters of trees within each plot
should be measured.
4.6.3 Special observationsfor evaporimeters and lysimeters

Observations of evapotranspiration by evaporation pans with growing plants should be
accompaniedby a survey ofthe development ofplants in the pan and in the area surround-
ing the pan.
Phenological observations should be made visually when the pans are weighed and
when the monoliths are being replaced. Phenological observations include height,
Prqect: .................................... Observers: ...........................

S a m p l e wo.: .................... Location: ...................... Vegetation type: ...............
Exposure: ...................... S l o p e percent: ................. Elevation: .....................
Plor type and size: Transect 30m
30
= 3.2 cu.m per m
3.2 hectare-metres per h e c t a r e
Plot type and sire: Transect 15 m
0.096
15= 0.006 cu.m per rn
0.006 h e c t a r e - m e t r e s per hectare
FIG.4.60. Sample of field-datasheet and foliage-volumecalculations.Measurements are in

metres. The first column gives the first two letters of genus and species,e.g. Juut = Juniper
utahensis.
159
density, damage by climatological phenomena (hail, storm, frost, etc.), by insects and
other pests, diseases, etc.
The followingprincipal aspects of the growth ofagricultural crops should be surveyed :
seedlings,formationof leaves,formationof lateral shoots,flowering,ripening of seed,and
harvesting.
4.6.4 Root depth and root density

The root distributionis a very important factor in many hydrological studies,especially
in experimentalbasins.The root of a plant is that part which is normally below the surface
of the earth and its form depends on its shape and mode of branching; these in turn
depend on the type of plant, the soil structure and texture and the moisture character-
istics of the soil.
The primary function of the root is the absorption of water and inorganic salts in
solution and the conduction of these to the stem (see sections 4.2.4;5.3.4;6.1.2)and this
leads to soil-moisturestressees(see section6.1.4).Roots are agentsin increasing the poros-
ity of the soil and must be studied in conjunction with soil-physicalcharacteristics (see
section 4.7).
Root characteristicscan vitally affect water-balancefactors.In East Africa,for instance
the Kikuyu grass (Pennisetunz clundestinurn) roots to 6 m or more [loll and allows only
about an average of 12 c m of water beyond its root range out of about 340 c m (over five
years), despite a markedly seasonal rainfall distribution. Under maize (Zeanzuis),which
roots only to 2 m there is recharge of ground water nearly every year.
In N e w Zealand,pine trees (Pinus radiata), which normally have a shallow rooting
system,have developed rooting depths of an average of 5 m in coarse-grainedpumice
soils [12].
Root-depth and density measurements can be obtained by root-washingfrom a pit
face. Soil-moisturesampling techniques can give information about rooting depths in
dry seasons with a high evaporative demand.
Note that correlations of root depth and density with soil physical and hydrological
characteristicsare frequently difficult because of porosity in soil caused by decayed roots.
Frequent sampling may be necessary to obtain an insight into root development and root
decay.
For further details refer to the literature[32,1251.
4.7 Soil physical measurements

Soils differ greatly in their ability to accept,transmit or retain the water that reaches the
surface of the ground. A guide is given to those physical characteristics of soils which
may be relevant to the study of the hydrology of representative and experimental basins.
Standard texts [9,15,1481 should be consulted for fuller discussion of samplingand meas-
urement techniques.
There has been considerable progress in the last decade or two in the understanding
of the retention and movement of water in soil materials in terms of the interrelation of
conductivity,moisture content and potential energy.There is, however,a large gap be-
tween these fundamental studies and their application to the understanding of the role
ofsoilsin basin hydrology,although somerecent studies of the movement ofwater through
soils on slopes [95,130,2281 may indicate a closing of this gap. The number of measure-
ments required to characterize several horizons ofeach of the soil types present in a basin
requires the use of relatively simple and rapid methods to assess the properties,texture
and structure,which govern the interrelation of water content,conductivity and suction
160
Methods of observation und instrumentation
in the soil. Some definitions are given in section 1.5.Principles of subsurface flow and
unsaturated and saturated flow analysis are discussed in section 6.1.4and in the literature
Wl.
4.7.1 Texture and structure
The texture or size distributionof the primary solid particles may be assessed in the field
by the common method of moistening and manipulation [220]or by dispersion and frac-
tionation in the laboratory [54].There is sometimes a broad correlationbetween texture
and the hydrological behaviour of soils,but as a rule the arrangementof particles or soil
structure,must also be considered.
Soil structure has been defined [28]as ‘thephysical constitution of-asoil material as
expressed by the size,shape and arrangement of the solid particles and voids, including
both the primary particles to form compound particles and the compound particles
themselves’.The quantity,size and continuity of voids or pores are particularly important
in hydrology, since it is in them that water is stored or transmitted during the subsurface
phase of the hydrologicalcycle.The total porosity S t is the percentage of the bulk (macro-
scopic) volume not occupied by solids, and is given by:
t = 100 (i-D/d)
S (9)
where D,the dry bulk density,is defined as M /Vin which M is the mass of soil (oven dry)
in bulk moist volume V ;d,the particle density,is the ratio M /VP where VP is the collective
volume of soil particles excluding pore spaces between them.
Bulk density is a widely used value and is needed for converting water percentage by
weight to content by volume. Owing to swelling and shrinkage with changes of moisture
content,the moisture content at which the bulk volume Vis measured must be specified,
the volume at field-moisturecontent at the time of sampling commonly being used. Bulk
density is convenientlydetermined in stonelesssoils by carefulsampling of soilcores using
a thin-walled metal cylinder;in stony soils alternative methods, such as weighing the
soil removed from a hole of known volume,must be used [27].For sampling soils,special
drills and special cylinders with given volume are often used. In the U.S.S.R. the BP-50
with 100,250and 500 cm3cylindersis used.At depths exceeding 3 m soilsamplesare taken
from drill holes and are subsequently immersed in paraffin. Transmission or scattering
of gamma rays may be used as an in situ method. Details of these methods and their
limitations(see section 4.11) are given in the literature [15].
Particle density is commonly determined by liquid displacement using a pycnometer
[16].A value of 2.65g/cm3is commonly assumed for the particle density of mineral soils,
but lower values are obtained when the content of organic matter is significant.
While bulk density or total porosity may sometimes be adequate to characterize soil
conditions,it is often preferable to measure the pore-sizedistributionor the volume occu-
pied by the largerpores.The literaturegives methods that may be used for these measure-
ments,which are based on a capillary-tubemodel of the soil-porespace [227].In such a
model the suction,h, required to drain water from a capillary of radius r is given by the
capillary-riseequation:
h = 2y cos &/@gr (10)
where :
y =the surface tension of water;
a=the contact angle between water and the surface of the solid (often, but not
always, zero);
@=the density of water;
g =acceleration due to gravity.
161
3
Suction is conveniently measured in centimetres of water. When an initially saturated

soil is subjected to a suction h, the volume of water extracted is equal to the volume of
pores having an effective radius (air-entry value) greater than the value of Y obtained by
substituting the value of h in the capillary-riseequation. At 2WC,Y is equal to 0.15/h,
where both r and 12 are in centimetres.Pores draining at a suction of 50 or 60c m of water
[51,80,881 provide a useful index of soil structure and of the effects of cultural practices;
they are commonly termed large pores or air-spaceporosity. The curves obtained by
determining moisture contents of a soil sample at a series of suctions (the moisture-reten-
tion or moisture-characteristiccurve) have been used to derivepore-sizedistribution curves
[42,2111. Although the capillary-tube model is only an approximation of the actual
mechanisms by which water is held in soils,reasonable agreement has been found be-
tween pore-sizedistributionsobtained in this way and those obtained from microscopic
measurements on'soil samples in thin section.
The effects of cultural practices may be confined to quite shallow depths and special
samplingmethods may be needed to detect them.Profileexaminationby standard pedolo-
gical methods (see section 5.1.3), with particular attention to type and degree of structure
development,root distribution(see section4.6.4)and a visual assessment of porosity may
be a useful preliminary to sampling.
Methods for measuring water relations are not well developed at the present time.
Some theoretical and laboratory methods are described below.
4.7.2 Water relations

4.7.2.1 Moisture retention
Over the relevantrange of suction(O-15,000c m of water) a variety of methods is available
for determining moisture retention at a specified suction.Details of apparatus and tech-
niques commonly used are given in the literature [189].
The wide range of suction values that may be encountered on the field makes it con-
venient for some purposes to use the expression logdi,commonly termed pF. A suction
of 15,000c m of water (pF 4.2) is often used as an approximation of the condition (per-
manent wilting point) where water ceases to be available to plants. In soils that are drier
than this,water is too firmly held for the plant to be able to extract it. Water held at a
lesser suction is available for plants to absorb and transpire but, since the largest pores
may drain under gravity, water in them following infiltration is lost before the plants
are able to use it. Field capacity,the amount of water held in the soil after the excess
gravitationalwater has drained away and the rate of downward movement has materially
decreased,is commonly used to define the wet end of plant-availablewater. A method
for determining field capacity in situ is described in the literature[
1771.Laboratory meas-
urements of water retention at a suction of 330 c m of water (3 bar) have been used to pro-
vide an estimate of field capacity, but no particular suction can be said to correspond
accurately to field capacity [148,177,1891.Some limitationsof the field-capacityconcept
have also been described [229].One method of determining wilting point is to obtain the
soil moisture at the point at which the plant wilts irretrievably for the first time,despite
cool and moist conditions.Such analyses are normally carried out in the laboratory with
plants grown in special vessels.
Available water can be expressed as the difference between the water percentages at
the upper and lower limits,or if bulk density is known, the percentage values can be
converted to depth of water. Summing the available water in each layer or horizon gives
the total depth of availablewater in the profile or in the rooting range.It should be borne
in mind, however,that the accuracy of the available water content is dependent on the
reliability of the field capacity and the permanent wilting point to estimate the upper and
lower limits of water availability.
162
4.1.2.2 Moisture-retentioncapacity and saturation-moisturecapacity

Repeatable measurements of the quantity of moisture a soil can retain against gravity
may also be obtained by using the new,revised,standard method of testing for centrifuge
moisture equivalent of soils [4].This method requires that the temperature within the
centrifuge be controlled at 20°C f 1°C.Analysis of the data that necessitatedthis revised
method indicates that it is also necessary to maintain the relative humidity within the
centrifugeat or near 100per cent [184].The resultsobtained were duplicated on subsamples
of the same soils by others,using a standard subsoil centrifuge with a simple evaporative
cooler and humidifier installed in the drain hole [152].Centrifuge moisture equivalent
values,measured under conditions of controlled humidity and temperature,are referred
to as ‘moisture-retentioncapacity’values.
The moisture content of saturated soilsamples is also a water-relatedindex of the phys-
ical characteristics of soils [1901. It is, therefore,useful to differentiate soils. ‘Saturation-
moisture capacity’values provide a measure of soil porosity.
Using the bulk density (D)and specific weight (w) of soils, the value of the saturated
moisture capacity (Ps) may be calculated.Thisvalue corresponds to the maximum amount
of moisture which a soil may contain when all pores and voids are completely filled with
water.
D-w
ps=-. 100.
Dw
A n even closer approximation of the moisture-retentioncapacity of soil can be obtained

from saturation-moisturecapacity data.It has been found that a semi-logarithmicrela-
tionship occurs between saturation-moisturecapacity and moisture-retention capacity.
Such relationships can be determined with great accuracy when comparisons are made
between samples of similar geological origin. This is illustrated by the data in Figure 4.61
which were obtained from soils of generally similar origin.The correlation coefficient (r)
of 0.98,obtained from only 96 samples,indicates that a very close relationship exists.
Granitic and andesitic origin soils
[~Y = 69 log,o X -79.6 4
FIG.4.61. Relationship
between centrifuge-
moisture retention and
saturation-moisture
capacity of 96 soil
samples of similar
geological origin. Saturation-moisturecapacity
4.7.3 Watermovement
Direct measurement of the quantity of subsurface water in contrast to that of surface
water is not possible.Therefore,in studies of hydrologicalconditionsof subsurfacewater,
it is possible only to observe some collateral factors and it is the function of scientific
163
Representative und experimental basins
hydrological research to determine, on the basis of field experiments, the fundamental

laws that govern the changes in water movement and water resources, as well as the
quantitative values characterizing the subsurfacewater balance [219].
A n importantcharacteristicin subsurfacewater movement is the constantKin Darcy's
equation which is termed the hydraulic conductivity in saturated soils and capillary con-
ductivity in unsaturated soils (see sections 1.5 and 6.1.4).
In saturated soils,conductivity may be measured in the field by a number of auger-hole
and piezometer methods or in the laboratory on soil cores taken with a minimum of
disturbance [20,127, 1481.While measurements of the conductivity of saturated soilsmay
be useful, many of the processes of hydrologicalinterest occur in unsaturated soils.
The conductivity (K)decreases rapidly with water content,because movement of liquid
water is confined to the water-filled spaces. Attempts have been made to calculate the
conductivity of unsaturated soils (capillary conductivity) from'the pore-sizedistribution
obtained from the moisture-characteristiccurve relating water content to suction; the
results are in approximateagreement with those obtained by experimentalmethods [148].
The conductivity of unsaturated soil may be measured by steady-stateor unsteady-state
methods [128,1291.The techniques are,however,difficult and subjectto improvementin
many ways. The conductivity of unsaturated soils has also been obtained from field
studies of water movement using tensiometersto follow changes in suction 11661.
Curves relatingK to moisture content and suctionare used in the analysis ofinfiltration,
drainage and water transport towards roots of transpiring plants [69,178, 191, 1971.At
the present stage of development of the theory of water movement and of the measure-
ment of capillary conductivity,however,the full analysis is confined to relatively simple
cases (see section 6.1.4.3),It should by borne in mind that, owing to hysteresis effects,
differing relations of conductivity to water content and of water content to suction may
apply to wetting and draining processes.
4.7.4 Vapour versus liquid movement

When a moisture-stress gradient exists,water will move as a liquid through films or it
will flow through soil pores in response to hydraulic gradients. Water will move as a
vapour when a temperature gradient is impressed on a soil mass.A combined movement
as vapour and liquid is described in the literature [179].Soil moisture may move as a
vapour in one direction and as liquid-filmflow in the opposite direction,resulting in a
circulation that displaces soluble minerals in the direction of liquid flow. Where liquid
moisture is transformed to a vapour, dissolved salts are left behind and an accumulation
of soluble salts is deposited.At depths where moisture has migrated primarily as liquid,
cation exchange influences ion distribution more than solubility of salts. The net result
is a reduction in cation concentrationsin the direction of moisture flow.This process has
been called salt sieving [122].
When moisture moves through soils as a liquid,cationexchange processes cause depo-
sition of calcium salts with a release of sodium salts [156, 1571.The combined effect of
these processes results in a chemical imprint in the soils that can be analysed to define
the patterns of moisture movement in soils.
4.8 Measurement of soil frost and thaw

4.8.1 General
Observations of soil frost and thaw are required to study the influence of snowmelt on
run-off,on water losses by infiltration,on water accumulation and discharge in the soil
and on conditions of ground-waterrecharge.
1 64
Methods of observation and iristrunientation
Freezing and thawing of soils may improve soil tilth in agricultural lands but may also
increase erosion on poorly protected,steep slopes.Alternate freezing and thawing causes
a granulating action on soil clods [9]and moisture content of the soil plays an important
role in their effect. If the soil is dry during winter months, the effect of freezing on soil
structure is minimized. If rain occurs during the thaw cycle, dispersion of aggregated
material may occur [9].
Frost action on soils may cause either aggregation or dispersion, depending on the
nature of crystallizationof ice [120].Slow cooling produces aggregatesunder the influence
of ice-crystalpressure and dehydration. Rapid cooling produces dispersion, as the large
number of small crystals formed causes a breaking-up of aggregates. The freeze thaw
effects on soil erosion are accentuated by hillslope exposure, presence of swelling clays
and density of vegetative cover.
4.8.2 Measurement of soil frost and thaw
Observations of soil and subsoil frost and thaw should be carried out on one or several
sites,which differ by their vegetation or relief,Where a long frost period occurs,observa-
tions should start on the first day of negative mean air temperature and continue until
after the soil has completely thawed.
The following are suitable methods for the estimation of soilfrost and thaw:
1. The thermometric method, based on measurements of soil temperature at various
depths.
2. The physical-mechanicalmethod,based on an estimate of the changes in soil-physical
and mechanical properties due to a transition of the soil from the thawed to the
frozen state and vice versa (degree of soil-ice concrete,presence of ice crystals in
soil, etc.).
3. The electrometrical method, based on soil electrical resistancemeasurements.
4. The calorimetric method, based on measurements of the heat absorbed by the soil
while thawing.
The thermometric method requires observations of soil temperature at different depths
(in the U.S.S.R. observations are carried out at depths of 0.0, 0.2,0.4,0.6,0.8,1.2,1.6,
2.4 and 3.2m with soil thermometersenclosed in vinyl tubes placed in the soil). By inter-
polation of the readings the depth at which the zero temperature is observed, i.e.,the
boundary of soilîrost,may be calculated.
A cryopedometer,a special instrument for the estimation of soil frost or thaw depth,
is based on physical-mechanicalprinciples.The receiving part of the instrument is a rub-
ber tube with a centimetre calibration and has a diameter of about 1 cm. It is filled with
distilled water,closed at both ends and placed within a protective casing (pipe). This is
then dug into the ground to the maximum possible depth of soil frost.The height of the
column of ice indicates soil-frostdepth in the tube and the length of the thawed part of
the tube indicates the thaw.A piece of thread is run within the rubber tube;this freezes
within the ice and preventsmovement of the ice columnwhen the thaw starts.A cryopedo-
meter is permanently installed on one or more sites which are most typical of the given
basin.
Physical-mechanicalmethods include also measurements of soil-frost depth in open
pits.
The electrometrical and calorimetricmethods have a limited application and are used
only in experimental work.
Observations of soil frost and thaw vary in frequency according to the particular con-
ditions in each country and may need to be frequent when the variations in temperature
are very high.
165
4.8.3 Eflect of exposure on freeze-thaw cycles

Measurements of soil temperature and soil moisture at various depths show that the
exposure has a marked effect on micro-climateand plant growth.Quantities of soil mois-
ture availablefor plant growth vary irregularly over the year.In the western United States,
from April to June,the difference in soil moisture to a depth of 60 c m is not large;by
late July both north and south slopes are very dry,but the north-facingslopes have more
moisture available for plant growth.
The soiltemperatureon north-and south-facingslopesshowsimilartrends[84].Through-
out the year the soiltemperatures on the north-facingslopes at depths of 25 mm and 30 c m
are much lower than temperatures at the same depths on south-facingslopes.The greatest
differences occur during the winter months when north-facingslopes are generally frozen
for several months and soil on south-facingslopes is subjected to a freeze-thaw cycle
almost daily (see Fig.4.62).The marked differences in soil moisture and soil temperature
are reflected in plant communities and mass movement.
70
FIG.4.62. Graph of
soil temperature at a
depth of 25 mm below
surface on contrasting
slopes in mountains of hi
western United States. 18 December l'i63
Mass movement is apparent on both slopes but is greater on the north-facingslopes

because of the generally higher moisture content in the soil mantle. O n the south-facing
slopes the soil mantle is reworked and disturbed by more frequentfreeze-thawin winter.
4.9 Erosion and sedimentation

4.9.1 Scope und purpose
Erosion and sedimentation research projects that are undertaken in representative and
experimental basins should be directed to basic processes of hydrology and geoniorpholo-
gy.In the past,many studieshave been undertaken on small basins where severalhydro-
logical parameters have been measured and empirical equations derived to establish
exponential relationships between basin characteristicsand sedimentyield.The objectives
I66
Methods of observation mid instrumentation
of these studies have not always been met and, in many instances,have neglected the
cause-and-effectrelations so important to an understanding of basic processes. The pur-
pose of this section is therefore to define some of the basic processes of basin erosion
and to describe briefly some practical techniques for measurement of erosion and sedi-
ment transport, together with some applications of results.For definitions of terms, see
section 1.5.
The determination of sediment discharge has been the subject of much experimenta-
tion in laboratory flumes and in natural rivers and of instrument development work on
sampling equipment. Nevertheless, the measurement of sheet erosion, soil creep, and
rilling on hillslopes presents problems that have not been satisfactorily solved. in order
to arrive at a sediment budget for a small basin it is essential to determine the sediment
transport on hillslopes as well as in the stream channels [201].
4.9.2 Erosion studies

The typical experimental or representative basin is composed of a variety of hillslopes
bordering the stream channels and valley floors. The erosion on hillslopes, as referred
to here,is the quantity of soil removal between the basin divide and the junction of the
slope with the valley floor. Furthermore, the detachment of soil and transport of the
eroded material down the slope can be categorized according to the process.
4.9.2.1 Erosional processes

The quantity of sediment which is eroded from hillslopes and channels in a basin and
transported to a downstream measuring point is referred to as sediment yield. This
quantity is never the same as the total amount being eroded throughout the basin,
because of the many opportunities for redeposition between hillslopes and channels.
Therefore,the sedimentyield does not always answer questions of hillslope and channel
erosional processes.
4.9.2.1.1 HILLSLOPE EROSIONAL PROCESSES
Sheet erosion is initiated primarily by raindrop impact and the detached particles of
soil are transported by overland flow.The kinetic energy of rainfall has been expressed
by the following equation [62]:
E = KV4.33d1.07 i0.65 (12)
where :
E = the soil intercepted in splash samplers during a 30-minuteperiod;
V = the velocity of drops in feet per second;
d = the diameter of drops in millimetres;
i = the intensity of rainfall in inches per hour;
K = a constant.
Raindrop impact can detach large quantities of soil during a storm of high intensity but
other factors,such as vegetation and land slope,influence the transport of the detached
material.In certain areas,diurnalfreeze and thaw cycles may be a greater agent in causing
erosion than raindrop impact (see section 4.8.3).
A universal equation for predicting soil loss from sheet erosion has been developed
by the Agricultural Research Service in the U.S.A.12211.The equation is as follows:
A = RKLSCP (1 3)
167
Representaiive arid experimental basins
where :
A = the average annual soil loss in tons per hectare;
R = the rainfall factor;
K = a soil-erodibilityfactor;
LS = a slope-lengthand steepness factor;
C = a cropping and management factor;
P = the conservationpractice,such as terracing or contouring.
Mass movement or gravity erosion is often not as obvious as sheet or rill erosion, but
where slope conditions such as declivity, density of vegetation and soil moisture are
favourable,large quantities of material are moved downslope. If the mass movement
deposits are transported to the toe of a slope where stream action can undercut them,
the sediment yield may be high. In the south-western United States,measurements of
mass movement show that they constituteless than 1 per cent of the amount contributed
by sheet erosion [63]. In New Zealand,however,mass movement in its many types is a
major erosional process,contributing vast quantitiesof detritus to river systems by direct
movement into the river [37]. This detritus fills the river channel,widens it and enhances
its meander capacity,thereby further increasing the opportunity to cause erosion of toes
of slopes (see Fig. 4.63) [38].
FIG.4.63. Typical toe erosion in an aggrading stream (Ministry of Works, New Zealand).
I68
Methods of observation and Nisrrunieiztaiioii
4.9.2.1.2 EROSION STUDIES O N R U N - O F F P L O T S
Run-offplots which are used for overland run-offstudies can also be used for the study
of erosion.
Data on sediment discharge should characterize the natural conditions of a slope.
Therefore the plots should cover either the whole slope from the divide up to the edge
of the drainage network (large plots,up to several hectares) or the upper part of a slope
(small plots up to 1 ha).
Sediment-dischargemeasurements on plots are normally made by sampling at the flume
site.The number of samples required for a determination of the concentration depends
on the daily variations in it. When sediment discharge from the plots is high and when
the sediment discharge and flow vary considerably within a period of 24 hours,the water
should be sampled more frequently,for example, every one or two hours but not less
than four times per 24 hours. During storms the quantity of samples may be considerably
increased.
To establish relations between the sediment discharge and yield and the erosion on the
plot, micro-levelling of the plots should be done (say two or three times a year). For
this purpose permanent benchmarks should be installed alongside the plots.
With the above measurements of sediment and erosion, the following factors should
also be taken into account:the type and mechanical composition of the soil on the plot ;
the character of the surface (ploughed land,fallow land,virgin soil, etc.); the type of
vegetation and its distribution along the slope; the slope and aspect; the amount and
intensity of precipitation and flow;soil frost and thawing,etc.
4.9.2.1.3 S T R E A M - C H A N N E L EROSIONAL PROCESSES
When the sediment-watermixture from hillslopes reaches the stream channels,they not
only provide the conveyance for erosional debris from the uplands but contribute addi-
tional sediment through erosional processes. Channel-bankand channel-bederosion and
flood-plainscour are examples of the erosion caused by the concentrated flow of water.
4.9.2.2 Measurement of erosion

The measurement of erosion on hillslopes,channel banks,and channelbeds is considered
separatelyfrom measurement of sediment discharge in streams,primarily because of the
difference in the degree of refinement in techniques.Several relatively simple and inex-
pensive techniques have been developed in recent years for the measurement of changes
in channel configuration,channel aggradation and degradation,sheet erosion,and mass
movement of soils on hillslopes [63,1451.The principal advantages of these techniques
are that they require no specialized equipment and can be installed at sites having a
variety of physical environmentsand at a modest cost.
4.9.2.2.1 M E A S U R E M E N T TECHNIQUES
4.9.2.2.1.1 Hillslope erosion
A simple way to measure sheet and rill erosion on hillslopes, as it is used in several
countries and is here given as an example,is to take a 25 c m nail,slip it through a large
washer and drive it flush with the land surface.Subsequent erosion will undermine the
washer and let it slide down the nail a distance equal to the increment of erosion [63].
The nails and washers may be installed either in a specified grid pattern on a hillslope or
in transectson contour or downslope (see Fig.4.64).
Mass movement of soil down a hillslope can be measured in a variety of ways. It is
simplest to measure the downslope displacement of individual pins from a surveyed line
169
Representative and experinienfa1 basins
FIG.4.64. Nail-and-washertype of erosion pin on steep slope,showing erosion as indicated

by the nail protrusion.
30 m long with permanent benchmarks at each end [63].Steel pins,2-12mm in diameter

and 25 c m long,are placed at intervals of 1.5-3 m along the linei The pins are driven
approximately 20 c m into the ground. At the time of resurvey,the displacement of indi-
vidual pins can be measured.
A detailed survey of measurement methods of soil creep is reported in the literature
[200].
Other methods of measuring mass movement are with ‘Youngpits’ [230]or modifi-
cations of this method using thin aluminium strips or glass beads,0.50mm in diameter,
inserted into the soil with a hollow cylinder.This msthod produces a graphic picture of
the vertical profile of soil creep upon excavation of the pits.
4.9.2.2.1.2 Stream-channel erosion

The simplest way to monitor changes in stream-channelmorphology is by a surveyed
cross-section.It is absolutely essential that permanent benchmarks be established at each
end of the surveyed line. The types of datum that may be gathered from channel cross-
sections are aggradation and degradation of the bed,lateral shifting of the channel,and
changes in channel width. Resurveys may be made at yearly intervals (to observe long-
term changes) and several times a year after high flows(to observethe effects on the channel
of individual floods) [237].
Measurement of lateral migration of a stream channel or bank erosion can be accom-
plished by driving a series of steel pins horizontally into the bank. As the bank is eroded
more of the pin is exposed.This additional amount of pin exposure represents the bank
erosion between measurements [63].
Bed scour occurs in most stream channels during periods of high flow and refilling
occurs during flood recessions.A technique has been developed to measure the maximum
170
Mrthuds uf observation and insirumetirniion
amount of scour and subsequent fill that occurs as the flood wave recedes [63](see Fig.
4.65). A hole is dug in the channel bed and a length of chain is placed vertically in the
hole with a rock or some heavy object wired securely to the bottom of it. The hole is
then refilled.During high flow,when scour occurs,the chain is laid over in a horizontal
pojition to the depth of maximum scour.As the flow recedes,the horizontal portion of
the chain is measured from monumented survey lines and excavated after each flood.
Measurement of the increment of chain that has been laid over in a horizontal position
will give the mximuin scour.The depth of sediment on top of the horizontal chain rep-
resents the fill that has occurred.From these two measurements the net change in the
bed elevation may be determined.
FIG.4.65. Scour chain excavated in channel bed (depth of excavation represents the filling
occurring after scour).
4.9.3 Sedimentation studies

Studies of fluvial sediment transportation are generally concerned with the three modes
of transport: contact, saltation,and suspended [13].In most streams the three modes
occur simultaneouslyand determination of the fractionof the total load being transported
by each mode is often impossible.To simplify the discussion,the terms ‘suspendedload’
and ‘bed load’will be used. Bed load is the sediment that moves by saltation,rolling,
or sliding on or near the stream bed. Suspended load is sediment that is supported by
the upward components of turbulent currents and that stays in suspensionfor appreciable
lengths oftime [224].All particles that are small enough to be in suspension shift up and
down in the flow and presumably move readilyinto and out of the bed layer.
171
4.9.3.1 Techniques of sediment measurement

4.9.3.1.1 M E A S U R E M E N T O F SUSPENDED-SEDIMENT DISCHARGE
Ordinarily,suspended-sedimentconcentrations are based on laboratory analysis of the

water-sedimentmixture obtained with a suspended-sedimentsampler.Water dischargeis
computedfrom velocity measurements taken when the samples are selected.The suspend-
ed-sedimentdischarge per unit area at a point,or per unit width at a vertical,is computed
from the sediment concentration and stream discharge [224,231J.
4.9.3.1.1.1 Snnzpling equipment
There are more than sixty-fiveso-called suspended-sedimentsamplers described in the

literature (224).The six general types are: ordinary vertical pipe;instantaneousvertical;
instantaneous horizontal;bottle; pumping; and integrating.
Details of samplers now in use in the United States for measuring suspended sedimeni
are given in the literature [224].
Samplers should,in general, meet the following requirements: (a) the flow velocity
inside the orifice should correspond to that of the stream velocity;(b) to obtain a sample
of the variation in turbidity,the sampler should not be filled instantaneously,but grad-
ually and continuously;(c) the sampler should have a minimum disturbing effect on the
flow at the sampling point.
4.9.3.1.1.2 Sanpling procedure
Samples of suspended sediment are collected using two basic methods,depth-integration

sampling and point-integration sampling.Depth-integrationsampling is used in most
routine measurements of sediment discharge.The point-integrationmethod is used on
large rivers to define the vertical distribution of sediment.
Sampling by the depth-integrationmethod in streams less than 5 m deep is generally
done by lowering the sampler to the bottom of the stream and returningit to the surface
at a uniform rate. Most streams in experimental or representative basins will be in this
category.For deeper streams this procedure must be modified.
The advantage of depth-integrationsampling is that a single sample provides a dis-
charge-weightedconcentrationfor a complete sampling vertical [224].
Detailed sampling involves taking water samples from all verticals at five points,i.e.,
at the water surface,at 0.2,0.6and 0.8 of the depth and at the bottom. When the stream
is shallow,the samples should be taken at two points,i.e.,at 0.2and 0.8of the depth,or
at one point (at 0.6of the depth).
Streamflow should be measured simultaneously with suspended-sedimentdischarge.
The number of sediment-dischargegaugings should be sufficient to cover the entire range
of stream-dischargefluctuations.As a rule,20-40sediment discharges are gauged during
the year.
Nuclear instruments for measuring suspended sediment on a continuous basis are
under development (see section 4.i 1.1.4).
4.9.3.1.2 MEASUREMENT O F BED L O A D

4.9.3.1.2.1 Samplers
N o reliable devices for sampling bed load exist. All existing devices,when lowered to
the river bottom, disturb to some extent the natural regimen of sediment moving near
the bottom and do not give accurate results.
Although the samplers such as D o n (U.S.S.R.), Sphinx (Denmark) and VUV (Czecho-
slovakia) used for this purpose in various countries are amongst the best available,much
172
Methods of observation ami inscrimientation
doubt is cast on the accuracy of the sampling results. For sampling coarse bed load,
basket samplers and,for sampling sandy and gravel sediments,so-calledribbed samplers
are used (Fig. 4.66).The latter are suitable for bed load with a wide range of particle
size. Apparatus designed to trap and accumulate bed-load sediment must be calibrated
to determine its efficiency for any given condition [102].Efficiencies average about 45 per
cent for basket and pan samplers and about 65 per cent for pressure-differencesamplers.
For a complete discussion of apparatus and techniques for measuring bed load,refer
to the literature [102,2241.
In some cases it may be possible to divert the stream into a natural or artificial pond
or reservoir, or even a tidal basin. This offers the opportunity of estimating bed load or
total load by repeated surveys of the reservoir.
Nuclear techniques for measuring bed-load transport have been most successful (for
details,see section 4.11.1).
FIG.4.66. Don-type bed-sediment sampler.
4.9.3.1.2.2 Gauging methods

Bed load is measured for the purpose of calculating the annual sediment discharge;for
this reason, depending on the river regimen, bed load should be measured frequently
and preferably not less than ten times a year,includingat flood periods.
Streamflow measurements should be carried out simultaneously and samples should
be taken in all velocity verticals of the gauging section.The time for which the device
is left on the river bottom should be checked by means of a stopwatch and should be
such as to admit not less than 50 g bed load to the device. If the movement of the bed
load is very slow, the device should be left on the river bottom for not longer than i0
minutes. The sample of bed load should be dried.and weighed in the laboratory.By
dividing the sample weight by time for which the device has been left on the river bottom
I73
Representarive and experimental basins
and by the width of the receiver of the sampler,it is possible to obtain the bed-load dis-
charge per unit of river width at the gauging section.The bed-load discharge for the
entire cross-sectioncan be easily calculated from these data.
4.9.3.i .2.3 Bed-matzrial observations
Observations of bed load may include the sampling of bed material at several velocity
verticals ofthegauging sectionfor a determinationof the sizecomposition.In the U.S.S.R.,
sampling is done twice yearly (prior to the maximum flood and at its terminal phase).
In countries with rivers of complex-flowregimen,sampling times may be different.
For shallow rivers,where the bottom is composed of gravel and boulders,the charac-
teristics of the bed material can be estimated visually and expressed in per cent per
square metre of the bed area. For this purpose a device such as a steel frame (size 1 x
1 m), divided into 25 squares with 20 c m side,is placed on the river bottom at various
points along the cross-section.Photographic methods are also useful.
Depending on the character of the bed material, the following types of sampler may
be used: (a) devices for sampling bed material with a disturbed structure;(b) devices for
sampling bed material with an undisturbed structure.The first type is useful for sampling
gravel and the second for samplingfine gravelled,sandy and silty material.
4.9.3.1.3 DETERMINATION OF C O N C E N T R A T I O N A N D PARTICLE SIZE
The determination of concentration and particle-sizedistribution from suspended-sedi-

ment samples offers numerous problems. Samples may contain only minute quantities
of sediment of very small particle size;or the sample may contain such large quantities
of sediment that special procedures are necessary [83]. The two most commonly used
procedures for determining concentration are the evaporation method and the filtration
method. The filtration method may be somewhat faster than the evaporation method
for some samples,but it imposes a severe limitation on the quantity ofsediment that can
be analysed.
4.9.3.1.3.1 Filtration method
The filtrationmethod for the determination of suspended-sedimentconcentrationutilizes

a Gooch crucible in conjunction with various types of filter material. The Gooch crucible
is a small porcelain cup of approximately 25 cm3 capacity with a perforated bottom.
It is easily adapted to an aspirator system and vacuum filtration.Asbestos is generally
satisfactory as the filtering agent.For details of the evaporation and filtration methods,
refer to the literature [82].
4.9.3.1.3.2 Methods .for determining parficle-size distribution
Particle-sizedata has many uses. The size of suspended and bed material is importam
in the computation of total sediment discharge.The frequency and kind of particle-size
analysis should be adequate to describe the pertinent characteristics of sediment particles
so that satisfactory comparisons can be made between sediments collected at different
places or from the same place at different times [82].The three most commonly used
methods are: the sieve-pipettemethod; the visual-accumulation-tubepipette method;
and the bottom-withdrawal-tubeand visual-accumulation-tubemethod.
4.9.3.1.4 SEDIMENTATION SURVEYS IN S M A L L RESERVOIRS
Several different methods have beer? used for sedimentation surveys of small reservoirs.
A field determination of the emergency spillway contour and a selected lower contour,
174
together with data from several cross-sections or ranges, preferably parallel, provide
adequateinformation.A stage-areacurve is the most direct way to determine the capacity
of a reservoir and the volume of sediment deposits can best be obtained from capacity
differences.A detailed description of this survey method and computationalprocedures
may be found in the literature [224].
4.10 Measurement of water quality

4.10.1 Purpose of water-quality measurements
The dissolved and suspended materials carried by natural waters are characteristic of
the soils and rocks making up the basin over which the water flows or through which it
percolates.
The physiography of the basin affects not only the discharge and therefore the time
of contact of the water with soluble matter, but also its carrying capacity for suspended
matter.
The collection of information on the organic and chemicalcontent of flow from repre-
sentative and experimental basins permits therefore:(a) the establishment of benchmark
conditions for use in assessing the effects of natural and cultural changes;(b) determina-
tion of the flow contribution (this is frequently more economical than establishing gaug-
ing stations); (c) the analysis of long-termtrends in ground cover;(d) the development
of criteria aimed at economical use of the available waters.
In view of the usefulness of such data,consideration should be given to the necessary
instrumentation and available laboratory analysis facilitiesduring the planning stage of
representativeand experimental basins.
4.10.2 Frequency of measurements

As flow is a dynamic process on all basins, the frequency of sampling must be tailored
to take into account the changes in the flow characteristics of the station cross-section
(i.e.,the stage of the discharge hydrograph).
Water sampling should be carried out at the following stages;during the rise, at the
crest and during the recession of two or three floods and, in addition, once or twice
during low-flow.O n rivers with spring snowmeltfloods,samplingshould be carried out at
the beginning of the fiood rise, at the crest,during the recession,during summer low-flow,
at the crestsof one or two summer and autumn floodsand during winter low-flowperiods.
Continuous sampling,unless limited to conductivity or turbidity measurements, should
not be undertaken unless for a specific research project where facilities are provided for
analysis of the enormous amount of data which is collected.
Systematic sampling and synoptic sampling are desirable (the latter to highlight major
contributingareas before a programme of culturalchange is finalized). Regardless of the
objectives of the basin study,the advice of specialists should be sought on the frequency
of sampling required.
4.10.3 Location of sampling

The equipment involved in sampling water quality requires that the stations chosen
should be reasonably accessible and, in temperate zones,should preferably be serviced
with electrical power to facilitate winter heating and operation of electrical water-quality
equipment. To ensure that samples are representative of the chosen cross-section,and
thatuniformmixingis achieved,stationsshouldbelocated a sufficientdistancedownstream
175
Representative ami experimental basins
of tributaries and major controls. The minimum programme for representative basins
should include a quality station near the main gauging station.Experimental basins are
likely to require supplemental stations for help in isolating the effects of cultural change
on subsidiary streams and run-offplots.
Water-qualityprogrammes are aimed at the development of procedures for predicting
future changes in the chemical and physical properties of water so that these proven
methods may then be applied to other basins and regions and this should be borne in
mind during the planning stage of all basin studies.
4.10.4 Sampling methods

For water sampling,thoroughly cleansed glass bottles (1 litre) should be used. The bottle
is rinsed several times with water from the stream to be sampled and is then filled with
this water,corked and labelled with the name of the stream and the sampling date. As
it is importantto avoid contamination,all equipment should be clean.
4.10.5 Chemical and physical analysis

At the present stage of this science,manual sampling for research basins is likely to be
the most economical and satisfactory.Equipment is available for either point studies or
for section surveys,which integrate the content of samples taken at a series of points.
Quality data are influenced by temperature,biological action,oxidation and reduction
processes and the action of sunlight.Quality measurements may therefore require a high
sampling frequency.111 addition,samples aimed at determining the content of carbon
dioxide,hydrogen sulphide or dissolved oxygen, etc., will require field analysis. Field
analyses of waters with a high bacterial count are necessary to offset the possibilities of
freezing or large temperature increases during shipping to laboratories.
Regardless of the specific use of the sample,its date, time and temperature should be
recorded to assist in determining the effect of storage time.
Considering the natural changes expected in representative basins and the changes
planned for experimental basins,the following benchmark data should be collected :Ca++,
Mg++, Na+,Soe,COZ, Cl-,Noar,Sioz,K+and HCOB-concentrations,pH, colour and
turbidity.These properties can be determined from a half-litresample.If F’, Po4”‘,Mn++,
Fe++and Fe+++traces are also to be mrasured,an additional small sample will be needed.
The choice of element sampled depends on the purposes of the research.
As some of the above properties do not undergo rapid change,a reassessment of the
sampling programme should be made annually.For detailed methods of sampling and
laboratory analysis,refer to standard textbooks [i41].
For details of data processing of water-qualitydata,see section 5.7.For analysis and,
interpretation,see section 6.1.6.
4.11 Use of radioactive tracers

Techniques involving artificial reactor-producedradioisotopesin hydrological studies on
representative and experimental basins are available for streamflow, sediment, snow,
soil-moistureand ground-waterstudies and the relative success gained by some of these
methods indicatesa rapid development of such techniques.
Moreover, useful results are being obtained by the application of methods based on
the isotopiccompositionof natural waters to determine the source,mixing,rate of move-
ment and past history in the hydrologicalcycle of the isotopes.
176
Merhods of' observation and ìnstrurne~itariorz
4.11.1 Sediment transport measurements (see section 4.9)

Radioisotopes can be used in measuring bed-load material, to label the natural sediment
or a simulatorsuch as glass beads.There are at present two methods available,as follows.
4.11.1.1 Space-integration method

This method consists of injecting labelled material at a given section of the river and
following its movement either by lowering detectors to, or collecting samples from, the
river bed for subsequent measurement of radioactivity.It is then possible to draw iso-
activity curves for different periods and to calculate the rate of travel of the centroid of
the radioactive material. This gives the mean velocity of the bed material having simila
size distributionto that of the labelled material.
4.11.1.2 Time-integrationmethod
This method consists of monitoring the radioactivityin a section of the river downstream
of the injection section.The time at which the centroid of the activity curve is observed
gives the average time of travel from the injection cross-sectionto the sampling cross-
section.Detection is accomplished either by direct measurement of the radioactivity in
the stream bed or by collection of samples of bed material during the experiment for
subsequent measurement.It can be seen that,in both methods,only the velocity of the bed
material is obtained.To estimate the bed-materialdischarge,an estimate of the depth of
the moving layer must be made. If undisturbed samples are collected from the river bed
it is possible to estimate the depth of the moving layer.If the activities are measured by
lowering a detector to the stream bed,estimationof the depth of the moving layer becomes
difficult.
4.11.1.3 Equipment required

A variety of radioisotopes (such as gold-198,chromium-51 and rubidium-86) can be
used for bed-loadmovement studies. The half-livesof the radioisotopesused should be
relatively short,preferably only a few days or weeks. Scintillation counters are generally
used for the detection of the radioactivity. Specially designed counters are now available
for sedimentation work [104,106,107,11 1, 1121.
4.11.1.4 Suspended-sediment measurements

The use of nuclear instrumentsfor measuring suspended-sedimentconcentrationin rivers
is a promising technique in the development stage,because it dispenses with the labora-
tory analysesrequired in conventional measurements. Thus it is possible to increase the
number of points sampled.This method uses a radioactive source-detectorsystem and
observes the variation of the intensity of the radiation,which can be related to the sus-
pended-sedimentconcentration in the volume between the source and the detector. The
radioactive sources used are cadmium-109 and americium-241 [log, 1il].
4.11.2 Measurement of the water content of the snow pa'ck

(see section 4.2.2)
The measuring system consists of a collimated radioactive source at the surface of the soil
and a detector abovethe snow surface (see Fig.4.67).The radiation received by the detec-
tor is a function of the water content of the snow cover.This design can be reversed by
177
having the radioactive source above the snow surface and the detector at the soil surface.
Jt is possible to make a continuous record of the water equivalent of the snow cover and
to transmit the information by a telecommunicationsystem.
Geiger-Müllercounters can be used for the detectionof the radioactivity and the source
most commonly used is cobalt-60.This method is particularly suitable for basins with
poor access where an index of the snow pack is to be obtained.
A device which measures the water equivalent of the snow cover and which records
the diminution of natural radiation of the earth under it is being installed in the U.S.S.R.
O n a selected course before and after snowfall,impulses of a medium radioactive back-
ground of the earth activate a telemetering device. The measurements may be made by
terrestrial or aerial survey [235].
I
l I
I
' PROTECTIVE CASING
'IVE
FIG.4.61. Radioactive
snow gauge installed in
Kvilde in 1967 [149].
4.11.3 Soil-moisture and density measurements

Neutron moisture gauges are now widely used for soil-moisturedeterminations (see sec-
tion 4.4.1.1). The method is based on measuring the deceleration of neutrons emitted
from a fast neutron source in the soil.This deceleration is due to the collision of neutrons
with light elements such as hydrogen (which is essentially contained in the molecules of
water present as soil moisture). G a m m a gauges based on the back-scatter principle are
used for the computation of the dry bulk density (see section 4.7)and are useful as a
supplement to the neutron method. When only the change in moisture with time is of
interest,differences in the dry bulk density throughout the profile generally cause only
minor errors in the determination of moisture content (because different dry bulk den-
sities in the same soil cause changes in the calibration curve which can be considered as
parallel shifts as a first approximation [109,110,2341.
4.11.4 Ground-water tracing, velocity and direction

measurements, determination of the efective porosity
The use of radioisotopesin ground-watertracing (see section 4.4.2) can be considered
as a natural follow-up of the organic and chemical tracers. The main advantage of
I78
Methods of' observation and instrumentation
radioactive tracers over conventional ones is the high sensitivity of detection and hence
measurement in both field and in the laboratory. The amount of tracer required is
thus considerably reduced though the cost remains high. A tritium concentration as
low as 1 TU1 can be measured with suitable counting equipment. Tritium has also the
advantage that it is not subject to adsorption, as the tritium atom is incorporated in
the water molicule.
4.11.4.1 Simple point-to-pointtracing experiments for the determination of

the hydraulic continuity and direction and velocity of ground water
A tracer pulse is usually introduced at a water point up the hydraulic gradient and its
appearance is monitored at points situated down-gradient.The simple appearance of
the tracer is a definite qualitative proof of hydraulic interconnexion.If the appearance
of the tracer is monitored at severalpoints down-gradientit is also possible to determine
the directionof the ground-waterflow,as the maximum concentrationwould be observed
in the direction of flow.The determination of the actual ground-watervelocity using the
simple technique outlined above is straightforward.In this case,the distance between the
injection and the sampling points, divided by the time which elapses between the
injection and the observationof the centre of gravity of the activity curve in the sampling
point, gives the actual velocity of the ground-water between the two points, assuming
that they are located in the direction of the ground-waterflow.If the effective porosity
and the hydraulicgradient are known,the permeabilityof the sedimentscan be calculated.
4.11.4.2 Determination of the effective porosity and dilution techniques for the
determination of the direction and velocity of ground water
In addition to the point-to-pointtracing tests,radioisotopescan also be used in combi-
nation with conventional aquifer tests to increase the information obtained concerning
the flow and aquifer characteristics such as effective porosity [109].A n example of the
dilution technique is the so-called single-wellmethod. The tracer is introduced into a
perforated section of a borehole and its vertical movement is prevented by pneumatic
seals. The decrease of the activity in the section is monitored by a collimated or non-
collimated probe. It can be shown that the rate of decrease of activity in the perforated
borehole section through which ground water is allowed to flow is proportional to the
ground-watervelocity.By using a collimated detector to measure adsorption of the tracer
on the walls of the borehole,it is possible to determine the direction of the ground-water
flow,as the adsorption occurs mostly in the direction of the flow. This method of tracer
dilution may also be applied to larger water bodies such as small lakes and ponds which
are in hydraulic connexion with an aquifer.
The single-welltechnique is, however, in a development stage and serious difficulties
exist in the interpretation of the data obtained,owing to the existence of vertical flows
in boreholes [106,109,110,1121.
4.11.5 Environmental isotope techniques

The study of the isotopiccompositionof naturalwaters has now become possible because
of the development during recent years of sensitive analytical techniques. Typical en-
vironmental isotopes used in hydrological studies are tritium (3Hor T),deuterium (ZH
or D),oxygen-18 (leo) and carbon-14(14C). They are produced by natural and/or
1. A tritium unit (TU)is a concentration of 1 tritium atom in 1Ole hydrogen atoms.
179
Representarive and experiniental basins
artificial means, but are invariably distributed by natural phenomena and follow closely
the hydrologicalcycle.Time and space variations of the so-called‘isotopiccomposition’
of water are governed only by physical phenomena (i.e.,spontaneous radioactivedisinte-
gration,atmospheric circulation,temperature and evaporation,etc.) except for those of
carbon-14which are also influenced by chemical reactions. For tritium,deuterium and
oxygen-18,however, this problem is negligible and constitutes the definite advantage of
these natural tracers over non-isotopictracers and over carbon-14[106,108, 1121.
4.11.5.1 Deuterium and oxygen-18

The most abundant isotopic species of water occur in natural waters in approximately
the following ratios:
HP0 : HDleO :HPO = 997,680:320 : 2,000.
The deuterium and oxygen-18contents ô of natural waters are best analysed by mass
spectrometry and are generally expressed in deviations per thousand from an arbitrary
standard,Standard Mean Ocean Water (SMOW). For definitions, see section 1.5 [48].
Csample-CSMOW
ô= x 1,000. (14)
CSMOW
The concentrationsof these stable isotopes in natural waters have a range which is about
200 times greater than the accuracy of measurement; there is ample room therefore to
look for variations and differences and for their interpretation.
The particular interest of deuterium and oxygen-18to the hydrologist lies in the fact
that,although they follow closely the hydrologicalcycle and are not subjectto preferential
adsorption and biological uptake, a change of isotopic composition occurs whenever
there is a change of state such as evaporation or condensation.The liquid phase is thereby
enriched in the heavy isotopic species owing to the lower vapour pressure of water mole-
cules containing deuterium or oxygen-18.This process is called the isotopicfractionation
of water and the ratio of the vapour pressure of the isotopically lighter component
(HPO)to the vapour pressure ofthe heavier component(HD160 or HPO)is the so-called
fractionation factor of the isotopic species concerned.Under normal temperature and
equilibrium conditions,the values of the fractionation factor are 1.OSO and 1.O09 for
deuterium and oxygen-18respectively.In other words, the ratios of the deuterium and
oxygen-18 concentrations of the liquid and gas phases are 1.080 and 1.009,provided
that equilibrium conditions exist (i.e.,that the liquid phase is overlain or surrounded by
saturated water vapour) and that enough time is availablefor molecular exchange [49,521.
Although these phenomena are complex and not yet fully understood,their consequen-
ces in natural waters may be labelled quite simply :the precipitationbecomes isotopically
lighter (i.e.,depleted in heavy isotopic species) with increasing latitude and altitude and
during the colder seasons (Fig.4.68)and there is a significantpositive correlationbetween
the isotopic concentrationsand atmospheric temperatures (Fig.4.69). Evaporating water
bodies such as lakes and reservoirs are,on the other hand,isotopically heavier than the
precipitation,run-offand ground water in the same region (Fig.4.70).
The cyclic seasonalvariations of stable isotopeconcentrationsin precipitationare con-
siderably smoothed during the accumulation and melting of water in the form of snow
and during its movement in the unsaturated zone, so that ground-water samples are
found to exhibit no large time variations with respect to deuterium and oxygen-18 con-
tent. This makes the comparison of different samples more significant and facilitates
identification of the sources of surface and ground waters,estimation of the altitude of
180
Methods of observation and instrrunzentation
recharge areas and determination of connexions between different water bodies such as
lakes,rivers and ground water [172].
4.11.5.2 Tritium and carbon-14

Tritium and carbon-14are radioactive and both are produced in the upper atmosphere
by the reaction of cosmic rays with nitrogen.The atmospheric content of both has in-
creased greatly during recentyears,owing to atmospherictestingof thermonucleardevices,
and has been steadily decreasing since the last test series.Although this transient condi-
tion created by man has greatly disturbed the equilibrium conditions in nature and made
radioactive dating difficult,it has also given to hydrologists the opportunity of studying
O O
Gough Is.-+
6 ‘O = 0.695 ta - 13.6%.
-5 Marion Is. -+ -5
Valentia
(S D = 5.6 to -loo%=) Thorshavn -+
-1 o
Isfiord Svalbard + h CoDenhaqen
. - -10
Tvärminne
Gronnedal 610 N UI tuna
Angrnmgssolik 660N
Scoresbysund 70°NA -1 5
-15
-20 Little America 780 5 (300)A / -20
Station Nord 820N
-25 -25
South Greenland
-3c -30
North Greenland
-31 -35
-4( .40
-
x
-4l -45
O
2L4 -5( -50
-50 -40 -30 -20 -1 o o IO 20 oc
M e a n annual air temperature
FIG.4.68. Relation of mean air temperature to oxygen-18concentration in precipitation [52].

Figures in parentheses indicate the total thickness (cm) of the investigated snow layers.
181
Representative and expcvirnental basins
the tritium pulse introduced to the river and ground-watersystems and to surface reser-
voirs [53, 112,167).
The half-lifeof tritium is 12.26years. With the precision of available analytical tech-
niques it will be possible to trace recent tritium pulses in ground water for severaldecades.
Tritium concentrations in precipitation are now being measured in some countries and
in a number of stations included in the I A E A / W M O world-widenetwork and it is possi-
ble to estimatethe inputfunction,provided that enough data is avaliableon climatological
and hydrologicalfactorssuch as precipitation,run-off,evapotranspiration,etc. (Fig.4.71).
Some of the major rivers of the world were recently included in this network.Data are
published in the literature-[l05].
FIG.4.69. Monthly 1
oxygen-18and deuterium
concentration in Genoa
precipitation.
8 '80(in %o)
-17 -16 -15 -14 -13 -12 -11 -10 -9 -E -7 -6 -5 -6 -3 -2 -1 b +I 4
-70-
-80-
-90.
Legend -100.
o Lake samples.
A Ground-water samples.
x Monthly precipitation.
FIG.4.70. 8D-81B0
relation, Beysehir-konya
region,Southern Turkey.
Tritium gives unequivocal evidence of modern recharge to ground-watersystems. It

should be emphasized,however,that dating liquid water is basically different from dating
of solid material by radioactive decay. This is due to variations of tritium concentration
caused by phenomena other than spontaneous radioactive decay,i.e.,mixing of waters
having different histories,diffusion,hydraulic dispersion,etc.Time-samplingis therefore
of primary importance in tritium studiesto separatethe variations due to differentcauses.
182
Methods of'observation and instrumentation
Under favourable conditions an average age can be estimated,depending on agreement

of the model and hypotheses adopted with the prevailing hydrological conditions [106,
108,1121.
Carbon-14 has a half-life of 5,568 years and is most useful in determining slow rates
of flow in confined aquifers (the most common case), where its dating capacity ranges
up to about 30,000 years. Addition of dead carbon from rocks through solution and
exchange has a diluting effect on carbon-14content and must be taken into account in
correcting apparent ages obtained from carbon-14analyses.
104
103
Legend
1. Vienna.
2. Genoa.
3. Midway Islands.
4.Gough Island.
102
FIG.4.71. Tritium
concentration at selected
stations [105]. IO
4.11.5.3 General
The isotopic methods described above are additional tools for the hydrologist.These
methods requireprecise laboratorywork,and samplescollected in the field must be shipped
to a laboratory.The size of the sample required is not,however,prohibitive;tritium and
carbon-14,for example,require samples of about 1 litre,although for carbon-I4prelimi-
nary field concentration of a larger sample is required. For deuterium and oxygen-18,
only about 20 ml of water are required.
Isotopicanalysis in many cases provides unambiguous answers to questions that can-
not be otherwise resolved.Its greatest value is, however,as a supplement to conventional
hydrologicalwork and its applicationrequires an adequate knowledge of the hydrological
and climatic conditionsin the region studied.
4.12 Minimum equipment

The complete set of instruments and equipment described in this chapter is applicable
only if detailed studies are undertaken. Perhaps only countries with long-termexperience
in the organization of hydrologicalresearch on a well-equippednetwork of experimental
basins could carry out such extensive investigations.Countries which have only just
begun to organize hydrological networks and research are probably unable to carry out
such detailed programmes. For this reason, the minimum equipment required for ele-
mentary observations on representativebasins is given below.
Minimum equipment,naturally,depends on the programme of research.If a minimum
programme,as given for instance in section 3.1.1is used, this includes observations on
precipitation,flow and snow and the minimum equipment should include instruments
necessary for continuous measurement of precipitation (rain and snow), flow, and air
temperature (if snow plays an importantrole in stream flow).
The number of points for rainfall and snowfall measurements should be sufficient to
provide an accurate estimate of both the mean basin precipitation and the precipitation
I83
Representative arid experimental basins
intensity for individual storms. The water equivalent of the snow pack needs to be esti-
mated only at the time of its maximum accumulation before the spring snowmelt.
The density of precipitation gauges depends on the drainage area and on the required
accuracy of the precipitation estimation (see section 4.2.1).
Flow should be measured on a gauging station that has adequate facilitiesfor obtain-
ing maximum and minimum flows. This means that it is necessary to install a complete
flow-measuringstation on each representative and experimental basin. Such a station
should include:(a) a water-levelrecorder; (b) a staff gauge and three benchmarks;(c)
a flow-measuringstructure (cableway,footbridge); (d) an artificial control on representa-
tive basins if no natural control is available;(e) a precalibrated measuring structure on
experimentaland small representative basins (see section 4.3.4).
Apart from observations of precipitation and flow,it is greatly desirable,if reservoirs
or lakes are a feature of a hydrological region,to measure evaporation from water sur-
faces by means of evaporation pans. O n all basins it is desirable to measure suspended-
sediment discharge; this does not require great expenditure.
The minimum equipment should also include a simple climatological station for air-
temperature and wind-velocityobservations.
TABLE
4.3. Minimum equipment necessary for basic observations
Items observed Instruments and equipment required Minimum number
On a representative basin
Precipitation Precipitation gauges 3-5
Recording raingauges 1-2
Snow cover Snow-samplingequipment 2-3

Portable snow stakes 2-3
Flow Complete gauging 1 for every discharge site

station (staff gauge,
benchmarks, etc.)
Current-meterequipment
Water-levelrecorders 1 for every discharge site
Weirs 1 for every discharge site
on uncontrolled streams
Evaporation from water Evaporation pan 1 set (with a raingauge)
surface
Suspended-sediment Bottle sampling equipment
discharge Simple equipment for
analysing sediment
On climate station
Air temperature Thermometers 1 set

(minimum,maximum
and psychometric) in a
thermometer screen
Wind velocity Anemometer 1
184
Methods of observation arid iristrrinientation
-- Water divide. y Discharge site. coniferous forest.

>River, brook.
z
t/ Weir. 191 Deciduous forest.
/>-Intermittent stream. H Gauging footbridge. Mixed forest.
Y Precipitation gauge. W Flume and weir. Burned-out forest.
'I Totalizer. O Run-off plot. a Forest cut.
O Recording precipitation gauge. 8 Observation well. a

,. .-
II..
Thin forest.
@ Radio precipitation gauge. @ Observation well and measuring ;o0 0 ; Strip of forest.
.........
x-4 Snow course.
point of physical properties
of soil.
a I
Bushes.
e Triangular snow course.

A Measuring point of physical
Dwarf forest.
0 Snow=covermeasurement plot. properties of soil. @Garden, orchard.
? Meteorological station. U Soil-moisturemeasuring site a Vineyard.
IIII] Water-surface evaporation site. A Soil-frostobservation site, Meadow.
I Land evaporation site. A Cryopedometer. a Pasture.
B Staff gauge. Lysimeter site. 0 Ploughland.
!Water-level recorder. I_I Swamp.

Sands.
a Stone talus.
a Glaciers and firn
fields.
1 FIG.4.72. Example of the location of hydrological equipment on a hypothetical representative

I basin in the U.S.S.R.
185
The list of minimum equipment necessary for basic observations of a representative

basin (up to 50-100km2) is given in Table 4.3.For experimental basins,where observa-
tions are carried out according to individual programmes,it is impossible to compose
such a list.
4.13 Location of instruments and equipment in a basin

The location of instruments and equipment in any one basin is an important aspect of
research.The location depends in the first instance on the use of sound sampling techni-
ques (see section4.1.1).The selectionof points is also associatedwith the relativecomplex-
ity of the physical features of any one basin and for this reason it is impossible to give
general recommendations.As an example,Figure 4.72showsthe locationof instrumental
equipment in a hypothetical representative basin in the U.S.S.R.
References
1. ALTOVSKY, A. A. (ed.). 1957. Metodicheskoe rukovodsto
M. E.;KONOPLIANTSEV, PO
izucheniu rezhima podzemnykh vod [Methodologicalmanual on the study of subsurface
water regime]. Moscow, Gosgeolizdat.
2. AMERICAN SOCIETYOF CIVIL ENGINEERS. 1949.Hydrology handbook. New York.(Manual
of engineering practice, no. 28.)
3. AMERICAN SOCIETY OF RANGE MANAGEMENT. 1962.Basic problems and techniques in range
research. Washington, D.C.,with the Agricultural Board, Nat. Acad. Sci.,Nat. Res.
Council. (Publ. no. 890.)
4. AMERICAN SOCIETYFOR TESTING MATERIALS. 1967.Book of ASTM standards.Philadelphia,
Pa.
5. AMOROCHO, J. ; BRANDSTETTER, A. 1966. Characterization of gauge level precipitation
patterns. Amer. Geophys. Un.
6. ANGUS, D.E.1958.Measurements of dew.Climatology and microclimatology.Proceedings
of the Canberra symposium. Paris, Unesco. (Arid zone research, XI.)
7. BALEK, J.; HOLECEK, J. 1965.The influence of density ombrographic network on the accuracy
of hydrologic computation. Symp. Quebec. (IASH publ. no. 67.)
8. BARNES, O.K.;COSTEL, G.1957. A mobile infiltrometer. Agron. J., 49: 105-7.
9. BAVER, L. D.1956. Soil physics. New York, Wiley.
10. BEARD, J. S. 1956. Results of the Mountain Home rainfall interception and infiltration
project in black wattle. J. Sth African Forestry Assn, 27: 72-86.
11. BEARD,L. R. 1965.Use of interrelated records to simulate streamflow.Proc. ASCE,91
(HYS)Pt. 1 :13-23.
12. BEEKHUIS, J.; WILL, G.M.1965.Regional variation in exotic forest productivity in New
Zealand: influencing factors.Proc. 4th N.Z.Geog. Soc. ConJ Dunedin.
13. BENEDICT, P. C. 1957. Fluvial sediment transportation. Trans. Amer. Geophys. Un.,
36(6) :897-902.
14. BERTRAND, A.R.;PARR, J. 1961.Design and operation of the Purdue sprinkling infiltro-
meter. Purdue University Agric. Expt Stn Res. Eull., no. 723.
15. BLACK,O.A. (ed.). 1965. Methods of soil analysis. Pt. 1: Physical and mineralogical
properties. Wisconsin, Amer. Soc. Agron. (Agron. monograph 9.)
16. BLAKE, G. R. 1965. Particle density. In: BLACK(1965) q.v.,chapter 29.
17. BLOEMEN,G.W.1964. Hydraulic device for weighing large lysimeters. Trans. ASAE,
7(3) 297-9.
18. BLOK,T.; COLENBRANDER, H.J. 1966.Discharge measurements by measuring structures
in relatively flat areas.
19. BOCHKOV,A. P. 1965. O raschete osadkov na vodosborakh [On the computation of
precipitation in watersheds]. Materialy soveschania PO voprosam experimentalnogo izuchenia
sloka i vodnogo balansa rechnykh vodosborov,4-7 August 1964,p. 244-51. Valdai, GGI.
186
L. 1965. Field measurement of hydraulic conductivity above a water table.

20. BOERSMA,
In: BLACK(1965), q.v.,chapter 15.
21. Bos, R. J. 1961. The Romijn measuring weir. The Netherlands,ISO, 1, 2/27,36.
22. ___ . 1965.Recommendation on the measurement of liquid flow using broad-crested
weirs. (Draft proposal for ISO.) The Netherlands.
23. BOWMAN,D.H.; KING, K. H. 1965. Determination of evapotranspiration using the
neutron-scatteringmethod. Can. J. Soil Sci.,45(2) :117-26.
24. BOYER,M. C.1957. A correlation of the characteristics of great storms. Trans. Amer.
Geophys. Un.,38(2) : 233-8.
25. BRANSON, F. A. 1962. Botanical analysis and sampling: natural pastures and range.
Pasture and range research techniques. Comstock, New York, Joint Comm. of ASA,
ADSA, ASAP,ASRM.
26. ~ et al. 1962.Effects of contour furrowing,grazing intensities and soils on infil-
tration rates, soil moisture and vegetation near Fort Peck, Montana. J. Range Man.,
15(3) : 152-8.
27. ___ et al. 1965. Plant communities and soil-moisturerelationships near Denver,
Colorado.Ecol.,46(2) : 311-19.
28. BREWER, R. 1964.Fabric and mineral analysis of soils. New York, Wiley.
29. BRITISH METEOROLOGICAL OFFICE. 1952.Observer’shandbook.London,H M S O . (MD554.)
30. ~ . 1956. Handbook of meteorological instruments, pt. 1. London, H M S O .
31. BROWN,D.1954.Methods of surveying and measuring vegetation.Bucks,England,Common-
wealth Bureau of Pastures and Field Crops. (Bull. 42.)
32. BROWN,W.G.E.;LACATE, D.S. 1961.Rooting habits of red and white pine. Canadian
Dept. Forests, Forestry Res. Branch. (Tech.note 108.)
33. BULAVKO, A. G. 1965. Experimentalnoe izuchenie zaderzhania atmosfernykh osadkov
rastitelnym pokrovom [Experimentalstudy of precipitation interception by the vegetation
cover]. In: Nauka i Tekhnika (ed.). Voprosy vodnogo khoziajstva Belorussii, p. 67-72.
34. BURGY,R. J.; POMEROY, C.R.1958.Interception losses in grassy vegetation.Trans.Amer.
Geophys. Un.,39 :1095-100.
35. BUROV,V.S. 1964.Lizimetry i vesovye ispariteli Valdaickoy gidrologicheskoy laboratorii
[Lysimeters and weighing evaporimeters at Valdai hydrological laboratory]. Materialy
Mezhduvedomstvennogo soveschaniaPO Probleme izuchenia i regulirovania isparenia s vodnoy
poverkhnosti, 30 July-3 August 1963, p. 381-3. Valdia, GGI.
36. CAIN, S.A.;CASTRO, G.M.DE O.1959.Manual of vegetation analysis.New York,Harper.
37. CAMPBELL, D. A.1951. Types of soil erosion prevalent in New Zealand,p. 82-95.Brussels.
(IASH publ. no. 35.)
38. - . 1955. Down to the sea in ships. Wellington, SC and RCC. (Bull. no. 5.)
39. CARTER, R. W.; ANDERSON, I. E. 1963. Accuracy of current-metermeasurement. Proc.
ASCE,89 (HY4)Pt. 1 :105-17.
40. CHAPMAN, T. C.1964. Design and initial instrumentation of a rainfalì/run-offexperimeni
in the Alice Springs area. Australia, CSIRO.(Tech. memo. 415.)
41.;- MCINTYRE, G.A. 1964.Gauging of shallow streams with a velocity head-rod.
Symp. Canberra.
42. CHILDS, E.C.1940.Use of soil-moisturecharacteristicsin soil studies.Soil Sci.,50:239-52.
43. CLARK, O.R.1940.Interceptionof rainfall by prairie grass,weeds and certain crop plants.
Ecol. Momgraphs, 10:243-77.
44. COLENBRANDER, H.J.; VERSTRAATE, J. M.I. 1966.Een registrerende grondregenmeter,
waarvan de gegevens automatisch kunnen worden verwerkt. Cultuurtechnisch Tijdschrijt,
6(3).
45. CORBETT, D.M., et al. 1945.Stream-gaugingprocedure. (USGSwater supply paper 888.)
46. COURT, A. 1961. Arealdepth rainfall formulas. J. geophys. Res.,66(6): 1823-31.
47. Cox, D.R. 1958. Planning of experiments. New York, Wiley.
48. CRAIG, H. 1961. Standard for reporting concentrations of deuterium and oxygen-18in
natural waters. Science, 133 :1833-4.
49. ___ ; GORDON, L.I. 1965.Deuterium and oxygen-18variations in the ocean and the
marine atmosphere. Symp.on stable isotopes in oceanograph studies and paleotemperatrires,
Spoletto, Italy, p. 9-130.
50. CROUSE, R.P.,et al.1966.Methods of measuring and analysing rainfall interception by
grass. ZASH Bull., 11(2): 110-20.
I87
51. DAGG, M. 1962.Physical properties of the surface soils, infiltration and availability of
soil moisture. E. Afr. Agric. For. J., 27:68-70.
52. DANSGAARD, W. 1964. Stable isotopes in precipitation. Tellus, 16(4) :436-68.
53. DAVIS, G.H.et al. 1966.Seasonal variations in the tritium content of groundwaters of
the Vienna Basin, Austria. In: INTERNATIONALATOMIC ENERGYAGENCY(1966), 9.v.
54. DAY, P.R.1965.Particle fractionationand particle-sizeanalysis.In: BLACK(1965), q.v.,
chapter 43.
55. DENDY, F.E.;ASSMUSSEN, L. E.1963.Permeability measurements with small well points.
Presented to S.E.Section, ASAE. Tennessee.
56. DESI, F.et al. 1965.On determining the rational density of precipitation measuring networks,
p. 127-30. Symp. Quebec. (IASH publ. no. 67.)
57. DROST, H.1966. Pre-calibrated fibreglass flow-measurementstructures for experimental
basins. J. Hydrol. (N.Z.), 5(1): 20-4.
58. DROZDOV, O.A.,et al. 1965.Pogreshnost ucheta atmosfernykh osidkov [Errors in pre-
cipitation measurements]. Trudy GGO,vol. 175, p. 24-30.
59. DUBREUIL, P. 1966.Point de vue théorique sur le rôle du sol dans le cycle hydrologique.
Paris. (Cahier Orstom d’hydrologie,no. 6.)
60. DUVDEVONI, S. 1947.A n optical method of dew estimation.Roy. Meteorol. Soc. Quart.
J., 73 (317-8): 282-96.
61. DYGALO, V.S.1959.K voprosu o tochnostiizmereniaosadkov (PO materialam nabludeniy
stokovoistantsiiB.Sareevo) [On the problem ofthe accuracy of precipitation measurements
(by the data of observation on experimentalcatchment of B.Sareevo)]. Trudy GIP,vol.94,
p. 92-9.
62. ELLISON, W. D.1947. Soil erosion studies.Agric. Eng., 28:145444.
63. EMMETT, W.W.1965.The vigil network: methods of measurement and a sampling of data
collected, p. 89-106. (IASH publ. no. 66.)
64. ERDELYSZKY, Zs.; UBELL, K.1962.Experiments for determiningthe variations of ground-
water flow velocity with depth. Periodica Polytecnika, Budapest, 6(3) :205-17.
65. EZEKIEL, M. 1958. Methods of correlation and regression analysis. New York,
Wiley.
66. FEDOROV, S. F. 1950. Opyt primenenia dozhdevalnoy ustanovki dla izuchenia infiltra-
tsionnoy sposobnostipochvy [Someexperience of the sprinkler application for the study
of soil infiltration rate]. Trudy GGZ, vol. 24 (78), p. 109-21.
67. - . 1954. Experimentalnye izuchenie infiltratsii na slabopodzolistykh pochvakh
[Experimentalstudy of infiltrationrate for weakly podzolic soils]. Trudy GGZ,vol.46 (loo),
p. 48-72.
68. ~ . 1963,O vozmozhnosti ispolzovania lizimetricheskikh ustanovok dla izuchenia
elementov vodnogo balansa lesnykhvodosborov [Onthe possibility of lysimeter application
for the study of water balance components of afforested catchments].Materialy mezhduve-
domstvennogo soveschania PO Probleme izuchenia i regulirovania isparenia s vodnoy po-
verkhnosti i pochvy, 30 July-3 August 1963, p. 400-9. Valdai, GCI.
69. FEODOROFF, A.1965.Etude expérimentale de l’infiltrationde l’eaunon saturé.Ann. Agron.,
16 : 127-75.
70. FIERING, M.B. 1964.A multivariate technique for synthetic hydrology.Proc. ASCE, 90
(HY5) 43-61.
71. FILIPPOVA, A. K. 1964.Resultaty sjemok vlazhnosti pochvy na experimentalny vodos-
borakh stokovykh stantsiy [Resultsof soil-moisturesurvey on experimentalbasins]. Trudy
GGZ, vol. 92,p. 104-18.
72. FORSGATE, J. F.,et al. 1965.Design and installation of semi-enclosedhydraulic lysimeters.
Agric. Meteorol., 2 :43-52.
73. GEIGER, K.;HITCHON, B. 1965. Ground-watermeasurement.Proc. Hydrol. Symp. no. 4,
p. 245-65. Ottawa,Queen’s Printer.
74. GERTYSK, V.V.1957.Some data on the role of growing seasonprecipitation in replenishing
the moisture supply in the soil. Trans. V.V. Alekhin Central-Chernozen State Reserve,
Iss. 4.
75. GILLIES, A. J. 1964. Review of snow survey methods and snow surveys in the Fraser
Catchment, Central Otago. J. Hydrol. (N.Z.), 3(1).
76. GOLUBEV, V. S. 1960. O b uchete dozhdevykh osadkov razlichnymi priborami [On the
inventory of rainfall precipitation by various devices]. Trudy GGZ, vol. 81,p. 5-18.
188
Methods of observation and instrume~itarion
77. ~ . 1965. Nekotorye resultaty issledovaniy na osadkomernom poligone VNIGL

[Some results of investigations on pluviometric experimental area in VNIGL]. Trudy
GGI,VOI. 175, p. 81-95.
78. -- . 1966. Metody otsenki i rascheta sezonnykh i godovykh summ osadkov na
vodosborakh i sutochnykh summ osadkov na selskokhoziistvennykhpoliakh [Methods
of evaluation and computation of total seasonal and annual precipitation on catchments
and total daily precipitation on the fields]. Materialy seminare PO ras chetam vodnogo
balansa rechnykh basseinov i organizatsii komplexnykh vodnobalansovykh i agrometeoro-
logicheskikh nabludeniy, p. 13-27. Valdai, GGI.
79. GOODELL, B. C.1963.A reappraisalof precipitation interception by plants and attendant
water loss. J. Soil Water Cons.,18:2314.
80. GRADWELL, M.W. 1960. Changes in the pore-spaceof a pasture topsoil under animal
treading. M.Z.J. agric. Res., 3 :663-74.
81. GREIG-SMITH, P. 1964. Quantitative plant ecology. London,Butterworth.
82. GUY, H.P.,et al. 1960.Laboratory theory and methods. USGS.(open-filereport.)
83. HADLEY, R. F.,et al. 1961.Hydrologic effects of water spreading in Box Creek Basin,
Wyoming. (USGS water supply paper 1532-A.)
84. ~ ; BRANSON, F.A.1965.Surficial geology and micro-climaticeffects on vegetation,
soils and geomorphology in the Denver area, Colorado. Guidebook for one-day jîeld
conferences, p. 56-61.Int. Ass. for Quaternary Res. (VIIth congress.)
85. HAISE, H., et al.1956.The use of cylinder injîltrometers to determine the intake characteristics
of irrigated soils. USDA. (ARS 41-7.)
86. HAMILTON, E.L. 1954. Rainfall sampling in rngged terrain. Washington, USDA.(Tech.
bull. 1096.)
87. ___ ; ROWE,P.B. 1949.Rainfall interception by chaparral in California.Calif. Dept.
Nat. Res.,Forestry Div. and U.S.Forest Serv.,Calif. Forest and Range Expt. Stn.
88. HARROLD, L.L.,et al. 1962.Influenceof land use and treatment on the hydrology of small
watersheds at Coshocton, Ohio, 1938-57. USDA. (Tech. bull. 1256.)
89. HAYNEJ, J. L. 1940. Ground rainfall under vegetative canopy of crops. J. Amer. Soc.
Agron., 32:176-34.
90. HELVEY, J. D.1964.Rainfall interception by hardwood forest litter in the Southern Appa-
lachians. U.S.Forest Serv.,South-easternForest Expt. Stn. (Res.paper 8.)
91. ~ ; PATRIC, J. H.1965.Canopy and litter interception of rainfall by hardwoods
of Eastern United States. Water Resources Res., l(2) : 193-206. (Amer.Geophys. Un.)
92. -; ___ . 1965.Design criteria for interception studies,p. 131-7. (IASH pubi.
no. 67.)
93. HERSHFIELD, D.M. 1965. On the spacing of raingauges. WMOIIASH symp. on design
of hydrometeorological networks, Quebec.
94. HERMSMEIER, L.,et al. 1963. Construction and operation of a 16-unitrainulator. USDA.
(ARS 41-62).
95. HEWLETT, J. D.;HIBBERT, A.R.1963. Moisture and energy conditions within a sloping
soil mass during drainage.J. geophys. Res.,68 : 1081-7.
96. ___ ; et al.1963.Instrumentaland soil-moisturevariance using the neutron-scattering
method. J. Soil Sci.,96.
97. HOFFMAN, G. 1956. Measurement of rate of dewfall. Summary in Nature, 177 (4499):
118.
98. HOPPE, E.1896.Precipitation measurements under tree crowns. Transi. by A.H. Krappe.
US. Forest Serv. (Translation no. 291, 1935.)
99. HORTON, J. S.,et al. 1964. Guide for surveying phreatophyte vegetation.USDA.(Agric.
handbook no. 266.)
100. HORTON, R.E.1919. Rainfall interception.Monthly Weather Review, 47 : 603-23.
101. HOSEGOOD, J. 1963.The root distribution of kikuyu grass and wattle trees.E.Afr. Agric.
For. J., 29(1) :60-1.
102. HUBBELL, D.W. 1964.Apparatus and techniques for measuring bed load. (USGSwater
supply paper 1748.)
103. HLJDDLESTON, F. 1933. A summary of seven years’ experiments with raingauge shields
in exposed positions, 1926-32, at Hutton John, Penrith. British Rainfall: 274-93.
104. INTERNATIONALATOMIC ENERGYAGENCY. 1962.Isotope techniquesfor hydrology.Vienna,
(Tech. report no. 23.)
189
105. ~ . 1962-65. Tritium concentration in rain, rivers, oceans and other waters. Vienna.
(WP/17/1-5.)
106. ~ . 1963. Radioisotop& in hydrology.Proc. Symp. Tokyo. Vienna. (STI/Pub/71.)
107. ~ . 1964. The application of radioisotopes to the study of bed-load movement and
transport in rivers. Report on a meeting of experts. Vienna. (STI/Rep/lOl.)
108. ~ . 1964. Tritium and other environmental isotopes in hydrology, meteorology and
oceanography. Report of a panel meeting. Vienna.
109. ___ . 1965. Isotopes in hydrology. Report of a working group in Grenoble.
110. ___ . 1965. Radioisotope instrumentsin industry and geophysics.Proc.Symp. Warsaw.
Vienna. (STI/Pub/ll2.)
111. ~ . 1966. Safe handling of radioisotopes in hydrology. Vienna. (STI/Pub/l31.)
112. ____. 1966. Use of isotopes in hydrology. Proc. Symp. Vienna.
113. JACQUET, J. 1960. Répartition spatiale des précipitations ù l’échelle fine et précision des
mesures pluvioniétriques. Assemblée d’Helsinki.(IASH publ. no. 53.)
114. 1963. Comparaison des procédés de mesure directe de l’évaporationsur des bassins
versants expérimentaux. Assemblée de Berkeley. (IASH publ. no. 53.)
115. JEFFREY, W.W. 1964. Watershed research in the Saskatchewan River headwaters.Proc.
fourth synip. on research watersheds. Guelph,Ontario Agric. College.
116. JENNINGS, E.G.; MONTEITH, I. L.1954.A sensitiverecordingdew balance.J.Roy.Meteorol.
soc., 80 :344.
117. JOHNSTONE, D.;CROSS, W.P. 1949. Elements of applied hydrology. New York.
118. JONES, P. H.;SKIBITZKI,N. 1956. Subsurface geophysical methods in ground-water
hydrology. Advances in Geophys., 3 :241-300. (Amer. Geophys.Un.)
119. JONES, R.L. 1964. A simple dewfall integrator.N.Z.J. Sci.,7(1) :45-50.
120. JUNG, E. 1931. Untersuchungen über die Einwirkung des Frostes auf dem Erdboden.
Kolloidchem, Beihefte, 32 :320-73.
121. KAMIYAMA, K.;MORIGUCH, M.1953. Collectionof fog particles with fine fibre and infra-
red absorption of fog particles. Papers in Meteorology and Geophys.,3 :104-13.
122. KEMPER, W.D.1960. Water and ion movement in thin films as influenced by the electro-
static charge and diffusion of cations associated with clay mineral surfaces.Proc. Soil
Sci. Soc. Amer., 24(1) : 10-6.
123. KINDSVATER, E.; CARTER, R.W.1959. Discharge characteristics of rectangularthinplate
weirs. Trans. ASCE,124: 772-802, paper no. 3001.
124. KING, H. W.;BRATER,E. F. 1963. Handbook of hydraulics. New York, McGraw-
Hill.
125. KITTREDGE, J. 1948. Forest influences.New York, McGraw-Hill.
126. ___ , et al. 1941. Interception and stem-flowin a pine plantation. J. Forestry, 39 :
505-22.
127. KLUTE, A. 1965. Laboratory measurement of hydraulic conductivity in saturated soil.
In: BLACK(1965), q.v.,chapter 13.
128. ___ . 1965. Laboratory measurement of hydraulic conductivity in unsaturated soil.
In: BLACK (1965), q.v.,chapter 16.
129. ~ . 1965. Water diffusivity.In: BLACK (1965), q.v.,chapter 17.
130. ___ , et aí. 1965. Steady-statewater flow in a saturated inclined soil slab. Water
Resources Res., l(2) : 287-94. (Amer. Geophys. Un.)
131. KOLUPAILA, S. 1964. Discussion.Proc. ASCE, 90 (HYI):352.
132. KONSTANTINOV, A. R.;FEDOROV, S. F. 1960. Opyt primenenia gradientnykh macht dla
opredeleniaisparenia i teploobmena v lesu [Some experience of the application of gradient
masts to determine evaporation and heat exchange in the forest]. Trudy GGZ,vol. 81,
p. 91-114.
133. ___ ; KISILENKO, A. A. 1965. Experimentalnye issledovania tochnosti izmerenia
zhidkikh osadkov razlichnymi priborami [Experimental investigation on the accuracy
of liquid precipitation measurements by various devices]. Trudy GGO,vol. 175.
134. KONTORSCHIKOV, A. S.;EREMINA, K.A. 1963. Interception of precipitation by spring
wheat during growing season. Soviet Hydrol.,4 :400.
135. KOSCHMEIDER, H.1934. Methods and results of definite rain measurements:III. Danzig
report (i). Monthly Weather Review, 62 :5-7.
136. KOZLIK, V. 1965. Notions relatives aux recherches scientifiques du réseau nivométrique
représentatif, p. 168-78. Symp. Quebec. (IASH publ. no. 67.)
190
Meihods of observation and instrumentarion
I 137. KRESTOVSKY, O.I.,et al. 1966. Metody ucheta izmenenia zapasov vody na poverkhnosti
vodosborov v pochvo-gruntakh zony aeratsii i zapasov gruntovykh vod [Inventory
methods of water storage change on the surface of basins in soils and subsoils of the
unsaturated zone and underground water storage]. Materialy seniinara po raschetam
vodnogo balansa rechnykh basseinov i organizarsii komplexnykh vodno-balansovykh i
I
agrometeorologicheskikh nabludeniy, p. 67-102. Valdai, GGI.
138. KUZMIN, P. P. 1960. Formirovanie snezhnogo pokrova i metody opredelenia snegozapasov
[Snow-coverformation and methods of snow-storagedetermination]. Leningrad,Gidro-
meteoizdat.
139. . 1961. Protsessy tajania snezhnogo pokrova [Snow-pack melting processes].
Leningrad, Gidrometeoizdat.
140. LANGBEIN, W.B.; HARDISON, C.H.1955. Exíending stream-flowdata.ASCE. (Paper 826.)
141. LAZAREV, K.G.(ed.). 1962.Sovremennye metody analiza prirodnykh vod [Modern methods
of natural water resources analysis]. Moscow, Academy of Sciences of the U.S.S.R.
142. LEBEDEV, A. V. 1963. Metody izuchenia balansa gruntovykh vod [Methods of subsurface
water balance study]. Moscow, Gosgeolizdat.
143. ~- . 1964. Sravniteinaja otsenka dostovernosti nabludeniy za olementami balansa
gruntovykh vod PO lizimetrani [Comparativeassessment of reliability of observations of
underground water components by lysimeters]. Materialy mezhduvedornstvennogo soves-
chania po Probleme izuchenia i regulirovania isparenia s vodnoy poverkhnostin i pochvy,
30 July-3 August. Valdai, GGI.
144. LEONARD, R. E. 1961. Interception of precipitation by northern hardwood. U.S.Forest
Serv. North-easternForest Expt. Stn. (Paper 159.)
145. LEOPOLD, L. B. 1962. The vigil network. ZASH BuII.,7(2): 5-9.
146. LEVY, E.B.;MADDEN, E.A.1933.The point method of pasture analysis.N.Z. J. Agric.,46.
147. LEYTON, L.;CARLISLE, A. 1959. Measurement and interpretation of interception of pre-
cipitation by forest stands, p. 11 1-19 Symp. Hannoversch-Munden.(IASH publ. no.48.)
148. MARSHALL, T. J. 1959. Relations between water and soil. Harpenden, Commonwealth
Bureau of Soils. (Tech. comrn. no. 50.)
149. MARTINEC, J. 1961. Snowmelt-run-offforecasting on the Vltava River. Prague,Hydraul.
Res.Inst. Podbabska.
150. MAXEY, G.1964. Geology.Pt.I: Hydrogeology.In:VENTECHOW(ed.), 1964. Handbook
of applied hydrology. New York, McGraw-Hill.
151. MCMILLAN, W.D.;BURGY,R. H.1960. Interception loss from grass. J. geophy. Res.,
65 :2389-94.
152. MCQUEEN, 1. S. 1963. Coolev and humidifier for subsoil centrifuge. US. Patent Office,
patent no. 3,109,872.
153. . 1963. Development of a hand-portable rainfall simulator injìltrometer. USGS.
(Circ. 482.)
154. ; MILLER, R. F. 1966 (in press). Calibration and evaluation of a wide-range
method for measuring moisture stress in field soil samples.Proc. Unesco symp. on water
in the unsaturated zone, 19-25 June,Wageningen.
155. MILLER, R.F.,et al.1961. Soil moistnre under juniper andpinyon compared with moisture
under grassland in Arizona. p. 233-5. (USGS prof. paper 424-B.)
156. ; RATZLAFF, K.W.1961. Water movement and ion distribution in soils,p. 45-6.
(USGS prof. paper 424-B.)
157. ___ -- . 1965. Chemistry of soil profiles indicates recurring patterns and
modes of moisture migration.Proc. Soil Sci. Soc. Amer.,29 (23): 263-6.
158. MOLCHANOV, A.A. 1960. The hydrological role of forests. Translated and published for
U S D A and National Science Foundation,Washington, D.C. by Israel Programme for
Scientific Translations.
1 159. MORRISSEY, W.B. et al. 1966. Snow surveys. In: NEWZEALAND MINISTRY OF WORKS,
1961. Handbook of hydrologicalprocedures.
160. MUSGRAVE, G.W.; FREE, G.R.1937.Preliminaryreporton a determinationof comparative
I infiltration rates on some major soil types. Trans. Amer. Geophys. Un.,18 : 345-9.
l 161. ; HOLTAN, H.N. 1964. Infiltration.In: VEN TE CHOW(ed.), 1964. Handbook
of applied hydrology. New York, McGraw-Hill.
' 162. NEBOLSIN, S. I. 1929. Meteorological conditions in field vegetation: precipitation and
evaporation.Meteorol. Herald, l(2).
191
Represetitctive and experimental basins
163. ;- NADEEV, P.P. 1937. Elementary poverkhnostny stok (Surface run-off plots).
Leningrad,Gidrometeoizdat.
164. NECHAEV, I. N.1965. Poteri osadkov na smachivanie osadkomerov i metodika korrek-
tirovki summ osadkov [Precipitationlosses by gauge wetting and methods of total rainfall
correction]. Trudy GGO, vol. 175, p. 76-86.
165. NEWZEALAND. MINISTRY OF WORKS. 1965.Calibration of stream-gaugingweirs. Wellington.
(Central Lab. report no. 205.)
166. NIELSEN, D.R.,et al. 1964.Water movement through Panoche clay loam soil. Hilgardia,
35 :491-505.
167. NIR, A. 1964. On the interpretation of tritium 'age' measurements of groundwater.J.
geophys. Res., 69(12) :2589-95.
168. OSBORNE, B. 1953. Field measurements of soil splash to evaluate ground cover. J. Soil
Water Cons., 8 :255-60, 266.
169. OURYEV, V. A. 1953. Experitnentalnye gidrologicheskie issledovania na Valdae [Experi-
mental hydrological investigations in Valdai]. Leningrad,Gidrometeoizdat.
170. PACIFIC SOUTHWEST INTERAGENCYCoMMrr'rEE.1966.Limitarions in hydrologic data. Report
of Hydrological Subcommittee,USDI.
171. PARR, J.; BERTRAND,A.R. 1960.Water infiltrationinto soils.Advances in Agron.,12 :311-63.
172. PAYNE, B. R.;DINÇER, T. 1965. Isotope survey of the Karst region of southern Turkey.
Proc. sixth int. conj on radiocarbon and tritium dating. Pullman, Washington.
173. PENMAN, H. L. 1948. Natural evaporation from open water, bare soil and grass. Proc.
Roy. Soc., A193 :120-45,
174. PEREIRA, H.C.,ec al. 1958. Water conservation by fallowing in semi-arid,tropical East
Africa. Emp. J. expt. Agric., 25(103).
175. ~ . 1962. Assessment of the main components of the hydrological cycle. E.Afr.
Agric. For. J., 27 (special issue).
176. PEREZ, V. F. 1965.Design and operation of hydrometeorological networks in tropical regions.
WMO and IASH.
177. PETERS, D.B. 1965. Water availability.In: BLACK(1965), q.v.,chapter 19.
178. PHILIP, J. R.1966. Plant/waterrelations:some physical aspects.Ann. Rev. Plant Physiol.,
17 :245-68.
179. ___ ;DEVRIES, D.A.1957.Moisture movement in porous materials under temperature
gradients.Trans. Amer. Geophys. Un.,38(2) :222-32.
180. PHILLIPS, E. A. 1959. Methods of vegetation study. New York, Henry Holt.
181. PHILLIPS, J. F. V. 1926. Rainfall interception by plants. Nature, 118: 837-8.
182. POPOV, O.V. 1959. Lysimeters and hydraulic soil evaporimeters, p. 26-37. Symp. Hanno-
versch-Munden,8-13 Sept.(IASH publ. no. 49,)
183. ___ ; PUSHKAREV, V. F. 1964. Gidravlicheskie pochvennye ispariteli-pribory dla
izucheniavodnogo balansa pochvenno-gruntovoy tolschi [Hydraulicsoil evaporimeters-
devices for water balance study of soils and subsoils]. Sbornik rabot PO metodike issledo-
vaniy v oblasti Jiziki pochv. Leningrad,Agrophysical Res. Inst.
184. PRILL, R.C.; JOHNSON, A.I.1959.Effect oftemperatureon moisturecontentsas determined
by centrifuge and tension techniques. ASTM special publ. 254, p. 340-9.
185. RAUZI, F. 1963. Water intake and plant composition as affected by differential grazing
on rangeland. J. Soil Water Cons., 18(3): 114-6.
186. REE,W.O. 1965. Swiss channel-type gauging stations. USDA,Agric. Res. Serv.
187. --; CROW, F.R.1954.Culverts aswater run-offmeasuring services.Ag&. Eng..35(1).
188. REYNOLDS, E. R. C.;LEYTON, L. 1963. Measurement and significance of throughfall in
forest stands. The water relatiom of plants, p. 127-41. Dorking,Blackwell Sci.Publ.
189. RICHARDS, L.A. 1965.Physicalcondition of water in soil.In:BLACK(1965), q.v.,chapter8.
190. ___ et al. 1954. Saline and alkali soils. USDA. (Agric. handbook 60.)
191. RIJTEMA, P. E. 1965. An analysis of actual evapotranspiration. Netherlands, Rept. Inst.
Land Water Management. (Res.paper no. 659.)
192. ROCHE, M. 1963. Hydrologie de surface. Paris,Orstom.
193. RODDA, J. C. 1967. The systematic error in rainfall measurement. J. Inst. Water Eng.,
21 (March) : 173-7.
194. RODE, A.A.1960. Metody izuchenia vodnago rezhimapochv [Methods for studying hydro-
logical soil regimen]. Moscow,Acadexy of Sciences of the U.S.S.R.
195. Roo, S. S. 1965. Resultaty experimentalnykh issledovaniy infiltratsii na malykh vodos-
192
borakh v zone nedostatochnogouvlazhnenia [Expcrimentalresults of infiltration research

on small basins in the zone of insufficient moistening]. Trudy GGZ,vol. 125, p. 79-120.
196. ROTHACHER, J. 1963. Net precipitation under a douglas-fir forest. Forest Sci., 9 :423-9.
197. RUBIN, J.; STEINHARDT, R. 1963. Soil/waterrelations during rain infiltration. I: Theory.
I Proc. Soil Sci. Soc. Amer., 27 :246-51.
198. RUTKOVSKY, V.I. 1936. Investigation of interception of precipitation by grass and moss
covers. Meteorol. and Hydrol., no. 11.
1 199. RUTTER, A. J. 1963. Studies in the water relations of Pinus sylvestris in plantation con-
ditions. I: Measurements of rainfall and interception. J. Ecol., 51 : 191-203.
200. SELBY, M . J. 1966. Methods of measuring soil creep. J. Hydrol. (N.Z.), 5(2): 54-63.
201. SHAMOV,G.I. 1959.Rechnye nanosy. Rezhim, raschety i metody izmereniy [River sediments.
I
Regime, computation and methods of measurement]. Leningrad, Gidrometeoizdat.
202. SHARP, A. L.;HOLTAN, H.N.1940. A graphical method of analysis of sprinkled-plot
hydrographs. Trans. Amer. Geophys. Un.,21 :558-70.
203. SIMAIKA, Y. M. et al. 1962. Methods of measurement of discharge and equipment for
survey in the Nile Basin. U.N.conf.for the application of science and technology for the
benefit of less developed areas, paper E.
204. SNEDECOR, G.W. 1946. Statistical methods. Ames,Iowa State College Press.
205. SOUBOTIN, A.S. 1964. Obzor lizimetrov i osnovnye trebovania k ikh konstruksiam [Review
of lysimeters and general requirements for their design]. Trudy GGZ,p. 348.
206. - - . 1965. K voprusu ob izmerenii elementov balansa gruntovykh vod lizimetrami/
kompensatsionnymi isparitelami [On the problem of measurements of ground-waterbudget
components by lysimeters/compensating evaporimeters]. Trudy GGZ,vol. 125, p. 69-78.
207. SPECHT,R.L. 1957. Dark Island heath. IV:Soil moisture patterns produced by rainfall
interception and stem flow. Aust. J. Bot.,5(2): 137-50.
208. STEIN, C.1945. A two-sampletest for a linear hypothesis whose power is independent
of the variance. Annals Math. Stats., 16 :243-58.
209. STRUZER, L. R. 1965. Osnovnye nedostatki i puti ulutshenia metodov izmerenia atmos-
fernykh osadkov [General drawbacks and ways of improvement of atmospheric precipi-
tation measurements]. Trudy GGO vol. 175, p. 5-23.
211. SWANSON,C.L. W.; PETERSON, J. B. 1942. The use of micrometric and other methods
for the evaluation of soil structure. Soil Sci., 53 :173-85.
212. SWANSON, N.P.,et al. 1965. Rotating boom rainfall simulator. Trans. Amer. Soc. Agric.
Eng., 8 :71-2.
213. THORNTHWAITE, C.W.; MATHER, T.R. 1957. Instructions and tables for computing
potential evapotranspiration and the water balance. Publ. in Climatology, lO(3). N e w
Jersey, Drexel Inst. of Technology.
214. TOEBES, C.1964. Applied hydrology, 2v. Wellington, Govt. Printer, 24 p.
215. TRESTMAN, A. G.1960. N e w aspects of river run-of calculations. Translated from the
Russian for U.S.Dept. C o m m . and National Science Foundation, Washington, D.C.
216. TROSKELANSKI, A. I. 1960. Hydrometry. N e w York, Pergamon Press.
217. TURC, L. 1953. L e bilan d'eau des sols et relations entre les précipitations, l'évaporation et
l'écoulement. Inst. Nat. de la Recherche Agric., Univ. de Paris.
218. UBELL, K.1959. Moisture movement in unsaturated soils with special regard to the utilization
of lysimeter observations, p. 1953-65. Symp. Hannoversch-Munden.(IASH publ. no. 42.)
219. . 1965. Investigations into groundwater balance in experiniental areas, p. 388-98.
Symp. Budapest. (IASH publ. no. 66.)
220. UNITED STATESDEPARTMENT OF AGRICULTURE. 1951. Soil survey manual. (Agric. handbook
I no. 18.)
221. . 1961. A universal equation for predicting rainfall-erosion losses. Agric. Res.
Serv. (Special report, 22-26.)
~ 222. . 1962. Field manual for research in agricultural hydrology. (Agric. handbook
no. 224.)
223, UNITED STATESDEPARTMENT
OF THE INTERIOR.1953. Water measurement manual. Bureau
of Reclamation.
~ 224. UNITED STATES INTERAGENCY COMMITTEE ON WATER RESOURCES. 1963. Proc. federal
Interagency sedimentation conference. Reports 1 and 14. Agric. Res.Serv.(Misc.publ. 970.)
I 225. VAN BAVEL, G.H.M.,et al. 1963. Soil moisture measurements with the neutron method.
U S D A . (ARS 41-70.)
193
226. VILLEMONTE, I. R.1943. New-type gauging station for small streams. Eng. News-Record,
131(21).
221. VOMOCIL, J. A. 1965. Porosity. In: BLACK(1965), q.v., chapter 21.
228. WHIPKEY, R. Z. 1965. Subsurface storm flow from forested slopes. IASH Bull.,lO(2):
14-85.
229. WILCOX, J. C. 1965.Time of sampling after an irrigation to determine field capacity of
soil. Can. J. Soil. Sci.,45: 171-7.
230. WORLD METEOROLOGICAL ORGANIZATION. 1962. Guide to meteorological instruments and
observing practices. ( W M O - N o . 8. TP3.)
231. . 1965. Guide to hydrometeorological practices. ( W M O - N o . 168, TP 82.)
232. YOUNG, A. 1960. Soil movement by denudational processes on slope. Nature, 188(4745) :
120-2.
233. ZHELEZNIAKOV, O.V.; DANILEVICH, B. B. 1966. Tochnost gidrologicheskikh izmereniy i
raschetov [The accuracy of hydrological computation and measurements]. Leningrad,
Gidrometeoizdat.
234. ZOTIMOV, N.V. 1964. Opyt izmerenia vlazhnosti pochvy na lizimetre gammametodom
[Experience of lysimeter soil moisture measurements by gamma-method]. Trudy GGI,
vol. 109, p. 140-50.
235. . 1965. O nazemnom metode izmerenia zapasov vody v snege s pomoschju
radioaktivnosti zemli [On terrestrial method of measurement of water equivalent in snow
pack by means of radioactivity of the earth]. Trudy GGI, vol. 130, p. 148-62.
236. ANON.1955. A million random digits with one hundred thousand normal deviates. Glencce,
Ill.,Rand Corp.,Free Press.
237. ___ . 1966. Metodicheskie ukazania Upravleniam gidrometsluzhby [Methodological
instructionsto the offices of the Hydrometeorological Service]. No. 71,Primenie materialov
aerofotocjemki dla izuchenia ruslovykh protsessov i deformatsii beregov vodokhranilisch.
238. , 1966. Metodicheskie ukazania Upravleniam gidrometsluzhby [Methodological
instructions to the offices of the Hydrometeorological Service]. No. 72, Opredelenie

poverkhnostnykh skorostey techenia (raskhovod vody) rek s pomoschju aerofotosjemki i
sjemki s berega. Leningrad, Gidrometeoizdat.
239. . 1958.Nastevlenie gidrometeorologicheskimstantsijam i postam [Instructions for
hydrometeorological stations and posts]. No. 3, pt. 1, Meteorologicheskie nabludenia na
stantsijakh. Leningrad, Gidrometeoizdat.
240. , 1957. Nastevlenie gidrometeorologicheskimstantsijam i postam [Instructionsfor
hydrometeorological stations and posts]. No. 6, pt. 1, Gidrologicheskienabludenia i raboty

na rekakh. Leningrad, Gidrometeoizdat.
241. . 1952. Nastevlenie gidrometeorologicheskirn stantsijam i postam [Instructions for
hydrometeorological stations and posts]. No. 6, pt. 2, Nabludenia na malykh rekakh.
242. -_ . 1961. Nastavlenie gidrometeorologicheskimstantsijam i postam [Instructionsfor
hydrometeorological stations and posts]. No. 7, pt. 2, Nabludenia nad ispareniem s vodnoy
poverkhnosti. Leningrad, Gidrometeoizdat.
243. . 1959. Organizatsia i proizvodstvo nabludeniy nad vlazhnostju pochvogrunta na
malykh vodosborakh stokovykh stantsiy [Organization and performance of soil and subsoil
moisture measurements on small experimental basins]. (Metodicheskie ukazania GGI
Upravleniam Gidrometsluzhby,no. 52.)Leningrad, Gidrometeoizdat.
244. . 1964. Rukovodstvo PO gradientnym nabludeniam i opredelenie sostavliajuschikh
teplovogo balansa [Guide for gradient observations and determination of heat budget
components]. Leningrad, Gidrometeoizdat.
245. . 1966. Rukovodstvo PO poverke gidrologicheskikh priborov [Guide for checking
hydrological instruments]. Leningrad, Gidrometeoizdat.
246. . 1967. Rukovodstvo PO poverke meteorologicheskikh priborov [Guide for checking
meteorological instruments]. Leningrad, Gidrometeoizdat.
241. . 1963.Rukovodstvo POproizvodstvu nabludeniy nad ispareniem s pochvy i snezhnogo
pokrova [Guide for evaporation measurements from land and snow cover]. Valdai,
GGI.
248. - - . 1951. Rukovodstvo PO proizvodstvu nabludeniy nad ispareniem s selskhokho-
zajstvennykh Polei [Guide for evaporation measurements from agricultural fields]. Pt. 1,
194
Nabludenia nad isparenium metodom pochvennykh ispariteley. Pt. 2, Nabludenia nad

ispareniem gradientnym metodom. Leningrad, Gidrometeoizdat.
249. ~ . 1954. Rukovodstvo stokovym stantssijarn [Guide for experimental catchments].
250. -.1965. Ukazaniapoproizvodstvu snegomernykh nabludeniy na gidrometeorologicheskikh
stantsijakh i postakh [Instructions for snow-cover measurements on hydrometeorological
stations and posts]. Leningrad, Gidrometeoizdat.
5 Data processing
and publication
5.0 General
Routine data processing for representative and experimental basins includes the collec-
tion of reliable cartographicalmaterial, used for a physiographical description of basins
and the processing of observationalresults of major hydrologicalvariables.
General recommendationson data-processingmethods are given below.
Some observationaldata should be published for international exchange. This can be
in the form of tables of standard observationaldata on representativeand experimental
basins together with summaries and reviews on research carried out. The major part of
the data should not be published for internationalexchange and is either kept by national
committees for the IHD or with the hydrological or hydrometeorological service of the
country.
5.1 Mapping of representative and experimental basins

Basin maps provide a picture that the user may interpret visually as a model of certain
features of the basin, and a record of definite information regarding the positional rela-
tionshipsof the features.Below are pointed out special problems associated with the map-
ping of representativeand experimentalbasins and ways of making the maps most useful
are suggested.For mapping techniques refer to the literature[lo,18,44,621.
Ideally,mapping of all the pertinent features of a basin should be completed before
making the selection of the sites for instrumentation.Proper location of instrumentation
depends primarily upon knowledge of the physical features of the basin. Because map-
ping is a lengthy operation which must be done at the expense of data collection,it is
not always possible to follow this ideal schedule.
The choice of scales and features represented on the maps will depend on the size of
the basin and the purposes for which it is being instrumented,and also on the resources
available. An adequate working scale for a map of a 200 kmz representativebasin,to be
used with many others for a generalized study of annual maximum flood flows,might be
1 :50,000,but an adequate scale for a map of a one-hectarebasin to be used for detailed
studiesof overland flow or infiltrationmight be 1 :500.
Many potential problems can be avoided by using the same base map for maps of
topography,soils,geology,land use,etc. Ordinarily the map used for this purpose is a
topographical base map showing the basin divide,drainage net, civil divisions and any
other major cultural features that provide orientationfor the user.
196
Data processing and publication
5.1.1 Topographical mapping

5.1.1.1 Techniques
Topographical mapping may be by instrument survey or photogrammetric techniques.
, In either case,it is important that permanent,well-identified benchmarks be established.
i
The benchmarks should preferably be related to the mean sea level datum.
' 5.1.1.2 Features

A decision on the featuresto be shownon the topographical map must be made in advance
of field survey activities so that the appropriateinformation may be assembled.If resour-
ces permit, more, rather than less, information should be assembled,even though the
need for all of it cannot be foreseen at the time.
5.1.I .2.1 CONTOURS
Elevations of land surfaces,which are the basic data for establishing the basin divide,
basin area and other geomorphologicalcharacteristics(see section 6.1.1.2),are expressed
by contours.The following is a guide to a contour interval which may serve as an example
for mapping land of mild slope:
Size of busin M u p srak Coniour
(kni2) inirrval (m)
0.01 1 : 2,000 0.5

1 1 : 5,000 1
1 O0 1 :10,000 2
1,000 1 :25,000 5
In rugged,mountainous areas larger contour intervals are used, and in plains areas they
may be smaller.
5.1.1.2.2 BASIN DIVIDE
Special attention should be given to establishing the location of the basin divide. O n
large basins it may be located and established by inspection of the completed contour
map, but on small basins,on-the-groundinspection and surveys should be made to identi-
fy the natural divide. Moreover, on very small basins it may be necessary to construct
small embankmentsto make certain that the divide is permanently established.The arti-
ficial divide should then be surveyed and accurately plotted.
5.1.1.2.3 D R A I N A G E NET A N D C U L T U R A L FEATURES
Large-scaletopographical maps available in a country and showing relief,hydrological

network,swamps,etc.,which are indicated by means of accepted conventional signs,may
serve as initialcartographicalmaterial for representativebasins.
These maps should show not only the permanent hydrological installations but also
the temporary ones. For the estimation of the various characteristics such as length of
the first-orderstreams,stream density,etc.,it is desirable to identify the point at which
a first-orderstream has an incised channel.
Agricultural fields may be indicated by means of well-knownconventional signs for
ploughland; forest by means of signs for conifers and leaf-bearingtrees ; herbaceous
cover by means of a conventional sign for meadows,etc.
The map should also show the location and type of hydraulic structure,towns and
197
villages and other objects which may influence the hydrological characteristicsof the basin.
The basin divide and the location of ali the points where hydrological observations
are made are plotted on this map. T o indicate principal hydrological installations and
instruments (gauging stations, precipitation gauges, observation wells for ground-water
measurements,etc.), conventional signs are recommended. The size of these signs should
be sufficient to make them readable.
5.1.1.3 Construction of maps

5.1.1.3.1 BASE MAPS
Consideration should be given to constructing the maps as a series of two or more bases.
The first m a p can, for example, include the basin divide, the stream system and the
benchmarks. On a copy of this map, man-made features such as roads, railways and
towns can be superimposed. O n another copy of the m a p the contours can be superim-
posed. Thus, three different base maps are available with only a small amount of additional
effort. This variety of base maps would provide a selection on which to superimpose the
maps of geology, soils and land use, and on which to plot the locations of the hydrolo-
gicd installations.
5.1.1.3.2 SIZE OF M A P S
Maps of different sizes (from a relatively large size, useful as a working map, down to
page size) are often useful for publication. Figure 5.1 is an illustration of a page-size
m a p showing the various features that m a y be shown on topographicalmaps. The m a p
is a reconstruction of a larger scale m a p on which 1.54 m contours were shown. Note
that a graphical scale is a necessity if maps are to be reduced or enlarged from the original
size.
5.1.2 Geological and hydrogeological mapping

Surface and subsurface geological maps are essential to the understanding of the factors
involved in the hydrological performance of a basin. Geological investigation should be
reviewed carefully to place them in the proper context of study. Surface mapping should
be completed first, since it provides guidance in designing subsurface exploration. Stand-
ard techniques and procedures should be used wherever possible to aid in the transfer
of results to other areas.
The construction of hydrogeological maps for representative and experimental basins
is essential to characterize principal conditions of subsurface water formation and the
relations between surface and subsurface water.
If possible, the maps should reflect the following data on the hydrological peculiari-
ties of the hydrological regimen:
1. Location and stratification of major aquifers within the river basin and adjacent terri-
tory, if the subsurface basin of these strata goes beyond the limits of the surface divide
of the investigatedbasin.
2. Areas of recharge,movement and discharge of subsurface water, of importance to the
hydrology of the basin, and also areas of possible intensive evaporation of subsurface
water.
3. Karst and karstic-phenomena.
4. Location of free surface and piezometric level of subsurface water for individual
aquifers (characterized best by hydroisohypses and isopiezometric lines).
5. Lithology of the unsaturated zone, thickness and physical qualities of the strata.
198
Daia processing and publicaiion
6.Lithology of rocks in major aquifers connected with streams;data on permeability

coefficientsand water yield capacity of these rocks.
7. Location of principal observationwells for the study of the subsurfacewater regimen
and springs with data on their yield. Hydrogeological maps may give some informa-
tion on the peculiarities of the chemical composition of subsurface water to evaluate
its qualities and effect on the chemical composition of surface water [8,9,19,20,32,
37, 47, 48, 551.
Basin Characteristics
1. Size: 69.2 hectares 0.692 kmz.
2. Range in elevation (approx. mean sea level): from 349 to
380 m.
3. Prevailing land slope: 4 and I per cent.
4. Range in land slopes: from 1 per cent to 12 per cent.
5. Length of principal waterway: 746 m.
6. Average slope of principal waterway: 2.2 per cent.
N W 114 S W i14 Section 17 7. Total number of waterways: 4.
8. Number of hectares per waterway: 17.0.
T 6 N R 2 W of 4lh P M 9. Total length of waterways: 1,509 m.
onq 90" 39'W Lai 43'00'N 10. Drainage density (length of waterways per hectare): 21.8 m/h.
II. Form factor: (AiLa)0.76.
Note: Road ditches drain only the road and are not included
in the waterways.
WISCONSIN
I
LOCATION MAP
,/
o Contour Interval: 3 metres
Fig. 5.1. Topographical map of Fennimore,Wisconsin,catchment IV (United States Depart-

ment of Agriculture) [61].
199
Repvesenrative mid experimental basins
5.1.2.1 Mapping of surface geology

5.1.2.1.1 TECHNIQUES
Prior to any detailed surface mapping,it is advantageous to make a thorough reconnais-

sance of the basin and the surroundingarea. Some form of base map is required for this
phase. Areas where specific types of detailed studies are to be conducted should be deli-
neated during the reconnaissance survey. Cost and efficiency of the field-mappingpro-
gramme may be greatly influenced by the proficiency of the reconnaissance.Procedures
for the use of aerial photography in preliminary surveys are given in the literature
[38,49].
Geological mapping on aerial photographs and/or topographical maps is preferred
because of the speed atwhich the observer can work and the ease with which he can recog-
nize natural and cultural features.Semi-detailedsurveys of certain structural and litho-
logical characteristics can be made from aerial photos prior to a field survey.Although
some field verification is required,a considerable amount of time can be saved by this
procedure.
5.1.2.1.2 FEATURES
Those geological factors which exert specific influences on surface-watermovement can

be broadly classified as lithologicaland structural.The former is broken down into text-
ural components,compositional fabric,and the stratigraphical sequence of rock types.
The latter may be characterized by the attitude of bedding, faults,folds,joints and, to
a somewhat lesser degree, rock cleavage.
The colour ofrock is one ofits most distinctivefeatures,but it is very difficultto evaluate
consistently.For this reason a standard colour chart [65] is recommended and may be
taken in the field as an example to determine precise definition of colour.
When mapping partially indurated or unconsolidated sediments,a handbook of sand-
size particles which represents different textural classes would help to eliminate subjective
estimates and some of the need for laboratory analysis.
Detailed check lists of features to be examined at an outcrop are given in the literature
í171.
In areas of complex structural features,rosette diagrams and equal-area nets [7]are
useful techniques for portraying primary and secondary structures. Figure 5.2-of a
small experimental basis [62]-shows both the lithologicaland structural features.This
map is used to select areas for detailed survey of rock-fracturepatterns. The equal-area
diagrams of the structural joint system on various parts ofthe same basin are shown in
Figure 5.3.A close relationshipis inferred between the fracture patterns and the drainage
net of the small channels on this basin where there is little or no glacial overburden.
This is one instance in which structuralfeatures influence the hydrologicalregimen and
many more may become apparent as the mapping programme progresses.
5.1.2.1.3 CARTOGRAPHY
Surfacefeaturesare most clearly shown on plan-viewgeological maps. Columnarstrati-

graphical sections also serve to show weathering surfaces and typical slopes of exposed
strata. Structural features may be so numerous that extremely large maps are required
for sufficientdetail.Naturalfeaturessuch as the location of perennial streams and springs
should be printed on the ñnal map as particular structural or artesian significance may
be attached to these features.Standard geologicalmapping symbols should be used where
possible.
200
Data processihg and publication
O.
ao
,
I
m
l GR Granite Strike and Dip of Bedding
DGM Gile Mountain Formation: 3 Strike and Dip of Foliation

l Dark and Light Grey Schists
I F
IAMP Waits River Formation 2 Amphibolites
:
'
'
- ...... .. Approximate Contact Boundary
l of Hornblende, Plagioclase and Minor *-----/ Accurate Contact

Amounts of Garnet Interred Crest Line and Directian
DWC Waits River Formation :Quartz Mica ( of Plunge
- - of Willoughby
- Arch
Schist, Gray Quartzite and Quartz
Feldspar Granu I ¡te
..
-----. .
Watershed
I OWP Waits River Formation :Quartz Mica

Schist and Minor Micaceous Quartzite
1 0 1
scale in kilometres
2 3 4
II FIG.5.2. Lithological and structural features, Sleepers River, Vermont [64].
l PO I
Represeniative and experimental basins
FIG.5.3. Equal-area diagrams,,

selected areas, Sleepers River, Vermont.
5.1.2.2 Subsurface mapping

Subsurface investigationson experimental and representativebasins may be broken down
into two distinct categories:hydrogeological,and geologicalengineering.The hydrogeolo-
gical investigation and mapping is designed to define quantitativelythe location of ground
water and its relationship to lithological and structural features.The investigation and
mapping carried out for geological engineering purposes is designed to report on the
suitability of specific sites for the location of structures [2].
5.1.2.2.1 TECHNIQUES
5.1.2.2.1.1 Hydrogeological mapping
A preliminary survey of the available information should include:maps of geology,soil

surveys,geochemical maps and reports,water-wellrecords,structure-test-holelogs and
samples,seismic-shot-holelogs,reports of artesian flow,oil and gas-welllogs and samples
and records of major water users,such as townships,industries,or irrigationcorporations
that derive their water from underground sources.From this information,a hypothetical
model of the physical framework and the flow system can be derived as a guide for the
location of detailed investigationsites and the installation of an effective instrumentation
network (see section 4.4.2.1).
202
Dafa processing and publication
5.1.2.2.1.2 Geological engineering mapping
Techniques used in hydrogeological investigations apply also to geological engineering

mapping. Irrespective of the type of structure to be installed at a site (see section 4.3.4),
the most important mapping techniques to be utilized are the drilling and the seismic
methods.
If bedrock is anticipated at shallow depths, either trenches or test pits should be dug
on the centre lines of the proposed structure. Machine or hand-dug excavations permit
first-orderlogging of profiles, and selection of samples. If shallow bedrock is not antici-
pated, the excavations should be offset from the centreline to avoid damage to the found-
ation of the structure. This technique is used to delineate the rack surface beneath the
main spillway in abutments, and in exploring emergency spillway materials.
Bore holes for site investigations would probably be less deep than those drilled for
regional mapping. The various methods range from simple hand-augerboring to complex
rotary drilling.
In the mapping of structure sites, smaller portable refraction equipment is usually
sufficient for the determination of bedding continuity between borings. W h e n the shock
wave is created by a hammer blow, the instrument is limited in resolution of material to
a depth of about 16 m . Small explosive charges can be used to increase the depth range.
A combination of boring, seismic and resistivity techniques is especially useful in deli-
neating gravel lenses, bedrock and the ground-watertable.
Field-penetration tests are useful in the delineation of various lithological units. The
relative densities of foundation materials are indicated by the number of mallet blows
needed to drive a sampling tube through 30 c m of subsurface material.
A vane attached to the end of a rod is forced into the ground and rotated at a constant
rate by a torque wrench. The torque that is required to turn the rod is an indication of
the shear strength of cohesive soils.
5.1.2.2.2 FEATURES
In addition to the features discussed in section 4.4.2.1, the depth to aquifers and their
confining aquicludes should be recorded. The permeability and porosity of each forma-
tion should be quantitatively described as it is obtained from logging techniques and core
samples. Surfaces of shallow ground-water tables and piezometric contours should be
mapped. The three-dimensionalextent of the ground-water flow system should be deli-
neated as controlled by the basin topography,lithology and structure.
Depth and extent of the alluvial fill in a valley are primary factors in site investigations.
Materials can be classified by conventional geological methods or by the unified soil
classification system [59].A proposed check of possible requirementsin site investigations
would include: cohesive or non-cohesive soil materials, consistency of the soil, density,
moisture content,permeability,coefficient of permeability, consolidation,shear strength,
gradation, texture, rock structure, strength of rock, colour, stratigraphy,type of deposit,
geological age, depth, thickness, continuity, structure and paleontology.
5.1.2.2.3 CARTOGRAPHY
Cartographical representation is basically the same for hydrogeological and geological

engineering features. M e a n sea level is used as a datum plane for stratigraphical and
structural cross-sections.Panel or fence diagrams can be used to show the relationship
between exploratory wells not located on a straight line. This diagram does not usually
have consistent horizontal or vertical scales. The diagram m a y be structural or strati-
graphical, and is based on the columnar section determined by coring or well logging.
The isometric projection as a means of representing certain stratigraphical variations that
203
Representarive aiid experimental basins
are not apparent on the conventionally oriented panel diagram is described in the liter-
ature [il]. Well logs can also be portrayed graphically.
Structure maps, isopach maps, facies maps, geophysical maps and geochemical maps
are useful in portraying subsurface features of interest to research on ground water.
Methods are described in the literature.
Statisticalmodels for use in trend surface analysisand various other aspects of statistical
sampling or analysis of subsurfacedata may be important [28].
Figure 5.4and Table 5.1show the methods which may be used to illustrate graphically
the hydrogeologicalconditionsin a typical basin study.
Standard hydrogeological mapping symbols have been prepared by Unesco [57] and
are suggested for use on representativeand experimental basins where applicable.
5.1.3 Pedological mapping

A broad-scalereconnaissancesurvey,including a generalized soilmap,is usually required
for the selection of representative or experimental basins. In addition,research on experi-
mental basins in particular requires an intensive physical inventory of the selected area,
including detailed knowledge of the areal distribution of the various soils,their inherent
characteristicsand their relative positions on the landscape.This informationis commonly
displayed on a large-scalesoil map, accompanied by a legend and a bulletin or soil-
survey report. Several manuals giving detailed instructionsfor soil surveys and soil-map
preparation have been published [40,51, 53, 55, 581.
Single-valuemaps showing magnitudes of single soilmorphological characteristicsmay
be useful. For instance,contour maps showing the depths of the A,or any other horizon,
may be of great value in infiltration and water-balance studies on experimental basins
(seealsosections2.2.3and6.1.1.3).
5.1.3.1 Techniques
The soil map should be prepared by a soil surveyor experienced in local soil mapping
and familiar with the current concepts of soil genesis and classification.Sufficient time
must be allotted to study the area and to take sufficient samples for field examination
and laboratory analysis.Mapping units are based upon observation of the landscape and
examination of the texture,structure,colour and consistency of samples obtained using
a hand auger or spade to a depth of approximately 1 m.
Soil-identificationsymbols,soil boundaries, streams and drainage-ways,roads,houses
and pertinent reference points may be plotted on the aerial photo or an overlay as shown
in Figure 5.5,a soil map on an aerial photograph of USDA-ARS experimental basin
No.1, Fennimore,Wisconsin.
5.1.3.2 Features of a soil map
The base for the finished soil map in Figure 5.6may be a standard topographicalmap of
appropriatescale.The boundaries of each significantsoiltaxonomic unit should be clearly
delineated on this base map. Within the boundary of each such mapping unit, a symbol
or symbols should be printed designating the type of soil,the surface texture,the degree
of erosion, and the mean land slope. Soil boundaries and soil symbols should be in
permanent black ink, complete and legible. Boundaries between soil units may be
solid lines;boundaries between slopeand erosion phases may be dashed and dotted lines
respectively.
204
Data processing and publicution
Stratigraphical distribution of water quantity and quality
SINOSTONL.
"El" LITTLE
W E
4
- 0 4 8
kilometres
1 2
UPPER PART ..
~
FIG.5.4. Geological map, Washita River, Oklahoma.
205
I Data processing and publication
FIG.5.5. Aerial photograph used for mapping soils, Fennimore, Wisconsin [60].
5.1.4 Vegetation maps

The nature and composition of the vegetative cover of a basin regulate its hydrological
characteristicsand therefore affect the components of its water balance.Detailed vegeta-
tion surveys must therefore be completed for all research basins where theories are to
be evolved which correlatechanges in basin cover with changes in the hydrological regi-
men.T o ensurethataccurateinformationis availableforthe basin,as a whole,it is essential
that aerial photographs be taken at the commencement of each basin study. Detailed
ground surveys of vegetation and land use should be carried out,as outlined below, but
additional photographs should be obtained at intervals of not more than five years or
at different intervals depending on type of vegetation and following any changes in land
use (see section 5.1.6).For detailed vegetation studies,see section4.6.
5.1.4.1 Vegetation surveys

Vegetation surveys should provide at least information on the composition and density
(in forests also the age), and the relation between vegetation and its milieu.
(a) Composition:the survey intensity will depend on the basin size and the study object-
ives.In range-vegetationareas,standard range-surveytechniques are adequate to deter-
mine the botanical composition and forage weights may be readily obtained by clipping.
In forest-vegetationareas, existing forest-survey methods are adequate. These ground-
survey resultsshould be combined with vegetation mapping obtained from aerial surveys.
207
Representative und expeviniental basins
I kilometre
Legend
Symbol Soil fype Sìopc percenfagc Erosion
ChB
JuB
Chaseburg silt loam
Judson silt loam
3-6
3-10
-
-
DsF2 Dubuque silt loam 20-30 moderate
DsE2 Dubuque silt loam 15-20 moderate
DtC2 Dubuque silt loam (deep) 6-10 moderate
DuD3 Dubuque soils 10-15 severe
DuE3 Dubuque soils 15-20 severe
DvC3 Dubuque soils (deep) 6-10 severe
DvD3 Dubuque soils (deep) 10-15 severe
DbD2 Dodgeville silt loam 10-15 moderate
DcC2 Dodgeville silt loam (deep) 6-10 moderate
DdC3 Dodgeville soils 6-10 severe
DdD3 Dodgeville soils 10-15 severe
DeC3 Dodgeville soils 6-10 severe
TaA
TaB2
T a m a silt loam
T a m a silt loam
0-2
2-6
-
moderate
TaB3 T a m a silt loam 2-6 severe
Tac2 T a m a silt loam 6-10 moderate
Tac3 T a m a silt loam 6-10 severe
FIG.5.6. Soil map of experimental basin, Fennimore, Wisconsin. Taken from an aerial
photograph (redrawn from Grant County Soil Survey, USDA, SCS, 1961,plate 46).
(b) Density and age should be determined by ground survey,in conjunction with aerial
photography.The information is most economically inventoried by aerial photography.
A variety of scales is available,including 70 mm photography, and these are in many
cases sufficient to show micro-arrangementon a sample basis.
(c) Association with environment: the mapping of the vegetative cover in relation to
the basin’s soils and physiography is often useful and illuminating.A clearer impression
of the effects of environment on cover development may be obtained by using overlays
of vegetative cover,aspect,elevation,physiography,surficial geology,etc.
In most research basins,periodic resurveys,at aboutfive-yearintervals,will be required
to evaluatechanges in conditionsand &rendclusters in range vegetation.As forestvegeta-
208
tion is subject to insect and disease attack (causing defoliation,change in composition

and death of individuals,groups or whole stands) annual appraisals of insect and disease
conditions are required,with mapping of affected areas where appropriate.
5.1.5 Geomorphological mapping

The geomorphologicalstructureis one of the factorsthat affects the hydrologicalregimen
of representative and experimental basins.
Geomorphologicalcharacteristicsof a basin may be representedby means of a geomor-
phological map; this map should give the basin division into regions according to its
principal geomorphologicalcharacteristics of the landscape,reflecting the genesis of va-
rious relief forms. The map should first of all show such effects ofwater (water erosion),
as ravine formation,slumps,sheet erosion,etc. (see section 4.9.2).
In addition,the map should reflect the effect of ancient and contemporary glaciers
(frontal moraine, eskers, kames, etc.), karst, seas and lakes, wind effect (various types
of dunes), mountain formationand permafrost (glacial karst).
If possible,the map should give data on the intensity of exogenic processes (ravine-
formation rate, sheet erosion,eolation,etc.).
5.1.6 Land-inventory mapping

The many and varied effects of vegetation,land use and culturalpracticeson the potential
use and hydrology of an area demand an accurateand detailed assessment ofthese factors
in hydrologicalresearch.These are perhaps the most difficultfeaturesof a basin to charac-
terize;thus,as much information as practicable should be obtained on present vegetation
composition and distribution,conservation practices and historical changes. Moreover,
the economics of the region and the original division settlement of the land play an impor-
tant part in understanding the arrangement and distributionof present-dayland use and
cultural practices. Land-inventory mapping is not required for experimental basins or
for small representative basins which are used for studying the effects of natural changes
on the hydrologicalregimen or for fundamentalresearch only,For such basins detailed
vegetation studiesare required as given in sections4.6and 5.1.4.
5.1.6.1 Techniques
Land use is usually mapped at the same time as the soils and to avoid confusion of
boundaries and symbols,the land-usemap is often drawn on an overlay to the soil map.
This is necessarysince soil and land use so often coincide and overlap.
5.1.6.2 Features
Classes of land use may be adapted from a publication such as the Soil Survey Manual
of the USDA (1951), and each class may be given a symbol (e.g.,L for cropland,O for
orchards), though these may prove to be insufficiently detailed.
5.1.6.3 Cartography
If drawn on soil maps, land-use boundaries and symbols are conventionally drawn
in red ink,often with broken lines.The land-useunits may be coloured and the colours
and symbols keyed to a legend which gives detailed descriptions of each land cover or
cultural practice.
The land-usemap should be revised periodically to record changes in land use,forest
209
Repuesentalive and experimental basins
cuttings,constructionof conservationworks and evolutionofimproved cultural practices.

In addition,records-monthly or even more frequent-should be maintained of crop
cover conditions,noting stand density and vigour, height and percentage cover of the
vegetation.Dates of significantclimatical events and their effects on the vegetation should
be noted (especially killing frosts and erosive storms).
5.1.7 Aerial and terrestrial photography

Aerial and terrestrial photography provide especially valuable data for the investigations
on representativeand experimental basins.
Aerial photos,because of the expense involved in obtaining them,are usually repeated
at periodic intervalsof at least a few years.They provide reliable data on the topography
of the basins investigated, their geomorphological and pedological characteristics and
determine the location of the occurrence of specific natural phenomena.
Terrestrial photos are taken more frequently and are generally used for a qualitative
recordingoflocalconditionson smallbasins and for a documentationof a given hydrolog-
ical phenomenon.
Where possible,two levels of aerial photography are desirable. The high flight covering
the entire basin provides a small-scalesurvey;the low flight provides a large-scalesurvey
of individual areas of the basin showing the most interesting hydrological phenomena.
5 1.7.1 Terrestrial photography

Sequentialphotography is frequently employed to map or record changes on small study
areas. This method of inventory is much less expensive than aerial photography and
easier to acquire.Like aerial photography,stereoscopic coverage is a great aid in inter-
pretation (see Fig. 5.7).Ideally,a 60 per cent overlap of photographs will provide the
optimum stereo vision. This can be accomplished by using an instrument known as the
stereo-tachapparatus (which can be attached to a 35 mm camera), or any other conven-
tional method.
FIG.5.7. Terrestrial stereogram of gravel deposit above weir, Switzer Creek, New York.
210
If sequential terrestrial photography is desired,a marker should be set in the grourid

to facilitate relocation at any time in the future. The subject, date, time of exposure,
direction of view and any other pertinent information should be recorded and the photo-
graphs should be catalogued.
5.1.7.2 Aerial photography

Geological,pedological and land-usefeatures are easily identified on aerial photographs.
On small open basins in plains (without forest), where the determination of the basin
divide by the usual methods does not give the required accuracy, aerial-surveymethods
are of great importance. If an aerial survey is made after heavy storm3 or snowmelt,it
gives an exact determination of the basin divide if the flow (overland and/or micro-chan-
nel flow) pattern is studizd.
Aerial-surveymethods are also useful for a determination of high-watermarks and
boundaries of submerged areas during inundation;water areas affected by aquatic vege-
tation; areas where vegetation damage occurs owing to fire,drought, floods,insects or
disease;and other similar features.
A n uncontrolled mosaic, the lowest-orderphotomap, is often useful for orientation
purposes.It shows qualitative details and is the simplest to construct. No true distances
or directions can be measured on the uncontrolled mosaic. The controlled mosaic is the
next order photomap. It is developed from photographs that have been rectified, i.e.,
distortion of all types is corrected as much as possible [3, 41.
5.1.8 Cartographical description of physiographical

characteristics
The cartographicalmaterial described above may be used for a physiographical descrip-
tion of representativeand experimental basins. Descriptionscould,for example,include:
The geographical location of the basin;country,region,co-ordinates;
The basin shape and its aspect to the four cardinal points;
Basin-reliefcharacteristics with a description of the origin of its morphology;mean ele-
vation above sea level,slope exposure;definition of the basin divide (this may reveal
possible inaccuracies in the determination of the basin area) ;
Geomorphologicalcharacteristics of the basin (see section 6.1.1.2);
Distribution of soils and subsoils (sands, sandy loams, loams, clays, etc.), and soiì-
salinity rate,etc.;
Characteristics of the vegetation (forest belts, composition of forests,age,density, etc.)
and the distributionof various vegetation types on the basin;
Characteristics of principal aquifers (distribution of individual strata over the basin ter-
ritory;their location in respectto one another and the drainage depression;cornpoition
and thickness of rocks containing water ; permeability and water-yield coefficients of
these rocks,similitude of surface and subsurface basins;location of free water surface
and piezometric level; lithology,thickness and physical qualities of soils and subsoils
in the unsaturated zone;karst,springs etc.);
Climatic characteristics of the basin with normal annual,seasonal and monthly przcipi-
tation;air temperature and water equivalent of the snow pack;
Hydrological characteristics of the stream : normal flow and its annual distribution,run-
off fluctuations,maximum and minimum discharges; depletion and freezingof streams;
lateral and subsurfaceerosion; peculiaritiesconnected with man’seconomical activities
(water intakefor irrigationand other economic needs,drainage,snow retardation,etc.) ;
Year and month of start ofobservations.
211
Representative and experimental basitu
Maps, hydrogeological and pedological profiles, described in sections 5.1.2and 5.1.3,

should be part of the physiographicaldescription of a basin [34].
Because continuous observations are made on representativeand experimental basins,
inventoriesas given above should be kept up to date.
5.1.9 Inventory of observations and research made on

representative and experimental basins
For the most efficient use of any data collection,an inventory of natural and/or cultural
changes in a basin should be made. This inventory should be made for every representa-
tive and experimental basin in a special register and give the following(see sections 1.4.3
and 1.4.4):
Name of basin,basin area and geographicallocation;
Topic of research;
Date of start of observations;
Informationon the observationpoints and instrumentation; dates of beginning and end
of observations with these instruments;
Changes in the observational programme;
Natural and cultural changes on the basin; forest planting and cutting;ploughing of
waste lands;swamp reclamation;road building, etc.;
Information on significant and unusual hydrological phenomena (catastrophic floods,
mudflows,ice dams,etc.);
Summariesof the investigationscarried out for each year;
Date of completion of observations;
Final observationalresults.
5.2 Recommendations on routine data-processing

methods
Hydrological data are processed to reduce observed data to a suitable form for analysis,
publication and storage. For representative and experimental basins more detail and
greater accuracy is required in data processing than is normally the case for gauging
stations for water-managementpurposes [71].
Itis recommendedthat the data-processingmethods discussed in this chapter be follow-
ed. Whichever methods are used, however, it is important to keep in mind the ultimate
automation of all data processingsince hydrologicaldata,per se,produces so many digits
that, unless data processing is partly or wholly automated,it would be very difficult,
because of the magnitude of manual processing,to obtain full value from all the recorded
data.For some countries which have no computer facilitiesexcellent research work can
be done with manual computation.Moreover,for research aspects where parameters are
ill-defined,mechanical computation may not be necessary initially.
Data processingconsistsof three principalstages:(a) calculation of the observed values
for individualpoints on a basin;(b) calculation of mean areal hydrologicalcharacteristics
of a basin;(c) tabulation of observationaldata on forms suitable for use in analysis and
for publication.
Data processing for individualpoints on a basin could include the following:
1. Initial processing:estimation of corrections for instrumentalreadings; initial proces-
sing of field records;tabulation of observed data;plotting of continuous graphs for
individual variables.
212
Data processiiig and publicarion
2. Checking of observational results.This refers to the estimation of errors in measure-

ment,malfunctioning of instrumentation,gaps in observations and evaluation of data
reliability.Such checks may be made by correlationof data between severalpoints and
by a comparison of temporal variations of the observed variables and possible causes.
All cases of considerablenonconformityin the data may be checked with climatological
data (precipitation,temperature,etc.). All doubtful observations should be repeated
if possible (for instance,dischargeat a given water level,water equivalent of the snow
pack in the absence of precipitationor thaw,etc.).
3. Tabulation of data for different time intervals (e.g.,month, year). A clear indication
should be given if any data of insufficientaccuracy is included.
The calculation of average basin values should be made according to methods described
below. The mean values of individual variables (precipitation,flow, evaporation,etc.),
should be compared for each basin so that possible large errors are detected.
Finally data are tabulated on forms for international exchange as recommended by
Unesco (see section 5.10).
Note that the final accuracy of the data depends on the observationaland processing
errors. An objective evaluation of the final error may be made by consideringallpossible
systematic and random errors. Systematic errors may be detected easily by carefully
checking the observational data. The estimation of random errors should be made by
statistical means. This may be difficult to apply in some cases because of short records,
infrequent occurrence of given phenomena or because of interrelations between the hy-
drologicalvariableconsideredand basin characteristics.When comparingvariousmethods
of areal averaging,the water-balancemethod may be a useful checking medium in some
cases.Further research into an improvementof the relative errors associated with hydro-
logical data is urgently needed [15,221.
5.3 Climatic data

5.3.1 Precipitation
5.3.1.1 Consistency of records
No record of precipitation should be assumed to be consistent throughout,unless it is

known to be devoid ofchangesinexposure,observer,location and equipment.The follow-
ing are some common sources of error and methods of detecting them.
(a) Improper exposure;if a gauge is not installed with the correct exposure,speed and
direction of wind as well as catch of gauge are likely to be abnormal.The abnormality
can be identified and evaluated by comparing catches of adjacent gauges between windy
and non-windyperiods,certainof the gauges being properly exposed (see (b) below).
(b) Change in location,change of gauge,etc.:a change in location of a station may
divide a precipitation record into parts that are not consistent one with another. Such
errors may be very large and could also result if the gauge is replaced by a different type,
if a windshield is added or removed, etc. Station history should always be carefully
reviewed and the record should be compared with those from nearby stations considered
to be in the same environment,either by double-massplotting [46]or by the ratio method,
which compares the records for various periods of time to determine whether the ratio
is consistent.If precipitation values are considerably different from winter to summer,
it is desirable to test consistency for each season separately.
Ifinconsistenciesare apparent,appropriateadjustmentsaresometimeswarranted.How-
ever,caution should be used in such adjustment and the safest approachis to adjust only
if presumptive causes coincide with dates of inconsistencies.In the case of a long record
at a station documented as having had changes in location,and if double-mass analysis
21 3
shows each to be consistent within itself, the record for each location should be treated
separately.
(c) Wetting: the correction value for wetting may be estimated by means of frequent
comparison of a definite volume of water poured into a dry bucket and the amount of
water poured out of this bucket some minutes later. The mean correction is obtained by
taking the average for a large number of tests.
(d) Evaporation from the bucket: corrections for evaporation of precipitation from the
bucket are determined by special experimental data.
(e) Height of gauge: in some cases gauges are installed with their orifices some height
above the ground and the precipitation recorded may be incorrect because of wind effects.
The correct precipitation may be estimated by comparing the data with gauges with
their orifices level with the land surface.
For a thorough analysis of observed data, readings should be compared with those of
nearby gauges. Such comparisons m a y aid in filling in missing records or in detecting
bucket leakage and other malfunctioning of the gauge,irregularitiesin observational dates
and incorrect information concerning the precipitation type (solid, liquid). It is essential
to take into account the c o m m o n type of precipitation that may occur and weather con-
ditions.
For recording raingauges the data should be compared with those of the check gauge
and with the time of the start and end of any storms that may have been recorded by other
recording gauges.
5.3.1.2 Point rainfall record

Each representative and experimental basin should have at least one recording precipita-
tion (see section 4.2.1). Data (hereafter referred to as rainfall) from such gauges should
be tabulated in sufficient detail as required for subsequent analysis. Figure 5.8 shows a
possible method for tabulating time and rainfall depths.
5.3.1.3 Routine methods of determining mean basin rainfall

For research on representative and experimental basins, a first requirement is mean basin
rainfall for short and long periods. Arithmetic means, isohyetal and Thiessen diagrams
are used. For details, refer to the literature [23,681.To obtain some advantage from the
isohyetal method, the Thiessen polygons can be adjusted to fit along typical isohyets.
T h e method of squares is also useful. The basin area is divided into a number of equal
squares and the rainfall depth is shown inside each square as observed at a raingauge
occurring within the square. If there are several points within the square,the arithmetic
mean of the values is taken. The values for blank squares are obtained by means of inter-
polation between two neighbouring rainfall points. The mean depth of rainfall of the
basin is obtained by summing the amount of rainfall in all squares and dividing by the
number of squares.The correctness of the calculation may be checked by repeating it
using the method of different squares. The deviation of the results of any such calcula-
tions should not exceed 5-10per cent.
If only recording gauges are installed in a basin,mean basin rainfalls m a y be calculated
easily for any period using data as calculated in section 5.3.1.2.
Jf a mixture of gauges (manual, storage and recording) has been installed, rainfall
intensities may have to be calculated for the manual and storage gauges. These have to
be derived by reference to the mean basin rainfalls (or intensities) derived from all record-
ing gauges in the basin, or from one or more recording gauges which are thought to
correlate most closely to the manual or storage gauge under consideration.Detailed proce-
dures are given in the literature [5q.
214
Zrnins 5mins 10rnins i5rnins 2Omins 30mins 1 nr 2hr 6nr 12hr
DEPT H m m
5 7 8 9 12 16 27
I INTENSITY
mml nr 60 *L 32 27 24 i6 13.5
5.3.1.4 Other methods of determining mean basin rainfall

In steep terrain which is subject to orographic precipitation,an alternative method to
calculate mean basin rainfall is to establish a precipitation-elevationcorrelation and to
group the elevation in a number of classes (e.g.,200 m intervals). From a contour map
of the basin, the area in each class of elevation is planimetered and the corresponding
rainfall computed.
Another approach is to describe the pattern of gauge catch by a ‘filteredinput field’
capable of mathematical expression [6].
The field is produced by a process of smoothing point data in time and space to reveal
hydrologically useful trends in a storm.Curve and surface fitting are based on the three-
parameter gamma distribution and low-order non-orthogononalpolynomials. Studies
such as these may contribute significantly to a more correct approach to mean basin
rainfall [141.
The advent of radar has opened up the possibility of establishing more accurately
mean basin rainfalls. On larger basins in particular the extent of storms could be deter-
mined to provide a basis for modifying the Thiessen polygons more accurately.
A simplewater balance (precipitation = evapotranspiration + run-off)can give a rea-
sonable first check on the validity of precipitation measurement.Evapotranspiration
values can be calculated by methods given in the literature[54],and the balance calculated
on an annual basis. If the precipitation is not sufficient to balance the equation for the
water balance, this can be explained by an inadequate precipitation net, water entering
the basin from an external source,or depletion oflong-termstorage.
5.3.1.5 Other methods of processing precipitation data

For some studiesin representativeand experimental basins other methods of processing
precipitationdata such as depth-durationanalysis and intensity-frequency-duration
stud-
ies are useful. For detailed procedures refer to the literature[68].
5.3.2 Snow cover

The final objectives of routine data processing of snow observations on representative
and experimental basins are as follows:
1. The determination of the mean water equivalent of snow, ice and free (snowmelt)
water under the snow pack for an entire basin and individual parts of a basin.
2. The calculation of the percentage of the coverage of a basin and its individuai parts
with snow,ice crust and water.
3. The determination of an integral curve of the distribution of the water equivalent in
snow on the surface of a basin.
For basins with a relatively homogeneous vegetative cover, the mean water equivalent
of the snow pack may be estimated as the product of mean snow cover height and mean
snow density.If there are forests and open fields in the basin,the mean water equivalent
of the snow pack may be estimated as the mean weighted value from the snow storage in
the basin,taking into account the length ofsnow coursescrossing variousphysiographical
conditionsor consideringareas occupied by various types of vegetation (forest,grassland,
etc.). The mean value of the thickness of the ice crust and the depth of snowmelt water
under the snow cover is estimated similarly.The mean water equivalent of the ice crust
is estimated by means of multiplying its mean thickness by the density;for crystalline ice
this density may be assumed to be 0.9,and for snow ice with air bubbles in it 0.8.
The mean depth of snowmeltwater under a snow pack on the surface of basin should
be estimated by dividing the total depth of snowmelt water observed in snow density
21 6
Data processing arid publication
measurementpoints into the total number of these points.The mean quantityof snowmelt
water under the snow pack should be estimated by multiplying the mean depth of water
by the porosity of the snow saturated by water (1 -ds), where ds is the mean density of
snow not saturated by water.
The total mean water depth on the surface of a basin (or any part of a basin) is estima-
ted by means of summing the corresponding mean amount of water in the form of snow
ice and free (snowmelt) water under the snow pack. When so-called ‘landscape’snow,
surveys are carried out the procedure is as follows.
For every landscape snow course,the mean depth and density of the snow cover,the
water equivalentof the snow pack,the mean depth,the water equivalentof the ice crust,
the water equivalentof the snow pack saturated with water and the total amount of water
on the soil surface should be estimated.
For a small basin (up to 100-200 kmz)the total mean amount of water on the soil
surface is estimated as the mean weighted value of the total water accumulationon indivi-
dual landscapes,taking into account their areas in the basin.
The percentage of landscape elements in the basin may be determined with the aid of
large-scalemaps and plans.
For a large basin, the total mean water content on the soil surface is estimated as the
mean weighted value from the total mean amount of water accumulated on individual
parts of the basin, taking into account their areas. If individual parts of a basin have
approximately equal areas,then the total mean amount of water for the basin is estimated
as the arithmetic average of the total mean amount of water, accumulated on the individ-
ual parts.
The coverage of snow, ice crust and free water for a basin on individual landscape
elements is determined by defining the ratio between the number of points with snow
(ice,water) and the total number of measuring points and is usually expressed as a per-
centage.
The distribution of the areal snow storage required for the analysis of spring floods
and for the calculationof snowmelt intensity may be given in the form of a snow storage
frequency curve. Such a curve is made by plotting areal snow survey data obtained late
in winter (before spring snowmelt) as a frequency curve. The frequency curve shows
which part of the basin has an excessive water equivalentofthe snow pack.
Current analysis of snow survey data is directed at plotting and comparing combined
chronological graphs of temporal variations of major snow-covercharacteristics and cli-
matological factors (precipitation,air temperature).
Snow-surveydata may be tabulated.Such tables could include snow-surveydates,data
on mean depth and density of the snow pack (for the whole basin and for individual
landscape elements), data on the water equivalent of the snow pack,depth of snowmelt
water on the soil surface,the percentage of snow coverage of the basin, the number of
observation points and frequency curves for water equivalent in the snow pack before
snowmelt occurs.
In rugged terrain average basin values are frequently meaningless and studies will have
to be carried out using index values only (see section 4.2.2).
5.3.2.1 Errors in snow data

A n ideally consistent snow-survey record should result from measurements at regular
intervals with no changein site,samplingpoints,equipment,or technique of observation.
All too frequently,however,changes have been made without adequate documentation.
Therefore,it should not be automatically assumed that snow records are homogeneous
and all records should be treated as though biased by some undocumented change along
the course.
Physicalchanges at a snow course site may be natural or man-made.Natural changes
217
consist mainly of increasesin tree canopy or growth of brush which changesthe intercep-
tion of snow and modifies the movement of wind.
Man-made changes include logging and construction.
Errors may be introduced by a failure to re-establishthe course on the same site on
successive surveys or to change the number of sampling points or by observationalin-
accuracy.
5.3.2.2 Tests for consistency

Standard statistical methods or the double mass curve technique may be used [46].An-
other method is to plot the historical records of all surveys at each sampling point of
a snow course. This may reveal a pattern characteristic of that course. Poor or missed
samples may usually be estimated from such a plot. Experience has shown that at any
given time the densities of individual samples along a snow course do not vary more
than a regular percentage;therefore,a comparison of densities may yield a clue to an
inconsistent sample.
Where resurveys are not possible, some questionable snow-coursesurveys may be
adjusted by direct correlation with records from other courses surveyed at the same time.
Very high correlations may occur between adjacent snow courses.
5.3.3 Interception of precipitation by vegetation

Interception results are normally reported as percentages of gross precipitation or as
regression equations. Percentage values are informative for comparing losses between
stands when rainfallamounts and distributionsare similar.However,percentagesderived
in one area are inapplicable to other areas having different rainfall patterns, because
lossesvary with storm size [26].It is better to report throughfalland stem flow as functions
of gross rainfall,i.e.,regressionequations.Regression coefficientscan easily be compared
independently of local rainfall patterns.As more ecosystem data become available,analy-
sis by various statistical or mathematicalmodels can be pursued (see section 6.2).
5.3.4 Evaporation
5.3.4.1 Water-balance method
The simplest method for estimating evaporation for a lake or reservoir is by the water
balance,which can be expressed by the water-balanceequation:
E=P+Z-OfAS (1)
where :
E = lake evaporation;
P = precipitation on lake surface;
Z = inflow,by streams of effluent ground water;
O = outflow,by streams or influent ground water;
AS = change in storage of lake.
The reliability of the method depends on the accuracy to which each item of the balance
can be measured or estimated and also on the magnitude of each item in relation to the
evaporation. For each of the water-balance items, the acceptable percentage error di-
minishes in approximately inverseproportion to the relative magnitude of that particular
item.Thus,unless evaporationis of aboutthe same magnitude on the largest ofthe water-
balance items,its determination,by the water-balancemethod, may not be reliable.
218
Water-balance values are measured by methods described in Chapter 4. In general,

such values can be considered more reliable on a percentage basis, over intervals of a
month or longer rather than over a day or less (it is almost impossible to calculate water
balances for a period of a day or less with any degree of accuracy).
The water-balancemethod may be used for natural basins;great difficulties will,how-
ever, be experienced in determining changes in the various storages and, naturally, the
method is suitable only for leak-proofbasins (see sections 2.3 and 2.4)[68].
5.3.4.2 Pan evaporation

Observed data for evaporation pans should be reviewed discriminately in case any are
unrealistic.Computed evaporation for days of heavy rain may be greatly in error because
of overflow of the pan,because of catch differences between pan and precipitation gauge
or because of splash.Strong winds may also blow water out of the pan.
Pan evaporation data,whether measured manually or continuously should be converted
to ‘freewater evaporation’,which is the evaporation loss from extensive water surfaces
with negligible storage of heat from one week or month to the next. In large reservoirs
and lakes,significant heat storage effects usually require adjustments to be made to free
water evaporation.
Evaporation between observationaldates is estimated by the equation:
E = Ah+P (2)
where :
A h = hi -hz (the difference in level between the correctedprevious (hi) and the current
(hz) readings)
P = amount of precipitation.
If during a severe storm some water was poured out from the pan to prevent overflow,
the evaporation value for the period between observational dates is estimated by the
equation:
E=h+P-AH (3)
where AH = the depth of poured-outwater.

In cases where water-levelfluctuations in the pan are estimated by means of water
which was added (or poured out) to equate the water surface to the edge of the needle,
the evaporation value is calculated by the equation:
E = P f AH (4)
where AH is positive when water has been added and negative when water has been
poured out.
Current analysis of observational data consists of the determination of relations be-
tween temporal variations of evaporation and hydrological and climatological character-
istics that may cause variationsin evaporation.Simultaneousobservationsof evaporation
pans and evaporation tanks give opportunity for comparing results.
Normally,ten-dayand monthly values of evaporation from water surfaces are esti-
mated together with mean values of causative factors.
Methods for converting pan evaporation data to free water evaporation are given in
the literature [31,43, 68,731.
5.3.4.3 Energy-balance method

The energy-balancemethod is discussed in section 5.3.5.
219
Represeii~ativeand experimental basins
5.3.4.4 Aerodynamic method

Many techniquesare availablefor computingevaporation by aerodynamic (mass-transfer)
methods [24,631.
One technique is based on an empirical equation of the form :
E = cil (eo -e,) (5’1
where :
E = evaporation;
c = a coefficient;
u = average wind speed;
eo = vapour pressure of saturated air at water-surfacetemperature;
ea = vapour pressure of ambient air.
This technique has been named the mass-transfermethod and its principal advantage is
that it eliminates advected energy and energy-storagechange as independentfactors.The
effects of these two factors appear in the water-surfacetemperature and so in the eo
factor of equation (5). Accordingly, potential causes of seasonal bias seem to be fewer
in the mass-transfermethod than in either the water-balanceor energy-balancemethod.
The mass-transiermethod, therefore,seems the most reliable of the three methods for
determining evaporation by months or shorter periods. For detailed computations,refer
to the literature [68].
5.3.4.5 Empirical equations

One method for computing evaporation from meteorological factors is based on a com-
bination of the aerodynamic and energy-balanceapproaches.
In its general form the so-calledPenman equation is :
AHfEay
Fo=
AtY
where :
EO= the estimated evaporation from a free water surface;
A = the slope of the saturation vapour pressure curve at ambient temperature Ta;
N = the net radiation;
y = the ‘constant’in the wet- and dry-bulb psychrometric equation (one that varies
with altitude);
E,- a parameter derived from wind speed u measured at 2 m and ea -ed (where ea is
~~~
the saturation vapour pressure at temperature Ta,and e d is the actual vapour

pressure at the same temperature or the saturation deficit at ambient air temper-
ature T, and dew point Ta).
Using Penman’soriginal units and terminology :
H=P,(I-~)0.29 cos e+o.a-n -01: [0.56-0.09Ved]) (o.i0+0.90~
( N ”> (7)
where :
R, = themean extra-terrestrialradiationexpressed in equivalentmm/day of evaporation;
r = the reflection coefficient or albedo of water;
= the Iatitude of the station;
n/N=the ratio of actual to possible hours of sunshine;
u = the Stephan Boltzmann constant (also in equivalentmm/day of evaporation);
Ta= the mean air temperature (absolute);
e d = the actual vapour pressure of the air in millimetres of mercury.
E, = (a + bu)(ea -ed) (8)
220
Daia processing and publication
where u is wind speed and ea-ea is determined from daily mean temperature and
vapour pressure.Wind observations may be adjusted to the standard height of 2 m from:
Heighr (m)
3 4 5 10 15 20
Ratio of wind speed at 2 m to given height 0.94 0.92 0.89 0.78 0.72 0.68
The value of the potential evapotranspiration (Et)from short turf may be determined
from :
Et =fEo (9)
where fis a seasonal factor which, for south-eastEngland,varies between 0.6and 0.8.
There have been a number of amendments to the original Penman method, mainly
concerned with the radiation and wind speed terms and with the effect ofaltitude.
If for the reflection coefficient r a value which applies to the actual surface in question
is used,the value for Et can be calculated directly without the intermediate step of calcu-
lating EO. When measurements of either solar radiation or net radiation are available,
they facilitate determination of H.In the original equation Penman showed that in the
British system of units:
Ea = 0.35(~+ bu)(ea-ed) (10)
where a = 1.0and b = 0.01.After the Lake Heffner experiments,Penman altered the

value for a to 0.5,but more recently he has reverted to his originalvalue of 1.O.
It is recommended that WMO guides which deal with this subject and recommenda-
tions from the I H D working group on the world water balance be used [67, 681.
Should the general form of the Penman equation be written as :
A Y
E--
'- ASY Ho + o+y EA
then the functions d/(d + y) and y/(d -t y) can be seen as weighting factorswhich are
functionsofmean air temperature and atmospheric pressure. Their variations with tem-
perature and altitude are shown in Table 5.2 [35].
TABLE5.2.
AIíAtv) YI@ 4-Y)
Ta PC)
Sea level 3000111 Sea level 3000111
10 0.545 0.633 0.455 0.367

15 0.615 0.698 0.385 0.302
20 0.678 0.754 0.322 0.246
25 0.733 0.800 0.267 0.200
30 0.780 0.836 0.220 0.164
5.3.4.6 Evaporation from soil and snow cover

Processing observational data on the evaporation froni soil (obtained by evaporimeters)
includesthe calculation of evapotranspiration,evaporation from soilunder the vegetative
cover and transpiration;depth of precipitation,the layer of water percolated through soil
monolithsinto the evaporationpans ;soil moisture and water equivalentin soilmonoliths;
data on the evolution stages and state of the vegetation.
221
For snow, evaporation data and the temperature of the snow surface are evaluated.
Soil evaporation is calculated by formulas given in section 4.2.4.2.The depth of the
snow evaporated is estimated from the weight of evaporated snow and the surface area
of the evaporation pan.
Current analyses of observational data on evaporation from soil and snow include
comparing evaporation data estimated by various devices for a similar time interval and
determining relations between temporalchronologicalvariations ofevaporation and tem-
poral variation of the major factors causing them.
In the U.S.S.R. evaporation from the soil,estimated by various methods,is tabulated
showing periods of observation,evaporation from under the vegetation,evapotranspi-
ration,transpiration,precipitation for individual time intervals and total values of these
items.
Tables of evaporation from snow may contain semi-diurnaland diurnal values of
evaporation and the results of regular observations of temperature,absolute and relative
air humidity,wind velocity and the temperature of the snow surface.
The estimation of mean depth of evaporation from representative or experimental
basins may be based on evaporation data from individual areas (forest,grass,etc.). Given
all these data the average areal depth of evaporation for different time intervals (month,
year) is estimated as a mean value, taking into account the size of the individual areas,
expressed as parts of the total basin area [29,301.
5.3.4.7 Actual evapotranspiration

Unless measurements are made of the actual vapour flux (for example, by the evapori-
meter), most determinationsof evapotranspirationwill provide an estimate of the poten-
tialwater transfer to the atmosphere.In representativeand experimental basins.for some
part of the year at least,evapotranspiration may be taking place at an actual rate which
is different from the potential.From initially wet conditions the soil will dry according
to conditions of climate,vegetation and topography as well as its own characteristics.
At some limitingvalue it is thought that the soil-moisturestress will cause the actual and
potential rates of evapotranspirationto diverge,but the time at which this occurs and the
differences caused are matters of controversy [66].
Various accounting techniques have been adopted to allow for actual evapotranspira-
tion,but they are somewhat arbitrary and relate to individual basins.
5.3.5 Other climatic observations, including energy balance

Processing of observational data on pressure, air temperature,humidity,wind velocity,
cloudiness,soil temperature, depth of soil frost and thaw is carried out according to
standard methods [70,73,741. Temporalvariations of individual climatologicalelements
are checked by comparing various elements.In particular,an evaluation of the reliability
of observations of the depth of soil frost and thaw is made by means of comparing tem-
poral variations of these characteristicsand temporal variations of air temperature and
snow depth.
Major climatological observations may be tabulated to show daily, weekly, ten-day,
monthly and annual mean values. As a rule, the climatological data measured at the
main climatological stationmay be assumed to be applicable to the entire basin. In some
cases auxiliary observations are required (see section 4.2.5.2.1.2).Data processing for
energy-balancecalculations depends on the purpose;for example,the estimation of eva-
potranspirationfrom an experimental basin during the most intensive snowmelt period,
etc. Note that climateobservations are often made at a fixed time of the day (see section
4.2).When correlating,such observations should agree with hydrological data which is
222
frequently based on a day from 00.00to 24.00 hrs. A standard day should be decided
upon-e.g.,00.00to 24.00 hrs-and all climaticvariables adjusted if possible.
5.3.5.1 Wind
Wind velocity may be recorded manually or automatically.In either case it is the practice
to express values in metres per second (daily,monthly and yearly).
5.3.5.2 Temperature
Temperature data is recorded manually or automatically. In either case data is to be
recorded in degrees Celsius to the nearest whole degree.The daily mean temperature can
be calculated approximately as the mean of the daily maximum and minimum tempera-
tures.
Other important characteristics are the monthly mean temperature,which is the mean
of all the daily mean values for one month, and the annual mean temperature,which is
the mean of all the daily mean values for one year.
For some types of climate the following definitions are useful;the monthly mean maxi-
m u m temperature is the mean of the daily maximum temperature for one month,and the
monthly mean minimum temperature is the mean ofthe daily minimum temperaturesfor
one month.
5.3.5.3 Humidity
Humidity is recorded with an hygrometer or wet- and dry-bulbthermometer.
Daily,monthly and annual values of relative humidity are recorded as for temperature
The relative humidity can be read directly from the hygrograph chart.Where only
dry- and wet-bulb readings are available,the temperature of the dew point is obtained
from the dry-bulbreading and the temperature difference between wet-and dry-bulbread-
ings. The latter is termed depression of wet bulb. Standard psychrometric tables are
used to convert wet-and dry-bulbreadings to values of relative humidity,expressed as a
percentage.
Of the other methods for expressing humidity (see section 4.2.5) the vapour pressure
oneis probably the most useful.Vapour pressure can be obtained,togetherwith dew point,
from wet- and dry-bulbreadings using standard psychrometric tables.
5.3.5.4 Radiation
Solar radiation data is tabulated in langleys (1 small calorie per square centimetre of
surface per minute) for daily,monthly and annual values.
Sunshine duration records are tabulated as hours of sunshine.
5.3.5.5 Energy balance

Energy aspects of a number of hydrological processes and, in particular,energy losses
by evapotranspiration from a basin, are evaluated by energy-balance calculations.It is
particularly desirable to make such calculations on experimental basins where detailed
research in all aspects of the hydrologicalcycle is carried out.
For energy-balancecalculations,data on the following observations are required:(a)
the residual energy of the radiation exchange (so-calledradiation balance) or its compo-
nents (Wp);(b) the energy flux directed vertically downwards through the surface (Ws);
(c) the energy exchange between the surface and the atmosphere (Wu); (d) the energy
i 223
losses by evaporation or the energy of condensation (LE),where L is the energy of the

condensation and E the rate of the evaporation.
In a number of cases,especially when a contouring effect occurs on irrigated fields,it
is necessary to take into account also the advectiveinflow and outflow of the energy (A).
Considered positive are the radiation fluxes when they are directed towards the surface;
the energy flux into soil;the energy exchange with the atmosphere and the energy losses
on evaporation when they are directed upwards.
Therefore the energy-balanceequation (not taking into account the advective inflow
of heat) may be written as follows:
Wp- Ws-LE- Wa = O . (12)
The value W p may be found by measuring the radiation-exchangecomponents,and the

value Ws may be estimated by means of observations on the distribution of the tempera-
ture profile in the soil.The accuracy of these calculationsis not high because coefficients
of the conductivityof energy and temperature of the soils vary with time and depth.For
this reason the direct measurement by thermo-electricaldevices (soil-energyexchange
meters) give more reliable results.
The value Wa may be determined by the equation:
dt
W=eC,a-
dz
and the value LE by the equation:
where :
e = the air density;
C, = the air heat capacity at constant pressure;
dt/dz= the temperature gradient;
dpldz = the specific humidity gradient along the vertical;
LY = a coefficientof turbulent air flow,determined by the vertical wind profile (taking
intoaccountthe temperaturestratification oftheatmosphereinthefrictionlayer.)
However,the calculation of the value (x at high above-adiabatictemperature gradients
does not give reliable results.For this reason it is better to determine the values Wa and
LEby a simultaneous solution of equations (13) and (14).
The relation between the values L and t may be approximatedusing the Smithequation :
L,= 569.9-0.580t cal/deg. (15)

Evaporation can be expressed by the equation:
wp-ws
E= mmlhr
At P)
(59.7-0.058t)+ 2 Q A/- --
where :
At = the temperature difference (in OC) between two altitudes;
Ae = the absoluteair humidity difference between the same two altitudes (in millibars);
PO = the normal atmospheric pressure;
P = the atmosphericpressure at the time of measurement.
The value of the energy exchange with the atmosphere may be calculated as the residual
of the energy-balanceequation (equation 12):
W U =Wp- WS-LE.
224
Data processing and publicution
Reliable results are not obtained by use of this equation where values of the difference
W p - Ws and of the value de are very small.
In this case the value W u should be calculated by equation (12), the value LE being
obtained as the residual term:
LE= W p - Ws- Wu.
Since the relationsof LEand Wu with temperature and humidity gradients are non-linear,
long-periodaverages should not be used (not even 24-hourvalues).
In the majority of cases, processes of evaporation and turbulent exchange may be
considered stationary or near stationary up to about one hour,and for this reason cal-
culations should be based on hourly averages.
In a number of cases,instantaneous values of the fluxes may be calculated and the
final results averaged. It is therefore necessary to make repeated measurements to obtain
reliable values of the energy fluxes for a given accuracy.The required repeatabilityof the
observations may be determined by means of a series of observations and subsequent
statistical analysis [45]. Further details on energy-balancecomputation are to be found in
the literature [36,431.
5.4 Surface water

5.4.1 General
Streamflowdata may be reduced to instantaneous discharges,mean and average dischar-
ges and run-offvolumes.Lake and pond data are normally reduced to volumes of storage.
For both representative and experimental basins,instantaneous discharges (for ana-
lysis purposes) and run-offvolumes (for analysis and water-balancepurposes) should be
calculated.Where representativebasins are used for prediction purposes,mean and aver-
age discharges are also valuable.
Methods of discharge calculation,using current meters, floats or tracers are given in
the literature [21, 42,68,69,70, 721.
5.4.2 Charts and tapes

Stage-heightdata are observed either manually or automatically on charts and tapes
(punched and magnetic), as discussed in section 4.3.Both manual and tape recording
are normally at fixed intervals;chart recording is continuous. The choice of recording
method must be adapted to the data-processingmethod used, especially if automated
procedures are followed (see section 5.8). Tape recording is based entirely on subsequent
automated data processing ; chart recording may be automated if a chart-to-digitalcon-
version unit is used.
Any method of recording or data processingrequirescareful annotation of charts and
tapes and correction for missing records,time and stage errors. In the case of automatic
data processing,particularly careful annotation is necessary to allow a smooth operation
of mechanized computation.
Detailed annotation and correction procedures are given in the literature [39,61J.
5.4.3 Tabulation of time and stage heights

Whether continuous or intermittentrecording of stage height is employed,specifications
must be laid down on the instantaneous gauge heights to be utilized for conversion to
225
discharge.The principle should be that the hydrograph is defined in such a way that
peaks, valleys and points of inflection (instantaneousdischarges) can be reproduced and
mean dischargescan be calculated with a given accuracy.For research on representative
and experimental basins this accuracy may be set at 2 per cent, although this will be
impracticable in many cases.
This consideration may lead to a fixed time interval for observation which may aid
in the selection of the punch-outinterval for punched-taperecorders or selection of take-
off intervals for automatic or semi-automaticchart to digital conversion units.
Where manual processing is done (with or without desk calculator), take-offintervals
depend on chart scales.
For run-offplots,a greater accuracy may be required than for the larger representative
basin where less frequent intervals may be satisfactory (especially if weekly or monthly
charts are used).
5.4.4 Stage-discharge relations and ponding corrections

To convert from instantaneousstage to instantaneousdischarge a relation must be devel-
oped between stage and discharge.In the case of precalibrated measuring structures(see
section 4.3.4)a curve of relation or a rating table will be available (unless,as is the case
on many run-offplots,the volumetric method of flowmeasurement is used). In all other
cases discharge measurements should be processed as soon as possible after they have
been made and,as a minimum,relationsshould be developed between the stage and the
instantaneousdischarge,the area and the mean velocity.
Ifthe scatter of the points around the mean curvesis significant,observations should be
repeated (sometimes at frequent intervals).
In some cases the discharge may deviate from the stage-dischargerating curve because
of ice phenomena,aquaticvegetation,etc.,causing backwater,or because of shiftingbeds
on a downstream stretch of a river,or unsteady flows.Flow estimation for such periods
should be made by using correction coefficients.For backwater,for example, the ratio
of the measured discharge(41)to the discharge(qk) according to the stage discharge curve
is:
The values of such coefficientsmay be interpolated, taking into account the dates of
discharge observations.The discharge at time t may be estimated as qt = C t qk.
In some cases the backwater may be variable and a family of curves will be obtained
for the discharge site,each curve relating to different backwater conditions.
In cases where it is impossible to obtain a uniform discharge curve (e.g.,the channel
is blocked by ice or snow) observations should not be neglected. Flow may,for instance,
be approximately estimated by interpolating from previously observed data, and rating
tables may be developed from the stage-dischargerating curves.Detailed procedures are
given in the literature [61,64, 681.
If a pond exists above the measuring structure,pondage corrections may have to be
applied to obtain discharges unaffected by ponding (the outflow hydrograph is corrected
to produce the inflow hydrograph).
If the pond area is less than 0.5per cent of the basin area,several storms must be
corrected on a trial basis and peak inflow plotted against peak outflow.If a straight line
resultsand no point varies by more than 5 per cent from a 45"line,no pondage correction
is necessary,but if the pond area is between 0.5 and 2.5 per cent of the basin area, a
correction is necessary.
226
If the pond area is over 2.5 per cent of the basin area,account should be taken of the
rain falling on the pond when calculatingrun-off[61].
5.4.5 M e a n discharges
The daily mean discharge is the mean flow over a period of one day from 00.00to 24.00
hrs.
The monthly mean discharge is the arithmetic mean of the daily mean discharges over
a period of one calendar month.
The annual mean discharge is the arithmetic mean of the daily mean discharges over
a period of one water year.
The monthly average discharge is the mean of a number (not necessarily consecutive)
of monthly mean dischargesfor a particular month.
The annual average discharge is the mean of a number (not necessarily consecutive)
ofannualmean discharges.
If stage intervals are taken off at irregular intervals,daily mean discharges are calcu-
lated by the method of balancing areas,as shown in Table 5.3.
TABLE5.3.
Instantaneous M e a n discharge Products
Hours (ma/sec) (cols. 3 x 4)
(hrs) discharge (m3/sec)
(1) (2) (3) (4) (5)
00.00
10.00 8 10 8 80
12.00 12 2 10 20
14.00 1O0 2 56 112
15.00 140 1 120 120
16.00 1O0 1 1 20 120
18.00 44 2 72 144
20.00 24 2 34 68
24.00 16 - 4
24
20 80
744
Daily mean discharge: 744/24 = 31 m3/sec.
5.4.6 Run-ofS calculations

Run-offis normally expressed in millimetre depths over the basin area. The calculation
to convert m3/sec to mm/hr is as follows:
Discharge in m3/secX 3600 sec
Basin area in km2 X 1000
= run-offin mm/hr
Figure 5.9shows an example of calculationof run-off.

In columns 6and 7the discharge is tabulated in i/s (corrected for pondage if required),
and in column 8it is enteredin mm/hr.Allvalues of mm/hrare tabulatedto three decimals.
Great care must be used when rounding off decimals to ensure as small an error as possi-
ble.
In column 9 average discharge is tabulated for time intervalin mm/hr (three decimals).
In column 10totalrun-offfor time interval is tabulatedin millimetres to three decimals.
227
Representative and experinienral basins
BASIN MAKARA No. I I AT MAKARA AREA 7.86 HECTARE

MAP REFERENCE: N 164:262287 STRUCTURE Im-H fiume RATING TABLENa Standard
GAUGE HEIGHT ZERO F L O W 0.0 GAUGE HEIGHT ZERO PONDAGE 0.0
GAUGE HEIGHT U)WEST INTAKE 0.0 GAUGE HEIGHT FLOAT AT REST 0.0
CONVERSION FACTOR BASIN AREA I litrekjec 0.0458 MMiHR. YEAR 1965
I 2 6 7 8 9 IQ II I2 I3
TI HE
ûATE INTERVA
a STAGE
TIME
NNVTE3
MAQCH
1130
1245 75 40.6
1400 75 38.I
1515 75 35.6
1700 105 33.0
2100 240
._
30.5
2400 i80 2 7.9
MARCH
5 ~~
30 O0 27-9
3130 90 27-9
3200 30 30.5
--
1224 24 33.0
3230 6 35.6 1 I I .287 0059 0.055 0006 0403
3233 3
3236
3239
0242
3
3
3
l. 587
,765-
I 939
0. ;-I-
0073 0069 10003 0.109
__
081 0077 10,0040.113
-t
O089 O085 10.004~-0117
3245 3 -
. - !.323 0.106 0.0% 10.0050.122
3248 3 55.9 ,748 0426 0.116 0.0060.128
3251 3 63.5 3 433 O 157 O 142 0007 0135
1254 3 76.2 4 760 0218 O188 10009 0.144
3257 3- -
88.9 5.294 0.288 0.253 '0013 0.157
I_
3300
-_ 3 104.1
~.
-~
__ 8.468 0.388 0.338,0,0170174
i303 3 119.4 3.952 0.5020.445 10-0220.196
I396 3 132.1 3-30I 0.609 0.556 10.028 0.224
3309 3 144.8 5.885 10.72810.669 0033 O257
COMPUTED BY B. R. P.
CHECKED BY J . M . H.
FIG.5.9. Form for calculating flow records for representative and experimental basins,Ministry of Works,N e w Zea
228
Data pi-ocessirig arid publication
I In column Il accumulated run-offis tabulated.Accumulation is from zero to zero for

ephemeral streams;for perennial streams daily,monthly or annual accumulations may
be used.
In column 12 accumulated total run-offfor each day and accumulated total run-off
for each month are tabulated.
Cross-checkingmust be done wherever possible and the credibility of a result consider-
1 ed. It is very simple to check monthly totals and to introduce some measure of certainty.
I
5.4.7 Approximate checks on validity of streamflow
Temporal variationsin the flow pattern as observed at a gauging station should be com-
pared with certain climatologicalfactorstocheckapproximatelytheirvalidity.Floodhydro-
graphs caused by rainfall only should be compared with precipitationand air temperature.
Flood hydrographs caused by snowmelt should be compared with water equivalent of
snow cover, precipitation and air temperature. Any variation in the normal low-flow
pattern should be checked with air temperature.
Where more than one gauging station is installed on a basin, it may be possible to
check approximately total seasonalor annual run-offvalues. The total run-offat a gaug-
ing station (QI) should equal the run-offfrom an upstream station (Qz) plus any inflow
between the stations (@). Where such inflow can be estimated with reasonable accuracy,
the difference between QIand 8 2 should not exceed 5 per cent. If greater discrepancies
occur,the flow measurements should be carefully checked.In some cases,great discrepan-
cies may be due to storage between the gauging stations (e.g.,karst phenomena) and/or
the occurrence of springs, etc. In such cases flow measurements between the gauging
stations may aid in clarifying any irregular differences.Unless a cultural or relatively
rapid natural change occurs in a basin,run-offrelations between successive gauging sta-
tions on the same stream should be relatively stable from year to year.
5.4.8 Lakes and ponds

Changes in stage are applied to lake or pond areas to obtain volumes of storage.In large
lakes with relatively steep sides and little or no overflow area, the surface area of the
lake is applicable to all stage heights. In smaller lakes and ponds with sloping sides and
overflow areas a detailed survey of surface areas at various stages should be made.
In lakeswhere unsteady stages occur (e.g.,seiches), several gauges may be required for
the calculation of lake storage.
5.5 Subsurface water

5.5.1 Water in the unsaturated zone
5.5.1.1 Soil moisture
Soil-moisturedata are of considerable value for research on representative and experi-
mental basins,since they reflect the moisture status of the zone of greatest hydrological
activity. The soil-moisturestatus reflects the condition of the soil immediately prior to
the occurrence of snowmelt or precipitation and in any studies of snowmelt or precipi-
tation-run-offrelations,soil-moisturedata are important. Soil-moisture data are also
essential in water-balanceand hydrological model studies.
Data processing of soil moisture may include: (a) estimation of moisture content of
different soil horizons; (b) determination of the water content of the zone of greatest
229
soil-moisturechangeforevery observationpoint or formaster sitesonly (see section4.1.1);

(c) determination of the soil moisture profile at every observation point or at master
sitesonly;(d) estimationof the variations ofthe basin soilmoisture between observational
dates. Sometimes more than one soil moisture observation point is located at a site.In
such cases, the arithmetical average of the values at all points serves as the mean soil
moisture status for that site.
For the determination of the basin average,other methods of averaging may have to
be employed,depending on the sampling method used (see sections 4.1.1 and 4.4.1.1).
The observationaldata is frequentlyutilized to determine the number of points required
to give a mean basin value with a given accuracy. Statistical methods, similar to those
mentioned in section 4.1.1 may be applied.
To detect samplingerrors it is useful to compare soil-moisturedata observed at similar
depths with adjacent observationpoints.Missing recordsmay often be replaced by cor-
relation techniques.A further check of soil-moisturevariations with precipitation will
indicate approximately the validity of increases in soil moisture.
Soil moisture is frequently tabulated in the following form:
Depth Average soil moisture

Date
Time represented (cm)
Weight Volume Water
From To (%dry weight) (%) (mm)
It is useful to tabulate certain soil physical characteristics such as wilting point, field
capacity,saturation,etc.,with these data (see section 4.7).
Soil moisture, expressed as a depth in millimetres of water in the soil,may be plotted
against time for each date on which soil moisture was sampled.Daily amounts of reten-
tion,computed as rainfall minus run-off,can be plotted on the same graph and these two
sets ofdata,retentionand soilmoisture,may be used to estimateaverage rates of moisture
depletion in deriving the moisture status between sampling dates.
Moisture-depletioncurves are useful for analysis purposes. Such curves are affected
by the soil-moisturetension which increases as soil moisture is depleted. Some workers
have described the construction of master depletion curves [5].
The advent of the neutron scatterer (see section 4.4.1.1)allows the depiction of soil
moisture throughout the soilhorizons.Patterns of moisture extraction are extremely use-
ful in analysis (see section 6.1.4)and provide important observations for water-balance
studies.
Data processing of soil physical characteristicsis carried out by standard methods [74].
5.5.2 Water in the saturated zone

5.5.2.1 Generai considerations
Normally a ground-waterinvestigationseeks to determine certain or all of the water-bal-
ance components of some particular area-for example, rates and yearly amounts of
recharge and of discharge,capability for sustained withdrawal, volume in storage and
fluctuationsin that volume and the rate at which water is transmittedthrough the aquifer
system [i].
5.5.2.2 Errors of measurement

Ground-water data consist of observations of piezometric heights or heights of water
level measured in wells (see section 4.4.2).Records normally include certain common
230
dimensions of wells such as diameter, depth, yield,drawdown, diameter and length of

casing, interval screened,etc. Many of these data are observed by untrained observers
and should be treated as suspect;as many checks by trained observers should be made
as is practicable.
Well records may appear to be perfectly satisfactory,but sometimes wells are partly
clogged and,especially if surface water can enter them,errors are very difficult to detect.
For detailed description of errors in ground-watermeasurement,refer to the literature
[43].Note that when a specialnetwork of wells is installed,the accuracy of ground-water
measurements may be as high as that for surface water. Correctionsmay have to be made
to water level observations in wells,piezometer readings and ground-watertemperature.
In some cases standard correction coefficientsmay be applied.
5.5.2.3 Methods of presentation of data

5.5.2.3.1 G R O U N D - W A T E R LEVELS
Data may be in tabular form or plotted as continuous hydrographs.The fluctuations of

the hydrograph give a general insight into the drainage and storagecapacity of the ground.
Large fluctuationsare generally associatedwith smallstoragecapacitiesand largeunsteady
subsurface flows,while small fluctuations pertain to a larger storage capacity. Naturally
climatologicalfactorsshould be taken into account;this is particularly useful ifthe record
extends over severalyears. Hydrographs may also indicate changes in the water balance.
5.5.2.3.2 FLUCTUATION DIAGRAMS
Another method of data presentation is by means of fluctuation diagrams. For this

purpose simultaneous readings of two observation wells are plotted against each other.
Certain wells will give a very good correlation,as shown in Figure 5.10~.Others will
show no correlation (Fig. 5.10b). The observations of Figure 5.10 were obtained from
an area ofabout 4kmz. The correlationbetween wells A and B points to similarhydrologi-
cal properties,while the conditions in well C are different.
160-
140-
120
._.
o
FIG.5.10. Fluctuation n
40 60 80 100 120 140 40 60 80 100 120 140
diagrams for ground- Depth in well A Depth in well A
water observations. a b
This method may be applied to a denser network of observation wells. After sufficient
data have been observed,correlationsmay be established and a limited number of wells
maintained forfurtherobservation.This Correlationis time-consumingbut may be reduced
by considering other factors,such as topography,soil characteristics,etc.
The method is very effective for the determination of the influence of drainage and
pumpage. For this purpose one reference well which is not influenced by such changes
is chosen.Figure 5.11 gives an example of the determination of the effect of temporary
231
~-
680
670
660
650
64C
L
630
3
o
1
+- 61C
FIG.5.11. Effect of .-Is)
a,
pumpage on depth of T I l I I l I I
water table (determined 600 610 620 630 640 650 660 670
from fluctuation
diagrams). River level
pumpage for building purposes on the piezometric head of the ground water. The latter
is plotted against river stage (the well was situated about 350 m from the river). The
original height of the ground water was determined by the river level.
The determination of mean fluctuations of ground water over a basin may be carried
out as follows.
According to the fluctuations of well levels and the specific yield of aquifers,the value
of ground-waterstorage fluctuation for a definite time interval may be determined for
individual points of a basin. After drawing isolines of the given variations on a map.
the areas between adjacent isolines are planimetered. The mean value of ground-water
storagefluctuationsfor each estimated area is taken as half the value of adjacent isolines.
For the entire basin, the mean ground-water storage fluctuations are estimated as a
weighted averagefrom the mean values of the variations for individualareas.If the water-
bearing strata in a basin are homogeneous,then the specific yield may be assumed to be
equal. Mean ground-waterfluctuations for the whole basin may then be estimated as
outlined above. Mean storage fluctuations are obtained by multiplying the estimated
mean fluctuations by the specific yield.
5.5.2.3.3 GROUND-WATER CONTOUR MAPS
An important factor in ground-waterstudies is the depth of ground-water below the

surface. This determines the uptake of soil moisture and the aeration of the soil. A map
indicating the depth of the ground-watertable may be obtained by combining,for a given
area,observed ground-waterlevels with a contour-heightmap (Fig,5.12).
232
Data pr.ocessitig atid publication
a b C
FIG.5.12. Construction of ground-watercontour map and map showingdepth of water table:

(a)contourmap,with observed depth of water table in wells; (b) ground-watercontours derived
from (a); (c) ground-water depth map, composed from (a) and (b).
Ground-watercontour maps may then be made by drawing contours connectingpoints

with the sameelevation of the water table.This type of map is very useful for the evalua-
tion of a basin divide but can also be used as a basis for the evaluation of the direction
and intensity of flow of ground-water.This is because water must flow perpendicularly
to the isohypses (lines of equal potential), since the driving force has its maximum value
in this direction (see also section 6.1.4).
5.6 Erosion and sedimentation

5.6.1 Sediment rating curves
The relatjon of suspended-sedimentload to discharge is often called a sediment-rating
curve and is usually shown as a plot, on logarithmic paper, of suspended sediment in
units of weight per unit of time against discharge of the sediment-watermixture (see
Fig.5.13). It will be noted that there is considerable scatter about the line in Figure 5.13;
suspended-sedimentload usually increases faster than water discharge at a given station.
It is believed that the slope of the curve showing the relation of suspended sediment to
discharge is the result of the processes of sheet,rill and gully erosion on the basin and
represents sedimentimposed on the major stream channelswhich must carry it away [33].
The sediment-ratingcurve indicates a mean condition and may suggest a much closer
correlation than exists in momentary or daily values. This is because the unit of yield,
tons per day, is itself a product of sediment concentration times discharge. When this
product is plotted against one of the factors (discharge), some degree of correlation is
inevitable.Statisticalor graphicalcorrelationof discharge against sedimentconcentration
alone is more realistic [12].
5.6.2 Calculation of sediment yield

If sediment yields are calculated from the average concentration of sediment in a series
of samples,a substantial error may result unless the data cover wide ranges in discharge
and in concentration.Experience has shown that 90per cent or more of a year's sediment
yield may occur in one or two days of flood flow.This being so,substantially less than
actual yield will be computed if these days are not included in the data; substantially
more than actual yield will be computed if the data cover the flood days but few others.
Errors of this kind are minimized if discharge-concentrationor sediment-ratingcurves
233
are applied to instantaneous discharges similar to daily mean discharge calculations (see
section 5.4).
A simpler calculation is to apply sediment-dischargerelations to a flow-duration curve.
However, for the relatively small representative and experimental basins, especially those
of a flashy nature, final yields may be underestimated since a flow-duration curve is
calculated from daily mean discharges and not from instantaneous discharges [69,731.
Discharge (rn3/sec.)
FIG.5.13. Sediment-rating curve for Brandywine Creek, Wilmington.
234
I
5.6.3 Relation to deposits in reservoirs
I
The silting rate of ponds,lakes and reservoirs depends on the annual amount of sediment
transported into the reservoir. For very small reservoirs,a reasonable estimate of this
value may be made by calculating the trap efficiency of the reservoir.This trap efficiency
may be expressed as:
C
TE= -CL (1 8)
G,
where:
TE = percentage trap efficiency;
C = the original reservoir storage capacity in m3/km2;
G, = the annualnet sedimentinflow into the reservoirin m3/km2per year;
CL = the annual silting rate or capacity loss in per cent per year.
As an index to probable trap efficiency, the ratio of storage capacity over basin area
(C/.) is frequently used. T o determine this index,a survey is made of as many reservoirs
as possible in a given hydrological region. This index should not be used outside the
region since,with the same C/Aratio,the trap efficiency will increase as the run-offper
unit area decreases.
For a given period of operation of the reservoir,the annual amount of sedimentation
per square kilometre of its area may be estimated by multiplying the trap efficiency by
the annual sediment inflow.By multiplying this value by the basin area and the number
of years of operation of the reservoir,the total sedimentation may be calculated.Sub-
tracting the latter value from the initial reservoir-storagecapacity,a new trap-efficiency
index may by calculated and a prediction of future sedimentationmay be made by using
this revised trap-efficiencyindex and the expected sediment inflow.
Volume and weight of sediment in reservoirs may not represent total or long-term
average yield from the basin.Reasons include the facts that the period spanned by data
may not be a period of average climatic conditions;that,not uncommonly,data on sedi-
ment yield are from reservoirs of low trap efficiency; that, as a substitute for actual
measurements,no accuratemethod exists for computing amount of sedimentpassing over
spillways; that some data are from reservoirs with ungated outlets (and that sediment
passing through such outlets cannot be calculated satisfactorily); that available data
commonly fail to cover deposits above spillway level (upstream from the reservoir); and
that,in some streams,coarse sediment is a substantial part of the total yield and much
of this may be deposited outside of reservoirs.
5.6.4 Interpretation of results

The present status of sedimentmeasurement and analysis suggests caution in interpreting
smalldifferencesbetween observationperiods or places,in interpretingdifferencesbetween
components of sedimentdischarge,and in inferring sedimentaryprocesses from sediment-
discharge measurements. For analysis methods, see section 6.1.5[60].
Observed basic data on sediment may be presented in the form of tables that contain
data on sediment concentration,water discharge, sediment discharge, particle sizes of
suspended sediment,bed load,and the rate of siltation of reservoirs.Data may be given
in monthly and/orannual amounts.
' 5.7 Water quality

Observed chemical water-qualitydata should be processed by calculating the difference
between the total milligramme-equivalentsand the estimated concentration of anions and
235
Represrnraiive and ezrpeuimental basiiw
cations.The ionic run-offis determined by plotting discharge ir1 I/sec on the ordinate
against ionic concentration in mg/lon the abscissa. A curve is drawn through the mean
of the points. Applying this curve to the discharge hydrograph,the daily mean ionic
concentration may be derived.By multiplying the daily mean ionic concentration (mg/l)
by the mean daily run-off(in litres), the daily ionic run-off(in mg) can be calculated.
Adding these values for corresponding periods, the total values of ionic run-offfor a
month, season and year may be derived; these are usually expressed in tons/kni2of a
representativeor experimental basin area.With the accumulation of new data the graphs
should be brought up to date.
The results of chemical quality observations may be tabulated. Such a table would
contain water sampling and analysed data,individualionicconcentrationsof samples,etc.
5.8 Automated processing of data

Before or immediately after publication of the basic observational data,it is desirable
to process them in the form most valuable to the users. The typesof processingperformed
in any country depend on the needs of the most important and most numerous users of
the data.
The publication of data subjected to data processing does not necessarily supersede
publication of the basic observational data. However,rapid storage and retrieval ofdata
(see section 5.9)by means of computerized techniques may eventually result in no need,
or at least much less need,for publication of data (see section 5.10).
5.8.1 Automated processes

5.8.1.1 Digital recorder
Digitalrecorders are being used more extensively each year (especially for recording river
stage) because the data from such recorders are usable for computation and analysis by
computer. The digital recorder now in use in several countries is a battery-operated,
slow-speed,paper-tapepunch that recordsa four-digitnumber on a sixteen-channelpaper
tape at preselected time intervals.The paper tape format requires rearrangement before
input to any of the presently available high-speed digital computers. For greater detail
on the digital tecorder and its application,refer to the literature [16,271.
Once the data have been translated into machine-usableform,they may be processed
by a computerthrough a wide variety of specialprogrammes.Some programmesavailable
at present are listed in section 5.8.1.4.2.Figure 5.14shows a flow chart for one applica-
tion of the digital recorder-the processing of streamflow data.
5.8.1.2 Pencil follower

A new developmentis the pencil follower which is a device for transferring either manu-
ally or semi-automaticallya chart record into machine-usableform (punchcards,paper
tape or magnetic tape). One type, developed in the United Kingdom (D.Mac Pencil
Follower) is suitable for all types of chart, including circular ones. The operation is
either manual or semi-automatic and an average chart record can be converted into
digital form in about five minutes. Another type, developed by the CSIRO, Australia,
is particularly suited for continuous charts.
5.8.1.3 Punchcard system

Observed hydrological data are frequently transferred to punchcards for data processing
and analysis.A method used in the United States of America to transfer hydrological
data to punchcards is given below as an example.
236
Samples of soil are often collected during sampling operations for hydrological studies.
Thesesamples are chosen to be representativeof the depths and soils of particularinterest
to the soil-moisturestudy.The hydrological and physical properties of these samples are
obtained by suitable analyses in the laboratory.
The data are recorded on special summary forms by a suitable system of numerical and
alphabetical codes. These coded data are transferred directly from the summary forms
to punchcards by the key-punch operator. The punchcards are used for automated
FIG.5.14. Flow chart for automated processing of streamfíow data.
231
print-outof tables of data,for data computationsand statistical analysis,and for other

data processing by computer techniques. One punchcard on which may be coded data
from analysis of soil-waterdata is shown in Figure 5.15.
333
444 4444444444 444444444444
555
666
777
888
999
Is Il JI
FIG.5.15. Punchcard for data on moisture properties of rock and soil samples.
5.8.1.4 Computer programmes

A wide variety of computer programmes is availablefor continuousprocessing of routine
field data,as well as for processing specialized data for a variety of research purposes.
Although the list given in section 5.8.1.4.2.
is not intended to be complete,it does provide
a cross-sectionof programmes available. Because of the special nature of the digital
recorder programme it is described in some detail below.
5.8.1.4.1 DIGITAL RECORDER PROGRAMME
In the United States,a family of programmes pertaining to the reduction of river-stage

data collected at gauging stations into figures of streamflow is available.The river-stage
data are obtained in the form of sixteen-channelpaper tape punched by the Fischer and
Porter digital recorder and are translated by off-lineequipment to seven-channelpaper
tape suitable for photoreading by the B-220computer.The initial computer operation
is a primary computation which produces daily mean discharges and other pertinent
data (which vary with the type of gauging station). For the most common type,the printed
output from the primary computationincludes the maximum, minimum and mean gauge
heights for each day;the datum and/or shiftcorrectionsused to correctthe gauge height
or rating table;the daily mean discharge and a list of the bi-hourlygauge heights within
the day;and the time of the maximum gauge height.In addition to the printed sheet of
preliminary results,the daily figures are stored on magnetic tape for use in later updating
or correcting any part of the year’s record.The final output consists of a table of daily
mean discharges with monthly and yearly summaries suitable for direct offset reproduc-
tion.An option in the programme gives a similar table of daily mean gauge heights when
required.
At present there are three different types of primary computation.The first type is for
regular river gauging stations where simple stage-dischargerelations can be used. The
second type is for ‘slopestations’where two recorders are used to record stage at both
ends of a channel reach and the computationsmake use of the interrelationshipsbetween
238
stage,fall,and discharge.The third type is for ‘deflectionmeter stations’where two record-

ers are used at a single site,one to obtain the usual stage record and the other to obtain
the deflection meter record.The deflection meter record gives the position of a movable
vane in the channelfrom which the instantaneousvelocity of the stream may be obtained.
In N e w Zealand a slightly different method for processing streamflow data (and also
suspended-sedimentand precipitation data) is followed.The basic data are either obtained
by Fischer and Porter digital recorders and translated to punched cards,or obtained by
charts which are converted into punched cards by means of a pencil follower.An optional
part of the processing system is the plotting of the stage or discharge hydrograph for
checking or analysispurposes.
5.8.1.4.2 E X A M P L E S OF P R O G R A M M E S A V A I L A B L E
(a) Computer programmes available from the United States Geological Survey, U.S.A.
B220 computer:
Multiple regression analysis. Backwater analysis.
General grain-size analysis. Run-offstatistics.
Flood-stage tables. Diffusion statistics.
Heat conductivity. Probability routing analysis.
Well pressure. Flood statistics.
Temperature effects. Flow through culverts.
Evaluation of aquifer constants. Flow variability.
Evaluation of aquifer constants card prepa- Hydrograph synthesis.
ration. Flow duration.
Drawdown tables. Current discharge.
p H curves. Stream cross-sectionanalysis.
Ionic activity. Random-walkanalysis.
Dissolved oxygen. Cross-sectionwalk analysis.
Ion-exchangechromotographical columns. 3-Drandom walk.
Static distribution coefficients. General least-square polynomials.
Static distribution coefficients II. Geochemical statistical analysis.
Static distribution Coefficients III. Curvilinear regression.
Cation-exchangecapacity. Multiple regression variable generator.
Chemical-wateranalysis. General regional gradients.
Multiaquifer flow. Trend contouring.
Irrigation networks. Orthogonal polynomial surfaces.
Tidal stream analysis. xy-symbol plot.
(b) Computer programmes available from the Water and Soil Division, Ministry of
Works, N e w Zealand, Elliott 503 computer (also available in metric units):
Flow programme for representative basins Current meter calibration (this includes a
(this calculates instantaneous discharge in curvilinear regression analysis of the cali-
cusecs,daily mean dischargein cusecs,daily bration points and the printing of a table
run-offin inches,monthly and annualmean of the velocities calculated from the
dischargesincusecsand maximum and min- equation obtained).
imum instantaneous discharges. Plotting Mean basin precipitation.
of stage and/or discharge is optional). Rainfall correlation.
Flow programme for experimental basins Thornthwaiteevapotranspirationcalculation.
(calculatesdischarge in in/hr,accumulated Penman evapotranspiration calculation.
discharge in inches,total run-offper day, Suspended-sedimentgauging calculation.
month and year in inches, daily mean Sediment-yieldcalculation.
discharge in cusecs, monthly and annual Iteration method of solution of unitgraph.
mean discharge in cusecs.Plotting of stage Matrix solution of unitgraph (least-squares
and/or discharge is optional). method).
Blackwater-curveanalysis. River stones analysis (sphericity).
Cross-section area calculation Multiple regression analysis.
239
Represeníutive and experimental busins
5.8.2 Some considerations regarding data automation

A computer may not do anything that could not be done manually,given enough person-
nel and enough time. However,the computer will do repetitivecomputationswith such
speed and accuracy that man can now afford to process many more data than ever before
and can try new methods he never could hope to use before.
Acquisition ofa computer,or even establishment of a computer programme,should not
be entered intocasually.Comprehensiveand detailedanalyses,both technicaland econom-
ic, must be made for each proposed applicationso that the feasibilityof the application
is based on sound judgemeat. Almost invariably, costs for automation are under-
estimated, and computer programming, debugging,and writing of instruction manuals
take longer than planned.Any large-scaleconversion of conventional to automatic data
processing (ADP)systemsrequiresadjustmentsin such items as organizational structure,
operational policy, and work habits of personnel. For example, field notes on water-
stage records must be prepared more clearly and completely under an ADP system,
because the records are processed some distanceaway from the field office and by comp-
uteroperating personnel unfamiliar with the detailed technical phases of the work.
Fully adequate financing (which may involve large sums) must be provided if an ADP
system is to be successfully inaugurated in an organization.The cost of computersmakes
it necessary to consolidate and centralize their uses by limiting their number within an
organization. On the other hand, a complete consolidation in only one ADP facility
within a large organization may lose its main advantage of speed of computation if
there is an extensive processing backlog. ADP equipment that is outmoded for its work-
load also can cause processing delays which negate the hoped-foradvantages of ,automa-
tion.
5.9 Data storage and retrieval

5.9.1 General considerations
A descriptionof procedures for the storage and retrieval of hydrologicaland climatolog-
ical data is given in the literature [68]. Some important points of the pertinent sections
are included here,along with a few notes of particular interest to the storage and retrieval
of data from representativeand experimental basins.
The first requirement is that all data are systematically organized so that they can be
easily found.
Because of the tremendous quantities of climatological and hydrological data being
collected in many countries,the storage of all original records may be precluded. Micro-
filming is useful in cases where storage space becomes a problem. Roll-typemicrofllm
requiresonly about 1/300of the storagespace needed for original records.
Transfer of data to punchcards permits efficient tabulationand analysis of large quan-
tities of data,but does not help the storageproblem. For example,if daily punchcards are
punched from monthly report form3 at ordinary climatologicalor hydrological stations,
the number of punchcards produced will occupy seven or eight times as much space as
the original documents.When punched tape is used as the medium for storage of data,
much less space is occupied than that required for punchcards. The paper tape tends,
however, to be less durable than punchcards. Magnetic tape or disc storage of datais
used in countries having access to computer facilities,because this provides the most
efficient storage and retrieval system.
Storage conditions should minimize destruction of the stored papers, cards, or films
by excessive heat,rapid temperature fluctuations,high humidity,dust, fire,and insects
or other pests.Non-inflammablefilmshould be used in microfilming.Because basic climat-
ic and hydrologicaldata are irreplaceable and invaluable,every precaution should be taken
240
I Data processing and publicarion
' to protect them from destruction.Whenever possible,duplicate sets of records should be

1 kept,one in the main collection centre and the other at regionalcentres.
The arrangementfor cataloguing or transferring data to microfilm is of particular con-
cern in hydrology. For some types of analysis-flood-peak discharge frequency at a
stream-gaugingstation,for example-the whole period ofrecord for the individualstation
should be readily available.This suggests the desirability of filing the reports for the years
of record from the one station in one location,or of transferring all the data for a partic-
ular station to one or several reels of microfilm. On the other hand, studies such as
storm-rainfallanalysis require that all the data for a region for a particular period be
filed or microfilmed together.The choice of arrangementshould be made only after the
most careful examination to determine which of these two general types of study is likely
to be the most frequent and most important.In some cases it may be possible to file the
original documents by station and the punchcard or microfilm records by month,or vice
versa.
Autographical records of water level,precipitation,etc.,play such an important role
in representativeand experimental basin research that thought should be given to the
best method of preserving the original charts. Although regular abstraction from the
charts may give the most frequently used information,the charts themselves will contain
additionalinformationwhich may be essential in some special studies.Thus the recorder
charts-either in their original form or on microfilm-or the punched paper tape for
digital recorders,must be retained.
5.9.2 Microfilm aperture punchcards

For those countrieshaving electronicsortingequipmentreadilyavailable,microfilm aper-
turepunchcardsmaybe used as a possibility for storage and retrieval of data.Figure 5.16
shows such a card,the 33 mm microfilm frame of which will store the equivalent of six
20 X 26 c m pages of data.These punchcards are especiallyusefulfor storageand retrieval
of laboratory and field graphical information,such as particle-sizedistribution graphs,
maps, geological cross-sections,hydrographs, and borehole logs, which does not lend
itself to coding on standard punchcards.
There are many types ofmicrofilm cameras which can be used for preparing the micro-
film aperture punchcards. One such camera automatically exposes,develops and washes
the microfilm and provides the completed aperture card within less than a minute,ready
for punching of the identification codes.This particular camera has a 10 c m depth of
focus and a 45 x 60 c m photo area.A reader-printer(Fig. 5.17)will permit the aperture
card,or 16 or 35 mm roll-typemicrofilm,to be viewed or printed at enlargements up to
1 about 25 times.Card-copyingequipment(Fig.5.18)may be used to make duplicate aper-
ture cards.The prints from the reader-printeror the duplicate aperture cards from the
copier are used to fill requests for information.
I 5.10 Publication of summaries

~ 5.10.1 Purpose of publishing
I The primary purpose ofpublishingeither basic or processed data is to provide tabulations,
maps and summaries in a form convenient to most users of the data. The format and
content of the publications are dependent upon the requirements of the majority of data
users.Availability of an electronic computer may also influence the amount of publica-
tion as well as the format and content.If the primary need of the users is for shorttabula-
1I tions or anaiyses of many different areas,publication may be unnecessary if the data are
readily available as high-speedprint-outfrom computer storage.
I
241
-" . * - ,
242
I
l Data processing and publication
FIG.5.17. Reader-printerfor aperture card or roll-typemicrofilm records.
Any system of publication or computer print-out must be designed to minimize the

number of data requests which are answered by manual retrievaland tabulation of data
from the office files.Publicationsmay also be produced by reproductionof print-outfrom
the computer.
Whether the data are made available by publication or by computer print-out,it is
important that a high level of data reliability be maintained and that some degree of
standardizationof format and units be achieved so that the data have maximum useful-
ness for international exchange of information.A system of quality control must be
, established to ensure that good-qualitybasic data emerge from the observation,collection
I and processing proceduresready for publication or computer storage.This quality control
is an essential part of any data-collectionprogramme and includes regular inspection of
data collection on stations,preliminary checking of data,and checking of internal con-
sistency of data and analysis.
The publication of data provides substantial advantages,including the safeguard of
irreplaceable records againstloss or destruction.The cost of data publication must,how-
ever,be considered in relation to over-allfinancialbudgets for data collection and to the
volume of requests for the supply of data of various types.In general,the cost of pub-
licationis smallinproportionto the total cost of data collectionand should be undertaken
whenever possible.
i 5.10.2 Publication requirements
Regular publication of basic or processed data may be advisable even though the collect-
ing agency makes only limited use of such data at that time. Publication usually makes
I 243
Representative and experiniental basins
FIG.5.18. Copier for production of duplicate microfilm aperture cards.
the data readily accessible and may prevent the loss of invaluable observations.Because
publication of data is more costly and difficult if postponed for a period of years until
the demand becomes very pressing,it should be started well before the demand for data
becomes critical.
If,on the other hand, data are readily available upon demand from computer storage,
little or no publication of basic or processed data may be necessary to meet demands.
Standard publication or demand print-out from the computer system can be evaluated
only on the basis of the needs of the greatest number of users and the relative efficiency
of the agency’s publication and computer systems.
One type of research on representativeand experimental basins requires hydrological
data on a daily, monthly, seasonal and annual basis. This requirement can usually be
satisfied by publications giving daily and monthly means and total of streamflow,precipi-
tation,evaporation and other climatic elements.
A second type of research requiresdetailed analyses of such items as individualstorms
and floods and short-durationdry periods. For such studies,hourly or more frequent
instantaneous data are essential.
The expense ofpublication to suit the second requirementis only occasionallywarranted
and the introductionof efficient data storage and retrieval systems makes detailed publi-
cation of this kind superfluous (see section 5.9). T o give an indicationof the kind of data
available,either a continuous annual hydrograph and/or one or more typical storms per
annum could be published.
Data from vegetation surveys,infiltration research,interception studies,etc., should
not be published in basic summaries,but in annual research reports (see section 1.4.3).
244
245
L
v
246
d)
d
O
N
241
248
Data processing atid publication
X
" f i
3 '9
z =
249
Representative and experimental b a s h
+
+
+,
+
Y
+
+
A I
8z
I
l
~
c
I
I
a
250
XI
I-
=I
Y
1-
- 1
I
.9
3
e
.zd
Li
251
Representative and experinierital basins
X
El
H
>
H
Y
U
252
Q
CI
~-
V
I
L
X
~ I-
1-F U
r(
I F
- I‘ L ,
I I
-I ü
Y
c
5
d
253
+
O
rci
O
254
255
8
h
3c
E
L1
c
o
-
T
L1
-
'6
C
c
L1
-
,E
c
L
h
E
h
5
c
,E
1
6
P
V
256
257
...................RIVER Mo. ............. IT ................. WLP REF. ........... REGION .........
IN OPERATION SINCE ................ FIELD AUTHORITY ..............
BASIN YAP: BASIN DEIAILS:
REFERENCES:
PAECIPI
GÏpT
TABULATION O F DAILY PRECIPITATION A N D RUN-OFF í.r n d d a ,,

vl
ló
11
12
13
14
15
16
11
18
19
20
21
22
23
24
25
26
21
28
29
30
31
1
Canversion to di herge 1 tor.
FIG.5.32. Publication form for representative basins, Ministry of Works, N e w Zealand.

258
' Monthly mean discharge for 19...

1 Average monthly discharge íor
.1
19... ____-
to 19...
Minimum monthly discharge íor
- -
lk.. ro19...
M s x i m m monthly discharge for
----- --__I
19... to 19...
Highest
Doi1 Mean D i s c h a r g e Percentage
E g u a f l e d or Erceeded Of of Time of T ~ m r
m'/sec
5.10.3 Frequency of Publication

It is recommended that publication of summaries be on an annual basis,from 1 January
to 31 December.
5.10.4 Contents and format

Contents and format for publication of data depend on the distribution (international
or national).
The Unesco Working Group on Exchange of Informationhas recommended a number
of forms for internationalexchange of standard observational data from representative
and experimental basins,covering the followingelements :atmospheric precipitation (Fig.
5.19); river discharge (Fig. 5.20); water content in the unsaturated zone (Fig. 5.21);
ground-waterlevels (Fig. 5.22); ground-watertemperature (Fig. 5.23); chemical quality
of ground water in observationwells (Fig.5.24);tank evaporation(Fig.5.25).In addition,
the Unesco Working Group on Representative and ExperimentalBasinshas recommend-
ed that data should be published in a similar form for:mean suspended sediment dis-
charge;temperature of river waters;chemical quality of river waters; pan evaporation;
climatologicaldata;and precipitation and run-offfor heavy rainfall (Fig. 5.26-31).
For national publications, more detail is frequently desirable.A n example for exper-
imental basins has been set by the United States Department of Agriculture [25].
For representativebasins,the formatshown in Figure 5.32could be used [50].
5.10.5 Publication of summaries

The national committeesfor I H D should annually or biennially publish summaries and
reviews of the research carried out in accordance with the IHD programme.Summaries
and reviews would contain preliminary results of investigations conducted on represent-
ative and experimental basins during the previous 1-2years.For details of such research
reports,see section 1.4.3.
References
I. ALTOVSKY, M.E.;KONOPLIANTSEV, A. A. (eds.). 1957. Metodicheskoe rukovodstvo PO
izuchenia rezhima podzemnykh vod [Methodological manual on the study of subsurface
water regime]. Moscow, Gosgeolizdat.
2. AMERICAN GEOLOGICAL INSTITUTE.1965. AGI data sheets, p. 147b. Washington, D.C.
3. AMERICAN Socmn OF PHOTOGRAMMETRY. 1960. Manual of photographic interpretation.
Washington,D.C.
4. ___. 1965. Manual of photogrammetry. Falls Church, Virginia.
5. AMERMAN, et al. 1966.North Appalachian experimental watershed Coshocton, Ohio. Annual
report 1965. USDA, Agric. Res.Serv.,Corn Belt Branch.
6. AMoRocno,J. ; BRANDSTETTER, A.1966. Characterization of gauge levelprecipitationpatterns.
Amer. Geophys.Un.
7. BADGLEY,
P.C. 1959. Structural methods for the exploration geologist. New York, Harper
and Row.
8. BENTAIZ, R.(comp.). 1963.Methods of determining permeability, transmissibility and draw-
down. Washington,D.C. (USGS water supply paper 1536-H.)
9. ___. 1963. Methods of collecting and interpreting groundwater data. Washington, D.C.
(USGS water supply paper 1544-H.)
260
I Data processing and publication
BIRDSEYE, C. H. 1928. Topographic instructions of the United States Geological Survey.

Washington, D.C., Govt. Printing Office. (USGS bull. 788.)
BISHOP,M. S. 1960. Subsurface mapping. N e w York, Wiley.
BROOKS, N. H.1965.Calculation of suspended load discharge from velocity and concen-
tration parameters. Proc. Fed. Interagency Agric. Sed. Conf.,p. 229-37. USDA, SCS.
(Misc. publ. no. 970.)
13. BROWN,I. C. 1963. Chemical methods as an aid to hydrogeology. Proc. Hydrol. Symp.
no. 3, p. 181-203. Nat. Res. Council.
14. BRUNET-MORET, Y .; ROCHE,M. 1966.Etude théorique et méthodologique de l’abatement
des pluies. Paris. (Cahier Orstom d’hydrologie, no. 3.)
15. BULAVKO,A. G. 1967. Otsenka progreshnostey pri vodnobalansovykh issledovaniakh
[Evaluation of errors in water budget research]. Sbornik rabot PO gidrologii,no. 7,p. 3-13.
CARTER, R. W. et al. 1963. Automation of streamflow records. (USGS circ. 474.)
COMPTON, R. R. 1961.Manual of field geology. Wiley, N e w York.
DAVIS, R.; FOOTE,F. 1953. Surveying-cheory and practice. N e w York, McGraw-Hill.
DE WIEST, R. 5. M. 1965. Geohydrology. N e w York, Wiley.
DOBRIN, M . B. 1960.Introduction to geophysical prospecting. N e w York, McGraw-Hill.
DUB,O. 1957. Hydrologia, hydrografia, hydrometria. Slovense vydavatelstvo technickej
literatury, Bratislava.
DUBREUIL, P. 1965. Méthodologie d’exploitation du bassin représentat$ élaboration et
classement des données d’observations, p. 262-74. Congrès de Budapest. Paris, Orstom.
(IASH publ. no. 66.)
GOLUBEV, V. S. 1966. Metody otsenki i rascheta sezonnykh i godovykh s u m m osadkov
na vodosborakh i sutochnykh s u m m osadkov na selskokhozaistvennykh poliakh [Methods
of evaluation and computation of total seasonal and annual precipitation on catchments
and total daily precipitation on the fields]. Materialy seminaraPO raschetam vodnogo balansa
rechnykh basseinov i organizatsii komplexnykh vodnobalansovykh ì agrometeorologicheskikh
nabludeniy, p. 13-27.Valdai, GGI.
HARBECK, G.E.1962.A practical field techniquefor measuring reservoir evaporation utilizing
mass-transfer theory. (USGS prof. paper 272E.)
HOBBS, H.W.; CRAMMATTE, F.B.(comp.). 1965.Hydrologic datafor experimental watersheds
in the United States 1960-61.USDA, Agric. Res. Serv.
HORTON, R. E. 1919. Rainfall interception. Monthly Weather Review, 47 :603-23.
JOHNSON, A. 1.; LANG,S. M. 1965.Automated processing of weather information. Proc.
First Annual Meeting, A WRA,Dec. 1-3, Chicago, Illinois, p. 324-50.
KRUMBEIN, W . C.;GRAYBILL, F. A. 1965.An introduction to statistical models in geology.
N e w York, McGraw-Hill.
KUZMIN, P. P. 1966.Teoreticheskaya skhema otsenki oshibok rascheta isparenia s pover-
khnosti sushi [Theoretical scheme to evaluate errors in evaporation measurements from
land]. Materialy mezhduvedomstvennogo soveschanìa po Probleme izuchenia i obesnovania
metodov rascheta isparenia s vodnoy poverkhnosti i sushi, p. 271-83.Valdai, GGI.
, et al. 1966.Opredelenie dekadnykh, mesiachnykh i sezonnykh velichin isparenia
s poverkhnosti vodosborov i selskokhoziajstvennykh Polei [Determination of decade,
monthly and seasonal evaporation values from the surfaces of basins and fields]. Materialy
seminara po raschetam vodnogo balansa rechnykh basseinov i organizatsii komplexnykh
vodnobalansovykh i agrometeorologìcheskikh nabludeniy, p. 28-66. Valdai, GGI.
KUZNETSOV,V. I.; GELBUKH, T. M.1966.Uchet vliania vodnoi rastitelnosti na isparenie s
vodoemov [The effect of aquatic vegetation on evaporation from reservoirs]. Materialy
mezhduvedomstvennogo soveschania po Probleme izuchenia i obosnovania metodov rascheta
isparenia s vodnoi poverkhnosti i sushi, p. 104-15. Valdai, GGI.
LAHEE, F. H. 1961. Field geology. N e w York, McGraw-Hill.
LEOPOLD, L. B.;MADDOCK, T. 1953.The hydraulic geometry of stream channels and some
physiographic implications. (USGS prof. paper 252.)
LITOVCHENKO, J. V.,et al. 1965.Raschety geomorfologicheskikh kharakteristik vodosborov
[Computationof geomorphological basin characteristics].Materialy soveschania PO voprosam
experimentalnogo izuchenia stoka i vodnogo balansa rechnykh vodosborov,p. 324-29.Valdai,
GGI.
MCCULLOCH, J. S. G.1965. Tables for the rapid computation of the Penman estimate
of evaporation. J. East Afr. Agric. For., 30(3) :286-95.
I 261
36. MILLER, D.H.1965.The heat and water budget of the earth’s surface.Advances in Geophys.,
11. N e w York, Academic Press.
37. MILLER, R.L.;KAHN, J. S. 1962. Statistical analysis in the geological sciences. New York,
Wiley.
38. MILLER, V. C. 1961. Photogeology. New York, McGraw-Hill.
39. MORRISSEY, W.B. 1964. Interpretation and correction of water-levelrecorder charts. (Hand-
book of hydrological procedures, prov. proc. 32.) Wellington, S C and RCC.
40. NOZIN, V.;PETROV,B.1959.Methods and techniques of field soil mapping.In:I. V. Tjurin
et al. (ed.). Soil survey-a guide to field investigationsand mapping of soils, p. 45-60. Transl.
from Russian by Israel Programme for Scientific Translations.
41. ORLAND, H.P.(ed.). 1965. Standard methods for the examination of water and wastewater,
12th ed. N e w York, (Amer.Public Health Assn., Amer. Water Works Assn., Water
Pollution Control Fed.) Amer. Public Health Assn. Inc., 769 p.
42. ORLOVA, V. V. 1965. Gidrometria (Hydrometry). Leningrad, Gidrometeoizdat.
43. PACIFIC SOUTHWEST INTERAGENCY COMMITTEE. 1966.Limitations in hydrologic data. Chapter
IV: Report of Hydrology Subcommittee, USDI.
44. RETZER, J. 1963. Significance of stream systems and topography in managing mountain
lands. In: C. T. YOUNGBERG (ed.). Forest-soil relationships in North America, p. 399-411.
45. RIOU, CH. 1966. L e calcul de l’évaporation par la méthode du bilan énergétique en zone
sahélienne. Paris. (Cahier Orstom d‘hydrologie,no. 7.)
46. SEARCY, J. K. 1960. Graphical correlation of gauging station records. Washington, D.C.
(USGS water supply paper 1541-C.)
47. SILIN-BEKCHURIN,A. T. 1951. Spetsialnaya gidrogeologia [Special hydrogeology]. M .
Gosgeolizdat.
48. 1965. Dinamika podzemnykh vod (s osnowami gidravliki). [Dynamicsof under-
ground water (with elements of hydraulics)]. Izd. Mosk. Univ.
49. SMITH, H.T. U. 1943. Aerial photographs and their applications. N e w York, Appleton-
Century-Crofts.
50. SOIL CONSERVATION AND RIVERS CONTROL COUNCIL. 1966. Hydrology annual no. 14. Wei-
lington, Govt. Printer.
51. SOIL SURVEY OF GREAT BRITAIN. 1960. Field handbook. London.
52. SWAINSON, O. W . 1938. Topographic mapping. Washington, D.C.,Govt. Printing Office.
(U.S. Coast Geodetic Survey, special publ. 144.)
53. TAYLOR, N.H.;POHLEN, I. J. 1962.Soil survey method. N e w Zealand,DSIR. (Soil Bureau
bull. 25.)
54. THORNTHWAITE, C. W.;MATHER, T.R. 1957. Instructions and tables for computing po-
tential evapotranspiration and the water balance. Publ. in Climarol., lO(3). New Jersey,
Drexel Inst. of Technol.
55. TJURIN, I. V. et al. (ed.). 1965.Pochvennaja sjemka. Rurovodstvo po polevym issledovaniam i
Icaiiirovaniu pochv [Soil survey-a guide to field investigations and mapping of soils].
Moscow, Academy of Sciences of the U.S.S.R.Transl. from Russian by Israel Programme
for Scientific Translations.
56. TOEBES, C. 1964. Daza processing for experimental basins. (Handbook of hydrological
procedures, prov. proc. 33.) Wellington, S C and RCC.
57. UNITED NATIONS EDUCATIONAL, SCIENTIFIC AND CULTURAL ORGANIZATION. 1963. lnter-
national legend for hydrogeological maps. Paris, Unesco. (NS/NR/20.)
58. UNITED STATESDEPARTMENT OF AGRICULTURE. 1951. Soil survey manual. Washington,D.C.
(Agric. handbook no. 18.)
59. ___ . 1960. Soil classification-a comprehensive system, 7th approximation. SCS.
Washington, D.C., Govt. Printing Office.
60. ~ . 1961. A universal equation for predicting rainfall-erosion losses. Agric. Res. Serv.
(Special report 22-26.)
61. ___. 1962. Field manual for research in agricultural hydrology. (Agric. handbook
no. 224.)
62. ____ . 1968. Hydrologic data for experimental agricultural watersheds in the United
Sfates 1962. (Misc. publ. no. 1070.)
63. UNITED STATESDEPARTMENT OF THE INTERIOR. 1952. Water-loss ìnvestigations;Lake Hefner
studies. Tech. report. (USGS prof. paper 269.)
64. ____ . 1965. Discharge ratings at gauging stations. Surface wafer techniques, USGS.
262
65. UNITED STATESNATIONAL RESEARCH COUNCIL.1948. Rock coloiir chart. Washington,D.C.,

Division of Geol. and Geogr.
66. VISSER, W.C. 1964.Moisture requirements of crops and rate of moisture depletion of the soil.
Wageningen, The Netherlands,Inst.for Land and Water Management (Res.tech.bull. 32.)
67. WORLD METEOROLOGICAL ORGANIZATION. 1962.Guide to climatic practice. (WMONo. 100,
TP 82.)
68. . 1965. Guide to hydrometeorological practices. (WMO No. 168, TP 82.)
69. ANON.1957. Gidrologicheskie nabludenia i raboty na rekakh. In Nastavlenie gidrometeo-
rologicheskim stantsiam i postam, vyp. 6, chast 1 [Manual for hydrometeorological stations
and posts, vol. 6, chap. 11. Leningrad, Gidrometeoizdat.
70. ~ . 1958. Meteorologicheskienabludenia na stantsijakh.In Nastavlenie gidrometeo-
rologicheskim stantsiam i postam, vyp. 3, chast 1 [Manual for hydrometeorological stations
and posts, vol. 3, chap. i]. Leningrad, Gidrometeoizdat.
71. -- . 1966. Metodicheskie ukazenia Upravleniam Gidromstsluzhby [Methodological
instructions to the offices of the Hydrometeorologic Service]. No. 73 : Raschet vodnogo
balansa rechykh vodosborov. Valdai, GGI.
72. - . 1952. Nabludenia na molykh rekakh. Nastavlenie gidromefeoroíogicheskim
stantsiam i postam, vyp. 6, chast 2 [Manual for hydronirteorological stations and posts,
vol. 6, chap. 21. Leningrad, Gidrometeoizdat.
73. ~ . 1961.Nabludenia nad ispareniem s vodnoi poverkhnosti.In:Nastavlenie gidro-
meteorologicheskim stantsiam i postam, vyp. 7, chast 2 [Manual for hydrornetorological
stations and posts, vol. 7, chap. 21. Leningrad, Gidrometeoizdat.
74. ~ . 1955. Rukovodstvo PO kontrolu i obrabotke nabludeniy nad vlazhnostj i promerza-
niem pochvy [Guide on the control and processing of soil humidity and soil freezing obser-
vations]. Leningrad,Gidrometeoizdat.
263
6 Analysis techniques
and interpretation
of research results
6.1 General
The purposes of analysis on representative and experimental basins are:(a) development
of quantitativefactors of the water balance;(b) development of relations between basin
characteristicsand hydrologicalcharacteristics;(c) elucidationofthe hydrologicalcharac-
teristics themselves;and (d) the development of relations between the elements of the
hydrological cycle.
Ideal objectives of such analysis can be summarized as follows:
1. To define a basin by a set of quantitativeparameters,thus permitting translation of
research results to other basins.
2. To define as accuratelyas possible the relevantbasin and climatologicalcharacteristics.
3. To determine mathematical models representing some or all of the hydrological pro-
cesses on representativeand experimental basins.
4. To develop in some caseswater belances for basins.
Although these ideal objectives are worth aiming at, neither the development of satis-
factory practicalmodels,nor the derivation of a quantitativeindex for each physiograph-
ical element of the basin is easy.It is always necessary to consider the ultimate translation
of results to other basins and the hydrological effects of natural and cultural changes on
hydrological prediction and translation of results to other areas.
Research on representativeand experimental basins is more and more directed to a
description of the hydrological processes by complex model systems (see section 6.2).
This tendency should not discourage organizations which are utilizing representative
basins but have not yet the means to permit more complex research.Reasonableresults
may often be obtained by more simple means but,in such solutions,it is important that
the scientistcompareresults with those obtained by previous experience.Simplermethods
are useful but require a great deal of common sense,
6.1.1 Basin characteristics

These comprise vegetational,geomorphological,pedological and hydrogeological char-
acteristics and determine, under the influence of the prevailing climate, hydrological
characteristicsand hydrological relationships in any one basin.It is therefore most impor-
tant that detailed research is undertaken to quantify all basin characteristicswhich may
be relevantin any hydrological study.Basin characteristics which have been used in the
past are generally the more simple ones and it may be necessary to consider the use of
new characteristicsor alternativemethods of determination of known ones.
264
I Analysis techniques a n d interpretation of research results
I To aid in translation of data to other basins,basin characteristicsfor ungauged basins

I . may have to be determined.Since in such cases the determination of basin characteristics
is not necessarily done by the hydrologists themselves,it is essential that the latter specify
precisely the characteristicsto be measured and the methods of determination.
6.1.1.1 Vegetation characteristics

Utilizing measurements of phytomorphological characteristics(see section 4.6)and vege-
tation surveys (see section 5.1.4),vegetation should be considered in two ways: (a) as a
primary factor where vegetation may be considered a basin characteristicaffecting hydro-
logical processes such as the amount of net precipitation, the redistribution of snow,
evaporation,overland flow,etc.;(b) as a secondary factor where vegetation serves as an
indicator showing rates and depths of soil wetting, area and duration of flood-plain
submergence,etc.
For the proper analysis of the effect of vegetation,it is essential to take into account
not only the characteristics and absolute area of a given vegetation in a basin but also
its location in the basin with respect to the basin outlet (for instance,a given forest area
in a basin may affect flow differently according to whether it is located near the outlet
or in the upper part of the basin).
A useful analysis method used in the U.S.S.R.is to plot mass curves for areas occupied
by a given type of vegetation (Fig.6.1).Analysis of such graphs,together with observed
phytomorphological characteristics,is a suitable way of quantifying vegetation character-
istics for subsequentderivation of relations with hydrological elements.
Legend
1. = total forest area.
2. = deciduous forest area.
3. = coniferous forest area.
N
E
Y
o
L
+
YI
'u
U
, FIG.6.1. Mass curves
of forest areas in a basin. Basin oreo ( k m * )
It is importantin mountain regions to note the adaptation of the vegetation to exposure

and altitude. As snowmelt characteristics are different for sunny and shady slopes,this
is especially relevantin regions with a seasonal snow cover. Moreover, vegetation varia-
tions,because of exposureand altitude,may affect infiltration,overland-flowand erosional
processes.
For scrub and herbaceous vegetation associations,it may be important to determine
the amount and rate of soil fixation during the experiment.Such studies,especially rele-
vant in arid,semi-aridand mountainous regions,are essential where erosion research is
undertaken.
Representative .and experimental basins
6.1.1.2 Geomorphological characteristics
The aim of geomorphological studies on representativeand experimental basins is the

determination of quantitative characteristicsto be related to hydrological factors and to
data on erosion and sedimentation [30].
Only some important characteristicsare discussed here. For a complete description of
geomorphologicalcharacteristics,refer to the literature [18,62,105,1121.
6.1.I .2.1 AREA-ELEVATION C U R V E - MEDIAN ELEVATION

T o obtain the area-elevationcurve,the area between contours is planimetered and each
contour elevation plotted against the area of basin below that elevation.The elevation
at which 50 per cent of the basin area is higher and 50per cent lower is the median eleva-
tion.
6.1.1.2.2 MAXIMUM A N D MINIMUM ELEVATIONS
To obtain the maximum elevation, either the highest surveyed spot height is recorded
or an approximate value is obtained from the area-elevationcurve. Often, the 5 or 10
per cent value of the elevation-areacurve is better for correlation studies.
T o obtain the minimum elevation,the surveyed altitude of gauge-heightzero flow at
the lowest flow-measuringstructure is recorded. The 90 or 95 per cent elevation is fre-
quently a more significant parameter.
6.1.1.2.3 ASPECT
The dominant-trenddirection of the principal channel is measured.Where the basin is

approximately rectangular,the aspect is the trend direction from the maximum elevation
to the minimum elevation. When curvature causes deviation greater than 45" from the
defined aspect,however,a single direction becomes meaningless and component direc-
tions are necessary.Aspect is most significant when related to dominant wind directions
and strengthsand to freeze and thaw cycles (see section 4.8).
6.1.1.2.4 MAXIMUM VALLEY SIDE-SLOPE
The maximum valley side-slopeis obtained by selecting the highest slope value normal
to the contours. Measurements should be made at intervals along the steepest valley
walls from the divide to adjacent channels,giving the equation:
f = II// m per ni (1)
where Zis the maximum valley side slope,h the vertical height and I the horizontal dis-
tance from the channel to the divide.
6.1.1.2.5 M E A N SLOPE C U R V E
This curve may be determined by planimetering the area (a)between two adjacnt con
tours and measuring the length (A) of each contour with an opisometer. The area (a)is
then divided by the mean contour length,+(A + A), to obtain the mean belt width. Be-
cause the length of a contour is rarely well defined,care is essential. Too simple or too
complex contour lines must be avoided.The contour heights are plotted on the ordinate
and the accumulated mean-beltwidth on the abscissa.The mean slope of the curve drawn
through the plotted points is the mean slope of the basin (see Table 6.1 and r ig.6.2)
266
Analysis techniques and interpretation of research results
TABLE6.1
Mean contour
Contour height Contour length L length +@+A) Area LX Mean-belt width
ím) ím) (m) (mz) (m)
268.5 45 - 263 -
270.0 128 86 821 9.45
271.5 223 176 1371 7.85
273.0 193 208 3095 14.89
274.5 153 173 1675 9.66
276.0 156 155 1533 9.86
277.5 142 149 1412 9.45
279.0 130 136 1331 9.66
280.5 60 95 1355 13.88
282.0 10 35 732 20.72
-
p 285
1
I-
- 1
I-Q
f 200
I
c
275
5?
W
I
2 70
œ
FIG.6.2. Mean slope O 20 40 60 00 100

curve. ACCUMILATED MEAN-BELT WIDTH (METRESI
6.1.1.2.6 HYPSOMETRIC C U R V E
To determine the hypsometric curve,the continuous function relating the relative height
(y) to the relative area (x) is plotted,y being the ralio of height of a given contour (h)
to total basin height (H) and x is the ratio of horizontal cross-sectionalarea (a) to the
entire basin area (A) (see Table 6.2and Fig. 6.3).
TABLE6.2.
O 0.00 0.00 0.00

3 0.06 2.60 0.02
8 0.16 10.70 0.08
13 0.27 24.27 0.18
18 0.37 54.88 0.41
23 0.48 71.42 0.54
28 0.58 86.60 0.65
33 0.68 100.58 0.75
38 0.74 113.72 0.85
43 0.90 126.91 0.95
48 1 .o0 133.15 1 .o0
267
6.1.1.2.7 SLOPE INDEX
The slope index is derived from two concepts, those of the equivalent rectangle and the
hypsometric surve.
The equivalent rectangle is a rectangle having the same area, the same hypsometric
distribution and the same coefficient of Gravelius as the basin, and a similar drainage
density and vegetation distribution.
100 -
90
70
60
-
o
%SO-
FIG.6.3. Hypsometric 5 ro,io(Horizontol cross-sectional ores

curve. 2: 100)
A Total borin oreo
If P is the perimeter of the catchment, A the basin area and K the coefficient of com-
pactness of Gravelius (K=0.28P/1/A), then the length and the width (I) of the equivalent
rectangle are given by :
L=- ""[1/
1.12
1+ 1 - l-z);1( -
and
It should be noted that, by transformation, the contours become straight lines parallel
to the smaller side of the rectangle. If ao, ...at, ...an are the elevations of the different
contours, then the average slope of the area between the lines marked a<-1 and a< is
-
(a< UZ-I)/X~, where xt is the distance between the lines on the rectangle. The so-called
slope index is given by:
If Bt is the fraction of the total basin area lying between the contours a* and ut-I,then:
268
The advantage of such an index is that the entire hypsometric distribution is taken into
account and no delicate or uncertain operation such as measuring lengths of contours
(which depends greatly on the accuracy of the contour lines on the map) is required.
6.1.1.2.8 D R A I N A G E CHARACTERISTICS
The maximum channel order may be derived by assigning order one to all unbranched
tributaries as determined from aerial photographs (maps frequently give insufficient de-
tails) and, where two first-orderchannels join, assigning order two to the composite
channel and so on (Fig. 6.4).The maximum order that occurs in the basin is recorded.
FIG.6.4.Channel order
and principal channel
length.
The bifurcation ratio may be calculated by taking the antilogarithm (base ten) of the
regression coefficient from a plot of the channel order against the number of streams
of each order (Fig. 6.5):
Channel order Number of sircams
1 20
2 6
3 1
The total length of channels of a given order is calculated by accumulating the lengths
(in metres or kilometres) of individual segments for that order.The first-ordersegments
are extended to the basin divide or to the internal divide as appropriate.
The length of the principal channel is defined by taking the distance (in metres or
kilometres) from the gauging station upstream of the main divide, selecting at each
junction the segment of the highest order. The principal channel normally begins at the
maximum elevation of the basin (see the unbroken line in Fig. 6.4).
269
FIG.6.5. Bifurcation
ratio.
CHANNEL ORDER
The drainage density is found by dividing the total length (L)of the channels in kilo-
metres by the basin area (A) in square kilometres.
The average junction angle is defined by measuring all angles between individual
channels and the principal channel at and immediately above the confluence. The stat-
istical average is recorded (see Table 6.3 and Fig. 6.6).
TABLE6.3
Below 60 m (degrees) Above 60 m (degrees)
1. 76 14. 45
2. 58 15. 25
3. 45 16. 25
4. 65 17. 27
5. 20 18. 25
h. 62 19. 32
7. 52
8. 12 Average: 29.83
9. 26
IO. 33
Average: 46.25 Average junction angle: 38.6"
6.1.I .3 Pedological characteristics

Soil-physicalproperties gave been discussed in section 4.7and soil surveys in section
5.1.3.The most suitable soil-mappingunit is the series because it is based on the morph-
ology of the profile-the number of horizons; their thickness; the physical, physical-
chemical and chemical properties of the horizons.
The analysis of pedological data should begin with a study of the spatial distribution
of the soils by calcualtion of the percentage of surface occupied by each soil series,
Such a study acquires greater significance when the distribution of soils is combined
270
Analysis techniques and interpretution of research results
with the distribution of vegetation in order to determine soil-vegetationcomplexes (see

section 2.2.3.)
Soil-physicalproperties are important for a study of the division between infiltrated
water and overland flow. Methods of analysis and interpretation on a basin-wide scale
still, however, remain to be clarified and research to this end is being undertaken at
present.
If an experimental or representativebasin is relatively uniform and has only one soil
type,it is not likely that the soil-physicalproperties will vary much from one point to
another.The sampling density can then be fixed in relation to the scale of the soil map.
If it is found that the soil-physicalproperties do indeed vary little, an average value
may be adopted for the whole basin. The comparison from basin to basin with climatic
and hydrological characteristicswill enable a determination of the relevant soil proper-
ties and their importance. Research on the influence of the soil on overland flow and
infiltration should therefore be done initially on a single site (e.g., on a master site:
see section 4.1.,)I a run-offplot or a small pedologically homogeneous basin.
FIG.6.6. Average
junction angle.
A n easy solution is not available when a basin is pedologically complex. It may be

possible to determine the physical properties by mapping and then to analyse the data
by statistical methods (analysis of variance).
Finally,the rate of erosion (see section 4.9.2)is frequently related to the following
characteristics: rainfall, soil (detachability, structural stability, granulometry), slope
(steepness and length), and vegetation (natural or planted by man).
6.1.1.4 Hydrogeological characteristics

Observations of hydrogeological characteristics are discussed in sections 4.4.2and 5.5.2
and hydrogeological mapping in section 5.1.2.
271
For water-balance studies, inflows, outflows and storage in aquifers are importan1
and, for a reasonable determination of such variables, characteristics of aquifers which
are connected to the relevant stream or reservoir should be determined.Such aquifers
should inchide ephemeral zones of saturation (see section 1.5).
Studies to determine which aquifer characteristics are most relevant in the supply of
base flow or interflow should be undertaken. Such characteristics should include the
location of aquifers the lithology of water-bearing strata, the depth of the ground-
water tables, the thickness of the aquifers, etc.
Simultaneous studies can be made of the ground-water balance in the regimen of
subsurface flow (including the temperature regimen and the chemical composition of
ground water).
Where confined aquifers occur, it may be important to study the contribution of
artesian water to flow and,in such a case,piezometric heights and their variations,and
the chemical composition of artesian water should be studied.
Generalhydrogeological characteristicsused in analysis are discussed in section 6.1.4.
In addition,data on discharge fluctuations of springs and on cultural changes affecting
aquifer performance(irrigation,water supply,construction of reservoirs,drainage,etc.),
may have to be studied.
In influent stream systems,channel losses in relation to subsurface flow may require
investigation.This should include the location of the aquifer and the depth of erosion
cut of the channel system.
6.1.2 Analysis of climatic characteristics a n d energy balance

The analysis of climatic characteristics generally involves the following different oper-
ations: (a) the definition of the climatic characteristics of a basin; (b) the study of vari-
ations of climatological factors in relation to hydrological characteristics or processes
(for instance, a flood hydrograph is analysed together with the storm generating the
flood hydrograph); (c) the establishment of an energy balance for the basin.
6.1.2.1 Definition of the climatic Characteristics of a basin

Detailed observations and data-processingtechniques are given in sections 4.2 and 5.3.
These include a description of methods to determine mean basin precipitation values.
For other climatic characteristics,data from one climate station are normally available.
Where auxiliary climate stations are employed (see section 4.2.5.2.1.2),the basin climatic
characteristics should be determined according to the soil-vegetationcomplexes occur-
ring in the basin, which are expected to have their own micro-climate.It is important
to consider a range of moisture conditions, each of which can be expected to have its
own climate.
Since, because of the short-termrecord,climatic observations on representativeand
experimental basins are not likely to produce,at least initially,accurate average values,
it is essential to establish correlations with the long-termrecord of the nearest national
climatological station.
It may not be necessary to study all climatic characteristics in great detail. In regions
with heavy precipitation and relatively low temperatures, for instance, evapotran-
spiration is not likely to be an important characteristic.
6.1.2.2 Study of variations of climatic characteristics in relation to hydrological

characteristics or processes
Detailed analysis methods are discussed in other sections of this chapter and relations
between characteristics are discussed in section 6.2.It is frequently difficult to specify
212
I Analysis techniques and interpretation of research results
sidered an important fact,for instance,it may not be certain whether maximum inten-
sity for the storm,average intensity for the storm,or average intensity for any period
of the storm is the most relevant and some or all these characteristics may have to be
considered.
6.1.2.2.1 PRECIPITATION
Main precipitation characteristics include intensity, volume or depth of precipitation,

the spatial distribution and the antecedent precipitation. Rainfall intensities should be
calculated over periods to suit any particular analysis (see section 5.3.1). For rainfall-
run-offrelations on representativebasins,quarter-hourhalf-houraverage intensities will
normally be found satisfactory.For studies on experimental basins (for instance,infil-
tration analysis or erosion research), smaller time intervals may be required.It is impor-
tant to consider whether the automatic raingauge density is sufficient to give,over small
time intervals.intensities that are typical for the entire basin.
The calculation of volumes or depths of precipitation is described in section 5.3.1.
Sometimes it is desirable to use only volumes and depths of ‘effective’storm rainfall-
i.e., that part of the storm which directly or indirectly contributes to run-off.In this
case,low-intensityrainfalls are assumed to be accretion to soil moisture only.
The spatial distribution of precipitation may be of great importance. Coefficients de-
fining the heterogeneity of the precipitation or an index corresponding to the position
of the epicentre of the storm can be used. A simple method, useful in any correlation
study for large or non-homogeneousbasins, is to divide any basin into sections and to
allocate an index to each section according to the position of the epicentre of the storm
under consideration.
Antecedent precipitation is often characterized by an index equivalent to the soil-
moisture index. Since in any translation of results to other basins,precipitation data is
likely to be the only data available,such an index should conveniently involve only
precipitation parameters.Indices are of many types and only thefinal analysis of rainfall-
run-offwill permit a reasonably precise specification of the index and of the value of
the parameters.A typical index is:
Mi = P kt (Kohler index) (6)
where :
Mi = the precipitation index;
P = the storm precipitation in millimetres;
t = the number of days counted before the storm studied(e.g., t = 1 for the first day,
t = 2 for the second,etc.);
kt = a parameter which varies from basin to basin.
1 Values for kt are given in the literature [65].
Another index frequently used is:
P
Mi=-$ (7)
The parameter n varies between 0.5 and 1 for various basins.In this index Mi does not
converge,but for basins in countries with dry seasons the computation should be stop-
ped at the beginning of the rainy season,without considering the previous rainy season.
In certain cases simplified forms may be used. If, for instance,the storms are of a
similar type and the soil is relatively impermeable,the time t a elapsed between the flood-
producing storm and the first previous storm causing a hydrograph rise may be used.
273
Representative and experimental basiris
Indices may be made more comprehensiveby using soil-moisturedata and base-flow

discharges.
6.1.2.2.2 SNOWMELT
Snowmelt studies involve relations between snow characteristics (snow depth, water
equivalent,nature of snow) and relevant climatic characteristics (air temperature,solar
radiation,wind, etc.). Daily variations, e.g.,of temperature,should be considered as
well as averages in any analysis [58].
6.1.2.2.3 TEMPERATURE
Temperatures are especially studied in relation to snowmelt and evaporation,but it in

likely that they play a significant role in infiltration processes also.
In any relations involving temperature,it is important to consider that temperature
values observed in any one basin are frequently point values and may not necessarily
representaverage basin values. Moreover, in snowmelt-evaporationstudiesit is frequent-
ly difficult to decide on the best index of temperature representing a daily value.
6.i .2.2.4 SOLAR RADIATION
For representativebasins, the observation of one climatic station may generally be re-
garded as typical for the basin, although problems may arise in mountainous regions
where the angle of incidence has to be taken into account in analyses (e.g.,in snowmelt
studies); for details,see section 4.2.5.3.
In experimental basins,'datafrom auxiliary stations should be used as discussed in
section 4.2.5.2.1.2.
6.i .2.2.5 HUMIDITY
Humidity data are important in evaporation studies where problems similar to those
discussed in section 6.1.2.2.3arise.
6.1.2.2.6 EVAPORATION
A problem in research involving evaporation data is often the utilization of data ob-
tained by various methods (see section 4.2.4). It is generally useful to compare with
formulas utilizing climatic elements, data from evaporation pans and evaporimeters.
If potential evapotranspiration data produces acceptable results, however, it does not
automatically mean that a workable relation exists between potential and actual evapo-
transpiration for all climatic conditions [59, 82, 1001.
6.1.2.3 Energy-balance
Energy-balancecalculations are discussed in section 5.3.5.5.These are particularly use-

ful for snowmelt and evaporation studies. .
6.1.3 Surface-water
Surface-waterstudies encompass the entire field of surface-flowphenomena. Analytical
work has been restricted mainly to recession studies, hydrograph analyses,infiltration
analyses and unit hydrographs. Not considered are individual studies which relate
214
hydrological characteristicswith each other or with basin and/or climate characteristics;

some aspects of these (component models) have been referred to in section 6.2,in which
unit hydrographs are also discussed. Frequency analyses have not been dealt with, but
detailed treatments are given in the literature [48,1361.
Before surface water analyses are carried out on data from representativeand exper-
imental basins,it is useful to classify a basin hydrologically even if the research has a
specific aim. This will allow a correlationwith other hydrological data available during
the course of the analysis and will facilitate the translation of results. Such a classifica-
tion is made using standard hydrological characteristics (a mrthod similar to those ap-
plied in climatology).
6.1.3.1 Standard hydrological characteristics
The most common hydrological characteristicsare run-offvalues, mean discharges and

low-and peak-flowoccurrence and frequency. Derivation of such characteristics is de-
scribed in Chapter 5 and in the literature [48,1361.
If the observational period on a representativeor experimental basin is short,exten-
sion of records correlating data with those from reference stations is necessary. Long-
record precipitation stations may usefully serve as such stations.
With the above characteristics,the hydrological regimen of any one basin may be
determined.
6.1.3.2 Various forms of flow, recession studies

A hydrograph recorded at any one gauging station is normally composed not only of
surface flow but also of baseflow and interflow and comprises an ascending and a de-
scending part.
The rising side of the hydrograph is known as the concentration curve and the falling
limb as the recession curve. The latter curve represents the diminishing discharge from
storage in the absence of further replenishment. Figure 6.7is a simple representation
Surface-flowrecession
(surface detention)
Storage in channels
(channel detention 1
(7
Interfbw recession
FIG.6.7. Simple representation of part of the hydrological cycle [121].
of part of the hydrological cycle and indicates the various storages from which stream-
flow is supplied.All recessions represent withdrawal from storage modified by channel
storage;the shape of each recession therefore depends not only on the nature and ex-
tent of the storage reservoir,but also on the nature and extent of the channels through
which the flow is routed.
For small basins with an area of less than 10-20km2,channel storage phenomena
are frequently neglected; this applies particularly to arid regions when small storms
275
Represeritative and experimental basins
produce surface run-offwithout filling the channel and without influencing the record
of the gauging station.
Surface flow,if generated by itself, produces a surface-flow hydrograph, the reces-
sion of which represents depletion of surface detention. Interflow is the outflow from
ephemeralzones of ground-waterbut is sometimes indistinguishable from and combined
with the surface flow. The combined flow is called direct flow, and has an associated
direct-flowrecession.Bank storage is sometimes considered separately ; it is the storage
caused by flow into the banks of a stream during rising stages and will deplete during
falling stages, sometimes causing a recession similar to that of interflow.Bank storage
from effluent streams draining relatively small basins may be insignificant,but it may,
however,become important when the point at which the stream changes from effluent
to influent moves considerably downstream during or shortly after heavy rain.
Baseflow is the outflow from more permanent ground-waterzones and normally forms
the tail end of a hydrograph;this tail end is referred to as the baseflow recession (see
section i.5).
Recession studies are invaluable in representative and experimental basin research.
Much work has been done on baseflow recessions for low-flow prediction and basin
comparisons.Research into surfaceflow and interflow recessions is required to improve
techniques of hydrograph analysis (see section 6.1.3.3).Moreover, storage curves,which
may be obtained by integrating surface storage (surface detention), aquifer storage and
soil storage respectively.
Distinction between various recessions is frequently very difficult. Subsurface-flow
studies (see section 6.1.4)will aid research in interflow and baseflow characteristics.
6.1.3.2.1 M A S T E R RECESSION C U R V E S
Any storm hydrograph is a short-termevent and its recession varies from the next one
because of variations in storage.A number of hydrographs of varying magnitude (thus
covering a period of varying storage) can be combined to give a master recession curve,
e.g.,combining individual baseflow recessions gives a master baseflow recession curve.
Three construction methods are available.
(a) The correlation method. This involves the plotting of qt against qt+N (qt,I? days
later) on log-logpaper and drawing a straight line or curve through the points plotted.
If a straight line is drawn through the points at 45" to the axis, the slope of the line is
log k and therefore a simple exponentia1 equation is presupposed (see section 6.1.3.2.2).
(b) The strip method [i 351. This involves plotting individual recessions on tracing pa-
per; they are then superimposed and adjusted horizontally until the main parts overlap.
A mean line through the overlapping parts is the master recession curve. This method
is generally accurate because visual control allows omission of those parts of recessions
which are too high (surface flow) or too low (snowmelt). When the individual recessions
are very flat, it is difficult to decide where they fit together, and the resulting master
recession curve may be either elongated or telescoped.This can be overcome by the
use of a large magnification of the discharge ordinate.
(c) The tubulat'ng method [49].This is essentially the same as the strip method. It in-
volves the tabulation of daily mean discharges of individual recessions in columns (or
instantaneous discharges at a fixed time, e.g.,midnight). The columns are adjusted ver-
tically until the discharges agree approximately horizontally.Subsequently the dischar-
ges are averaged horizontally and these mean discharges constitute the master recession
curve. This method gives a good control of the data,making it less probable that the
final curve will be too long or too short.Its disadvantage is that irrelevant parts of the
recession cannot be omitted without a detailed inspection.
To obtain the overland-flowrecession,the surface-flowhydrograph must be corrected
for channel storage.
276
Ana(yJ'is techniques und interpretation of research resulfs
6.1.3.2.2 RECESSION E Q U A T I O N S
Recessions can be represented mathematically with relative ease. Little theoretical work
has, however, been done on them.
The diminishing discharge from a confined aquifer (i.e., baseflow recession) can be
expressed [127]by the following equation:
qb = q o eëat (8)
where:
qt = the discharge at time t;
q o = the initial discharge;
e = the base of the natural logarithm;
a = a constant.
Normally ëa is replaced by k,which is called the recession constant.
Various empirical equations which represent recessionsmore or less successfully have
been proposed and a number of these are given in the literature [121].
6.1.3.2.3 STORAGE RELATIONS
Integration of equation (8) yields:
where S is the change in volume of water storage in the interval between the time of
occurrence of q o and qt, or the baseflow (or interflow or surface flow) discharged in the
interval between the times of occurrence q o and qt.
Equation (9) may be used to estimate the volume of baseflow (or interflow or sur-
face flow) which will be discharged in the time interval between any two discharges.
Taking qt as zero and 90 equal to the selected discharge,the volume of water remaining
in storage (and yet to be discharged) in an aquifer,or soil horizon or as surface deten-
tion,may be estimated.
In cases where empirical equations are fitted to recession curves or where no suitable
equation is available,graphical integration by planimetering the area under the recession
curve at suitable time intervals(starting from the end) and plotting these areas (storage
values) in millimetres depth over the basin against the discharge, can be resorted to.
6.1.3.3 Hydrograph analysis

6.1.3.3.1 F L O W SEPARATION
A total-flowhydrograph can be separated into baseflow,interflow and surface-flowhy-

drographs. By correcting the latter for channel detention (see section 6.1.3.5),the over-
land-flowhydrograph is obtained.
In many cases such a detailed separation is fraught with difficulties and the results
possibly meaningless. In some cases either baseflow including interflow,or baseflow
alone,is separated and the remainder is regarded as surface flow.
For ephemeral streams no flow separation is warranted, except where the surface-
flow hydrograph shows a long recession owing to a prolonged,slight discharge of sub-
surface origin.
Flow separation is a prerequisite for further surface-flowanalysis and is particularly
important in representativeand experimental basins when changes in flow distribution
are studied.
Methods available so far are empirical;assumptions made normally are that the be-
ginning of a rise in a hydrograph corresponds to the beginning of surface flow and that
a break in the recession indicates a change from surface flow to interflow or interflow
I 277
to baseflow. Assuming that a simple exponential equation (equation (8) )fits the reces-
sion,which can be represented by a straight line on semi-logarithmicpaper, detection
of any break is facilitated.
The actual separation consists of connecting the beginning of a rise with a break in
the recession by a straight line [88, 961 or by the alternative methods shown in Figure
6.8 /65,1361. Many other methods have been proposed, but suffer from their subjective
nature. When discharge variations are small, a curve may be suitably traced between
the beginning of the rise and the break in the recession as a function of time encompas-
sing both interflow and baseflow.The curve is drawn from data obtained by a study of
hydrographs which produced interflow only (i.e.,time elapsed between beginning of
Time-
FIG.6.8. Methods of
hydrograph analysis [65].
rise and maximum interflow,ordinates of this maximum, etc.). A close scrutiny of the
hydrographs,whether or not they are composed of interflow only, is essential and cor-
relation with the isohyetal pattern of the storm studied will indicate whether the hy-
drograph is in fact composed of interflow only or of surface flow from the upper pari
of the basin only.
A study of the hydrogeological conditions can be of great assistance and Figure 6.9
shows a number of methods of flow separation with respectto hydrogeologicalconditions
and subsurface ñow towards the stream channel.
278
a b at6 atbtc
II
III
IV
a aquifer the same according to model b
R confining stratum subsurface flow into the stream

from artesian aquifers
0surface run-off ground-water level
subsurface flow into the stream stream-water level

according to model a
FIG.6.9. Typical methods of flow separation with respect to hydrogeologicalconditions and
subsurface flow towards the stream channel:
I. Hydrogeological conditions of streamflow supplied by subsurface water: a = confined
ground-waterwith no hydraulic connexion with the stream bed; b = confined ground-
water with hydraulic connexion with the stream bed; a + b and n + b + c = confined
and artesian water.
II. Stream and subsurface water-levelfluctuations (H)in the stream bank.
III. Fluctuations of subsurface flow into the stream.
IV. Method of flow separation: t = time, q = discharge, T = period of bank storage:
-A and + A = subsurface run-off participating in bank storage.
In the literature [8] measurements of B-radioactivity of rainfall and flow at various

phases of flood waves are reported and such studies may be an excellent aid in deter-
mining factual baseflow,interflowand surface-flowhydrographs.Detailed studieson run-
off plots, designed to measure interflow(s) are also invaluable in studying hydrograph
analysis [1301. Moreover, analytical treatment of ground-watermay offer solutions;
some aspects of these are discussed in section 6.1.4.4.3.
For many analyses the method of separation used is not important, provided a stan-
dard method is used throughout the analysis. A W M O working group is currently
studying flow-separation methods with a view to recommending standard ones.
279
6.1.3.3.2 M I SCELLA N E ou s CH A R A CTERISTICS DE FINING T H E H Y D R O G R A P H
In studies involving the unit hydrograph (see section 6.2.5.3)or the isochrone method
11361,characteristicsother than the standard hydrological ones (see section 6.1.3.1)and
recession curves are frequently used.In such cases these characteristicsare derived from
surface-flowhydrographs; they could also be derived,however,from total-flowhydro-
graphs for use in correlationstudies,especially where relations with climatic and basin
characteristicsare considered.
In the case of the unit hydrograph,the storm is characterized by the depth of surface
run-offin millimetres and the hydrograph is expressed in terms of such characteristics
as time of rise, duration of flow, peak discharge, etc.,for subsequent correlation with
basin characteristics.
Such a characterizationis fairly simple for showmelt hydrographsand for cases where
the basin area is smaller than the area covered by the storm.In other cases complications
are likely to arise and,to avoid a number of trial and error attempts,it is useful to es-
timate in advance the order of magnitude of the characteristics.This can be done either
by estimating the minimum depth of precipitation at which surface flow occurs under
given antecedent moisture conditions,estimating the hydrograph characteristics by em-
pirical methods based on known relations with basin characteristics (synthetic hydro-
graphs) [136].The principal characteristics which can be defined are as follows.
(a) Time of rise. This is the interval between the beginning of a rise and the peak of
the hydrograph following this rise. In simple hydrographs caused by short,high-inten-
sity rainfalls the time of rise is usually less than the rainfall duration.
(b) Lug. This is the time interval between the centre of gravity of the effective rain-
fall hyetograph and the flood peak. On very small basins the lag may be shorter than
the time of rise.
(c) Duration offlow or base length. This is applicable in particular to surface-flow
hydrographs.In French terminology this duration corresponds to the time of concen-
tration of a unit flood.
(d) Peak percentage. This is the ratio of the volume of flow (run-off) occurring in a
given time smaller than the total storm period (but including the hydrograph peak) to
the total storm run-off.Any other dimensionless characteristicwhich expresses the peak
discharge in terms of a given run-off(for instance,the ratio of the peak discharge to the
mean discharge for the storm) can be used and will have a much greater magnitude
than the characteristic derived by the first method (except if a very short base length
for calculating the fractional run-offis used to derive the peak percentage).
According to the theory of the unit hydrograph, these four characteristics are in-
variant for a given basin for a storm of infinitely short duration.This may still be appli-
cable to a storm of a duration less than the time of rise and with a not-too-heterogeneous
spatial distribution.
The theory of the unit hydrograph is, however,an approximation which would only
apply if a number of conditions other than those of homogeneity were fulfilled. One
such important condition is constancy of the overland-flowvelocity irrespective of storm
rainfall depth and intensity and of the vegetation condition.
Since overland-flowvelocity is not independent of these factors, it may be useful in
some analyses to consider various types of hydrograph separately. Where distinctrainy
seasons occur and the vegetation cover is significantly different in the dry and rainy
seasons,hydrographs occurring in the beginning and at the end of the rainy season
can be used.A n alternative grouping is hydrographs with small flow volumes and hy-
drographs with large flow volumes.
If any one hydrograph characteristic appears to have rather large variations for a
given number of storms,it is useful to scrutinize the values in relation to the storm pre-
cipitation. This may aid in eliminating those derived from hydrographs which have
280
Analysis rechniqiies und interpueration of research results
abnormally long base lengths or which have been caused by too heterogeneous a
spatial distribution of the precipitation.
In any research where very large floods are of particular interest,hydrographs which
have been caused by homogeneous, high-intensity storms should be selected, even if
storm duration appears, at first sight,to be too long.
For any one basin either the lowest or the mean value of a characteristic is used in
analysis,depending on the purpose of the analysis.Alternatively the characteristics may
be expressed by curves for small or large flow volumes,etc.
In basins with a snow regimen, similar analyses may be made, but, for such basins
air-temperature and solar-radiationvalues are used rather than precipitation data pro-
vided that the snowmelt causing the hydrograph does not deplete the snowpackentirely.
Where the basin area is rather larger than the area covered by a storm,over-allbasin
analyses may be meaningless and more complex models may have to be used.
Procedures for the isochrone method [136], are not as well defined. The duration of
surface or direct run-offis normally studied for typical homogeneous or localized storms.
Such data are used to estimate the time corresponding to the intervals separating two
isochrones.
6.1.3.4 Volume of flow (surface or direct run-off)
After flow separation has been carried out (see section 6.1.3.3.1)the surface or direct
run-off,expressed in cubic metres or in millimetres depth over the basin area,is deter-
mined (see section 5.4.6).Except in basins with very intense run-off,the surface or
direct run-offwill vary according to the method used for flow separation. A standard
method of flow separation allows a correlation of depth of run-offwith causal factors.
Some workers have expressed this relation in a coefficient of run-off,which is the
ratio of the storm surface run-offto the total storm precipitation [98]. An alternative
approach is to relate the run-offto the effective storm precipitation (see section 6.1.2.2.1),
or singly to the most intense rainfall. This may lead to more significant run-off coeffi-
cients.
Run-offcoefficients for periods longer than the storm are relatively meaningless un-
less the rainfall intensity pattern is similar for the periods for which they are calculated.
More complex relations involve infiltration studies (see section 6.1.3.5) and general
model studies (see section 6.2).
6.1.3.5 Infiltration analysis

The variation of infiltration with time is expressed either in an infiltration-ratecurve,
showing the variation of the rate of infiltration (f)with time, or an infiltration mass
curve,showing the variation of the mass infiltration(F)with time. If the rainfall inten-
sity during the storm is at a rate greater than the infiltration capacity,the inñltration-
rate curve is called the infiltration-capacity curve. The initial rate of infiltration is
termed fo, the ultimate,approximately constant rate of infiltration fi (see section 1.5).
Infiltration analysis may be carried out using rainfall and flow data from sprinkling
plots, run-offplots or small natural basins and determines the variation of infiltration
with time for the duration of the storm.
Results from infiltration analysesvary from those obtained by infiltrometer (see section
4.5)as they give an integrated index for the basin and have errors associated with correc-
tions for subsurface flow and storage.Infiltrometer results give a point value which has
errors associated with lateral flow, splash,air trapped in the soil,etc.
Infiltration values obtained by analysis sometimes include interception,evaporation
during a storm and depression storage. The method of infiltration analysis chosen
281
Representative and experimenral basins
depends on whether the rainfall is artificial or natural and on the size of the plot or
natural basin. Five methods are listed below.
(a) Sprinkling-plotanalysis. This method is restricted to data obtained by the use of
a sprinkling infiltrometer (or when a uniform rate of rainfall occurs).
(b) Run-offplotanalysis. This method, while being best suited to data obtained from
studies of plots, can be extended in many cases to data from natural basins of up to a
few square kilometres,provided that the rainfall intensity for the storm analysed can
be considered uniform over the entire basin area.
(c) Natural basin analysis. This is a more approximate but simpler method requiring
uniformity of rainfall over the whole basin area and separate hydrograph peaks from
separate storm peaks. The method is restricted to basins up to about 5 km2.
(d) Natural basin analysis using a flow-linlitcurve. This method, for rather permeable
soils,eliminates interception and surface storage and results from this analysis may be
compared with those obtained by infiltrometer [15].
(e) Natural besin analysis by time-condenscztion method (Holtan method). This method
is perhaps more subjective than the others.It is,moreover,restricted to basins of up to
a few square kilometres and requires uniform rainfall over the basin.It is furtherrestrict-
ed in that only certain types of storms can be analysed.
Detailed procedures for carrying out infiltration analyses are given in the literature
[i20]. Infiltration analysis cannot be used on data from large basins, firstly because
rainfall is generally non-uniformover large areas,and secondly because the errors asso-
ciated with correcting the total-flowhydrograph to obtain the overland-flowhydrograph
are,in the case of any basin but the smallest,too large to produce worth-whileresults.
The effective basin area may, moreover, be a variable quantity in larger basins (see
section 6.3.2).
6.1.3.5.1 GENERAL M E T H O D O L O G Y
The principle of the analysis is based on the comparison of net rainfall (rainfall minus
interception loss) with its associated overland-flowhydrograph.The difference between
the two,correcting for surface detention (Da) and depression storage (Vd),gives a meas-
ure of the average rate of infiltration in the basin.
In the case of sprinkling-plotanalysis,the hydrograph obtained is the overland-flow
hydrograph and a correction has to be made to the gross rainfall for interception loss
(see sections 4.2.3 and 5.3.3).
When infiltration analyses are applied to natural basins, the total-flowhydrograph
measured may have to be corrected for: baseflow (see section 6.1.3.3.1);interflow (see
section 6.1.3.3.1);channel-precipitationflow; channel detention; and gross rainfall for
interception loss.In addition,correctionsmay have to be made to the infiltration curves
derived if, during a storm,rain occurred at a rate less than infiltration capacity.In this
case an actual rate of infiltration,which may be less than the infiltration-capacitycurve,
is derived. Correction is by so-calledtime condensation.For details,refer to the litera-
ture [120].
6.1.3.5.2 SPRINKLING-PLOT ANALYSIS
A test run and an analytical run typical of those made with the Type-F infiltrometer
described in section 4.5.1.4 are plotted in Figure 6.10,together with their graphicalanal-
yses. Since infiltration is constant throughout the analytical run, accumulated rainfall
(e)
excess (Pe)is determinate.The difference between Pe and accumulated run-off is due
to depression storage (Vd) and surface detention (Da). Approximate separations of V d
and D a can be made by relating D a to the rate of flow (4).During hydrographrecessions,
D a is estimated for any given q as the integral of the subsequent recession curve and is
282
Time (min)
10 I I I I I I I I
1
.I
8-
7-
ó-
5-
4-
/p-F i
Time (min)
FIG.6.10. Test run, analytical run and graphical analyses of Type-F infiltrometer hydro-
graphs [74].
283
Representative and experirnenlal basins
plotted versus q as shown in Figure 6.11.With Du determinate, Vd is computed as a

difference in the analytical run.
The process is then reversed in the test runs to derive accumulated infiltration (F)as :
F=P-Q-Da-Vd (10)
and differences in the Da/q relationship for rising and falling rates of flow are not con-
sidered to be significant errors on small plots. As discussed in section 6.1.3.5.8,the
storage-flowrelationship derived from the recession curve can also be applied to the
rising limb of the hydrograph on larger areas if a hysteresis is induced.
10.0
T7
I
I
I
I
l
FIG.6.11. Detention-
flow relationships for
Type-F infiltrometer run
[106].
.-i,
.- 0.5
L
-
o
.c
0.1
0.01
d
0.05
Averoge depth of deiention,
0.1
Do ( c m )
are based on measurements with a Type-Finfiltrometer.The rates implied by slopes of

these curves are realistic for the soils tested.
6.1.3.5.3 R U N - O F F P L O T ANALYSIS
0.5
,
The curves of Figure 6.12,showing the influence of land use on infiltration capacity,
In the literature [106],a method is described to extend the sprinkling-plotanalysis to

1.0 l .3
cases of variable rainfall on small,homogeneous basins.

In this method the mass-net-rainfallcurve, Pe-Is (rainfall corrected for interception
loss) and the mass-overland-run-off curve QOare plotted. The difference between these
two curves represents surface detention, depression storage and infiltration (Pe- Is-
QO= Da + Vd + F).A relation is developed between the discharge and the surface
detention and this relation is used to correct the Pe-Is-QO curve to give the Vd + F
curve. A subsequent estimate of the depression storage, Vd,produces the F curve.
Any storm and associated hydrograph may be analysed by this method.The nature
of the analysis varies with the hydrograph pattern and therefore some storms are more
suitable than others,Figure 6.13 shows a typical analysis.
6.1.3.5.4 N A T U R A L BASIN ANALYSIS
The principle of the natural basin analysis [46]is that for any storm burst resulting in
a hydrograph of surface flow, one average value of infiltration may be calculated for
284
Analysis techniques and interpretation of’research results
FIG.6.12. Mass infiltration curves based on Type-F infiltrometer tests on Cecil, Madison
and Durham soils [44].
the period of the storm burst. A series of storm bursts and associatedhydrographsmay
therefore be utilized to produce a series of average infiltration values which may be
combined in an infiltration curve [97].
6.1.3.5.5 N A T U R A L BASIN ANALYSIS U S I N G A F L O W - L I M I T C U R V E
Natural basin analysis may, in certain cases,be improved by using so-calledflow-limit

curves.
From the mass rainfall depth (P),a depth POis subtracted.This is the depth of precjp-
itation above which run-offoccurs and is called the ‘limitprecipitation for flow’.PO
is determined by plotting the depth of rainfall on the ordinate and an index of antece-
dent precipitation on the abscissa. Each storm is represented by a point and marked
R if run-offoccurred,N if no run-offoccurred and L if run-offhas been negligible.
The flow-limitcurve is drawn between points marked N and R (Fig. 6.14)115).
The antecedent precipitation index commonly used in arid zones is the number of
285
06’04rm’hi
-P
4-
,
IineC
3- line D
u
C
.-
2-
L
C
3
O
J
il&
a
:1 -
Jz
\
E
C
.-
2 -0 4:OOp.m. -
A U I I
FIG.6.13. Analysis of storm and hydrograph,7 September 1941 ; Control Plot No.6,Lacrosse,
Wisconsin [22].
hours or days (ta) between the storm stuaied and the first antecedent storm that gene-
ated run-off.Any other index, e.g.,the Kohler index,also gives good results.
Neglecting the influence of the rainfall intensity,storm with antecedent index ta will
produce run-offif its storm rainfall depth (P)exceeds the value PO, corresponding to the
value ta or the flow-limitcurve.
The flow-limitcurve is related to a particular soil-vegetationcomplex.If rainfall in-
tensities are plotted for a given storm and a vertical line giving an area to the left of it
equal to the POvalue is drawn,then the area under the curve to the right of this verti-
cal line equals infiltration plus surface flow. O n the remaining part of the hyetograph
the mean infiltration capacities may be determined as described in section 6.1.3.5.4and
Figure 6.15[16].
A curve drawn with various values of F,corresponding to various storni durations,
has always the same origin, indicating similar antecedent moisture conditions.
286
R i6
R9
R II
TL
R 32
+
R I?
+ R 33
c R Flow
+ 278
0L Almost negligiblc flow
/
I N NO flow
/ L 14/6
R ?5
,
r)
R I4
20
''I7 N 19,'
/ a N 16
O 2 1 6 a
ia (Time from previous storm in days)
FIG.6.14. Example of determination of flow-limitcurve for impervious soil (representative

basins of Koumbaka II, Mali).
Infiltration capacities determined in this manner are more easily compared with those
obtained by infiltrometer.The method can also be used for small heterogeneous basins.
It gives good results for very small basins and permeable soils.
6.1.3.5.6 NATURAL BASIN ANALYSIS BY THE TIME-CONDENSATION M E T H O D
In this method [43], the time of the rainfall period is condensed to give a constant rate
of rainfall;by condensing the time of the flow at periods of low rates (where surface
detention is at a minimum) in sympathy with the rainfall time condensation, a curve
of P-Qis obtained. This equals infiltration plus surface retention.A correction for sur-
face retention leads to the infiltration curve.
6.1.3.5.7 INFILTRATION EQUATIONS
Infiltration-capacilycurves may be expressed by exponential equations. A common em-

pirical one is the Horton equation:
f=fc + (fo-fc)e-L'" (11)
where fis lhe rate of infiltration at time t, e the base of the natural logarithm and
C a constant. The equation is widely used because it is simple to apply and has the
287
advantage that, in the limit and as t approaches infinity,the infiltration rate does not
become zero. It does not fit well when infiltration rates decrease rapidly and three
parameters are needed to express any particular curve.
Another equation [53]is:
F = cta (12)
where Fis the mass infiltration and c and a are constant.The equation for the rate of
infiltration is obtained by differentiation:
Its scope is also limited,because the value of a, obtained when the equation is fitted.
depends on the range of t. A theoretical derivation is given in section 6.1.4.3.4.
n
40
20
O
17 ie 19 Time (hours)
FIG.6.15. Example of utilization of fiow-limitcurve (storm of 4 September 1959 on repre-

sentative basin of Barlo, Chad; mass rainfall 27.1 mm).
The following alternative equation was recently derived [83]:

F = St * -At
or upon differentiation:
f=+St-&fA
where S denotes sorptivity (which is a measure of the absorption and desorption of

water-capillaryuptake or removalof water). The equationis a simplificationof one deriv-
ed by considering infiltration as a phenomenon of flow in porous media and employs
288
the concept that soil-water movement may be satisfactorily expressed by equations of

the diffusion type. For details, see section 6.1.4.3.4.
6.1.3.5.8 STORAGE-FLOW RELATIONSHIPS
Surface-detentiondata is obtained as a by-product from several methods of infiltration

analysis. Surface detention,Da,is plotted versus the discharge go in Figure 6.16 to illus-
trate the hysteresis so often encountered in storage-flowrelationships.The rising path is
variable, evidently a function of the rate or height of rise, but the recession path is quite
stable.
A procedure for inducing hysteresis by two sequential routings of rainfall excess (Pe
through half the storage indicated by the storage-flow relationship from the recession
curve) has been developed. Routing is by simultaneous solution of the continuity equa-
tion and the storage-flow relationship [45].
FIG.6.16. Developing
q versus D u relation
[122]. Procedure: plot
points and when only a
few points are plotted
and no very definite
trend is apparent, as in
this example, draw upper
-5
and lower envelope lines 5
as shown, using a slope ;
,
of 5:3. Draw the line of
relation for q and D a , -
i
o
using a slope of 5:3 at
half the distance between li: c
the envelopes. Surface detention, D a (cm)
For many basins, the recession-storage-flowrelationship is generally linear with zero

intercept. Then, for the first routing:
Da
Pe-q,= 7
and
Da -mql
or
Da - mq,
2 2
289
and, for the second routing:

ADa
41-92 =
and
Differences between Pe and the integrated q 2 curve plotted versus q 2 give the type of
hysteresis observed in Figure 6.16.
The effect of natural or cultural changes in basins, may, logically,have great effect
on surface-detention-discharge relationships.The occurrence of hysteresis and the diffi-
culty of deriving such relationships accurately with infiltration analyses make the use of
these in predicting effects rather hazardous.
A n alternative method is to select a number of hydrographs which have uninterrupted
recession from the peak to the end of surface run-off.In such cases surface flow past the
peak is depletion from surface storage only. Correcting the hydrograph for subsurface
flow and channel detention (see section 6.1.3.5.1)allows the development of a relation
between the run-offoccurring after the peak (surface detention) and the peak discharge.
Such a relation is quite stable for relatively stationary conditions in a basin and expres-
ses the relation between the maximum surface detention occurring during storms and
the peak discharge (and not,as in infiltration analysis,a relation between surface deten-
tion and discharge occurring during a storm). This latter relation is subject to great va-
riation from storm to storm owing to rainfall intensity,vegetation and soil conditions.
A n example is given in Figure 6.17.
Equation (9)indicatesthat a simpleexponentialequation may be fitted to the overland-
flow recession (see section 6.1.3.2.2).Experimental evidence, however, indicates that
equation (9) should be written:
q = K Dam. (21)
It has been found that rn = 3.0 for laminar flow and approximately 1.67 for turbulent
flow [47].
o IO..
0.09-
0-08 -
0.07 -
0-06 - p :0-0126 Di
0.05 -
O04 -
.
003-
E
v
e
o 0.02-
P
2=
.VI
-
FIG.6.17. Relation U
between peak discharge 5
and maximum surface 2 0-01
.e .91.0 2 3
I
5
I
6
I I
7 B9Ii
I I
290
I
I 6.1.4 Subsurface water

l
6.1.4.1 Generai
~
Subsurface water comprises all water below the surface of the ground,A useful way of
classifying subsurface water is to consider saturated and unsaturated conditions.When
soil horizons or strata are saturated,ground water is found;when they are unsaturated,
subsurfacewater is designed as soil water and intermediate water. The boundary between
ground-waterand soil-intermediatewater is a ground-watertable (see section 1 S).
In the unsaturated zone,infiltration and capillary rise are the most important phenom-
ena, apart from temporary storage of water. Ignoring the flow towards plant roots,
which is associated with water uptake by plants and evapotranspiration,the prevailing
flow is in a vertical direction.
The flow of ground-waterdepends on the readiness with which strata can transmit
water and hence on the sequence,thickness and permeabilities of the various strata or
soil horizons. Here horizontal flow will prevail.
For ground-waterflow a large number of analytical solutions for various boundary
conditions is known, while for unsaturated flow only some rather simple steady-state
analytical solutions and a limited number of numerical solutions are available,owing to
the inconstancy of the capillary conductivity.
Studies of subsurface water in the unsaturated zone consist mainly of an interpreta-
tion of rates and volumes of infiltration and computations of a possible rise of capillary
water. The latter is associatedwith water-uptakeby plants (see sections 4.4.1,4.7, 5.5..)i
Ground-water studies concern themselves with the determination of the subsurface
flow boundaries of a basin, an evaluation of the direction and intensity of flow and
evaluation of the hydrogeological characteristics.
For special investigations such as pumping test, well drilling,etc.,and for subsurface-
flow measurement,see sections 4.4.2,5.1.2,5.5.2.For detailed treatment of subsurface-
flow problems, refer to standard textbooks [41,75, 87,1191. Section 6.1.4.2deals with
the principles of subsurface flow,while sections 6.1.4.3and 6.1.4.4give some examples
of unsaturated and saturated-flowanalyses.
6.1.4.2 Principles of subsurface flow

An understanding of the principles of subsurface flow is important to explain clearly
the differences between saturated and unsaturated flow and to indicate possibilities of
analysing the flow of water.
Flow of water in the soil is described by Darcy’s law [23,931. This law states that
the flow velocity is directly proportional to the driving force and therefore the flow is
of the laminar type. The general form of Darcy’slaw is:
I where the minus sign is introduced to give a positive-flowvelocity in the direction of

decreasing potential 4. Expressing the potential in terms of a length of water column,
the factor K is called the hydraulic conductivity and represents the ratio of the flow
~
velocity to the driving force, the latter being the gradient of the velocity potential 4.
Since the potential gradient is dimensionless(L.L-l)K has the dimensions of velocity
I (L3.L-2.T-1).
K may be defined as the rate of flow per unit cross-sectionalarea perpendicular to
the flow direction under the influence of a unit driving force. Since flow takes place only
l
I in the water-filledinterstices and pores between the soil particles, and since in the case
l of saturated flow all these voids remain completely filled,K is a constant.
~ 291
In the case of unsaturated flow, part of the voids will be filled with air instead of
with water.Under these conditions the part of the cross-sectionalarea which is effective
in transmitting water will be smaller than that in the case of saturated flow.Moreover,
the effective part of the cross-sectionalarea,i.e.,the part of the pores filled with water
decreases with decreasing moisture content and hence, in the case of an unsaturated
zone,K is not a constant but a function of the moisture content.In this case K is called
the capillary conductivity.As saturation is approached,the capillary conductivity ap-
proaches the hydraulic conductivity (see section 1S).The velocity potential + consists
of two parts: one due to pressure (p) and the other to elevation (b). Therefore:
Ifp is expressed in length of water column, 4 also has the dimension of length and then
it represents the piezometric head, i.e.,the height of water in a piezometer placed with
its bottom at the point under consideration.The value of p then represents the height
of the water column in the pipe,
Soil water is held by forces of adsorption cohesion and solution and is therefore not
capable of doing as much work as pure, free water. For most flow problems it suffices
to deal with forces due to the attraction of the soil matrix. In order to cause the water
to move out of an unsaturated zone against the forces of the soil matrix,a certain suction
or tension must be applied.
In other words,soil water, in contrast to ground-water,has a negative pressure com-
ponent of the potential. Normally the matricial suction designated by y is taken as a
positive value and therefore equation (23) for unsaturated flow should be read as:
+ =-p + h = yfh. (24)
The flow of water in an unsaturated soil must obey the law of conservation of matter.
Hence the principle of continuity must be valid, and therefore:
av - --
_ a+
as at
which expresses the idea that the difference of the flowinto and out of an element equals
the rate of storage.In the latter equation 4 represents the volumetric moisture content
(i.e.,the volume of moisture per unit volume of soil); v stands for the flow velocity, s
for the flow path, and t for time.
Substitution of equation (25) into equation (22) yields:
a
!?al = -(K grad 4).
as (26)
Considering one-dimensionalflow only, we obtain for vertical flow into the z direction,
where M/az = 1,
at az
where z is taken to be positive in the upward direction. For horizontal flow, equation
(27)reduces to :
at ax
292
Analysis techniques and inierpretation of research results
Equations (26),(27) and (28) are the general equations for unsaturated flow. They can-
not be solved analytically,since K is a function of w. Moreover, hysteresis may occur
and then a distinction between sorption and desorption should be made.
The concept of diffusion in soil water movement is introduced by substituting [19]:
where D is called the diffusivity (dimensionsL2.T-l).

The term òw/ò4representsthe slope
of the so-calledsoil-watercharacteristiccurve,i.e.,the curve giving the relation between
the soil-watercontent and the corresponding suction y. Since D is a function of 4,the
water movement is a concentration-dependentprocess.
Using equation (29), equation (27) becomes:
(30)
For horizontal flow,the last term of the right-handside of equation (30) vanishes again.
Nevertheless,the same problems as mentioned previously apply for the solution of this
equation.
6.1.4.3 Unsaturated-flow analysis

Because of the inconstancy of the capillary conductivity and the time variability and
complexity of boundary and initial conditions,unsaturated-flowanalysis is difficult,and
known evaluation methods are limited. As an example, a discussion will be given
here of vertical upward and downward flow or infiltration (omitting,for instance,flow
towards plant roots).
Both rate and volume of infiltration determine to a large extent the magnitude of
overland flow.O n the other hand,infiltration governs the recharge of ground-water.In
fact,the unsaturated zone is the area of greatest hydrological activity and is therefore
of utmost importance in representativeand experimental basin research.
6.1.4.3.1 TYPES OF INFILTRATION
Vertical downward flow or infiltration may be divided into three types: (a) infiltration
at a constant rate;(b) ponded rainfall infiltration;and (c) ponded infiltration.
Type (a) occurs when rain is falling on the soil surface at a constant rate, which is
smaller than the saturated hydraulic conductivity.In fact,this is the same problem as a
constant evaporation of water from the soil surface.
Type (b) is the most common one;at the beginning of rainfall the rate of infiltration
is rather high,but with continuing rainfall the rate of infiltration will decrease gradually
and become less than the rainfall intensity.When this occurs, temporary ponding (sur-
face detention) will occur on the soil surface.
In type (c) infiltration,it is assumed that surface detention occurs during the entire
period of infiltration. This happens,for instance,during field work with infiltrometers
(see sections 4.5 and 6.1.3.5).
6.1.4.3.2 INFILTRATION A T A C O N S T A N T RATE, CAPILLARY R1SE
If water is flowing at a constant rate down a sufficiently long column of soil to a water
table maintained at a constant level, and sufficient time has elapsed for the rate of flow
293
Representaiive and experimental basins
to be the same everywhere in the profile so that no further changes of the moisture con-
tent + take place, the differential equation for the flow is given by:
Oar
K(:+I)=O
which follows from equation (27) with O@t = O.

Integration of equation (31) gives:
KObeing a constant of integration.N o w equation (32) simply gives Darcy's law, KO

being equal to the rate of flow. Rearranging equation (32) gives:
where the suction y is taken as a positive value and z is positive upwards. Integration
of the latter equation between the boundaries z = O, y = O (water table) and z = z,
y = y then gives:
for the relation between z and y. Evaluation of this equation is possible only when
K(y),i.e., when K as a function of y, is known.
For capillary rise of water from a water table held at a constant level the same
solution holds,but then a constant rate of evaporation -EOmust be substituted instead
of an infiltration rate KO.Hence for capillary rise we obtain:
Values of the integral of equation (35) are given in the literature [37]by assuming for
R ($4:
Another assumption for K (y) which is often used is:
a
K=-. (37)
Wn
The value a is determined mainly by the saturated hydraulic conductivity of the soil.
The value of n depends on the soil type and will be about 1.5 for heavy clays and up to
3.0 for coarse sand.Figure 6.18gives the relation between z and y for various values
of EOand KOfor the assumption that a = 200 and n = 2 and is compiled with the aid
of equations (34) and (35).
294
I Analysis techniques and interpretafion of research resuh
From the curves of Figure 6.18 the moisture profiles for each value of EOcan be re-
produced with the aid of the soil-moisturecharacteristiccurve.The line for EO= KO= O
represents the equilibrium moisture profile at which the height above the water table is
equal to the prevailing suction. For increasing infiltration rates, the suction decreases
gradually,causing a certain temporary storage of water in the unsaturated zone above
the water table. O n the other hand,increasing evaporation EOcauses an increasing suc-
tion, giving rise to a water shortage.
KO:10 K0:5 Kor2 K0.1 E,:O
L o g suction = pF
FZG.6.18. The relation between height above water table (2) and suction (y) for various
values of KOand E.
Equation (32) can be used to determine the capillary conductivity by measuring suc-
tion gradient and flow rate,under either laboratory or field conditions.Curves such as
I those given in Figure 6.18 may serve as an estimate of the prevailing water supply from
l ground-waterto the root depth and depth of water table. Generally a suction of 1,ooO
c m (pF = 3) is assumed at the lower end of the rooting zone.The exact magnitude of
this suction y is not so important,since the y curves above a suction of 1,000c m show
a very large gradient owing to the small value of K.If now the intersectionof the various
curves for EOand the line y = 1,000(pF = 3) are plotted against each other,Figure 6.19
'
~
is obtained. From this figure it is seen that a flow rate of 5 mm/day is possibleover a
distance of 29 cm. If, however, the water level is 50 c m below the lower end of the
~
rooting zone, the flow rate is reduced to only 1.75 mm/day.

For the above type of solution,other equations €or K (y) are equations (38) and
(39) [94]:
K = ae-"V'. (38)
I 295
For soils showing an air-suctionvalue, i.e., a considerable range of suction near the sat-
urated end of the moisture-characteristic curve, with constant moisture content,
where y a represents the air-entry value. This is the minimum suction to be applied be-
fore a saturated soil releases its first amount of water.
107
9-
8-
7-
6-
5-
I
h
>.
O
4- .\ \
l -o
l \ 3-
l E
l v
E
l a>
v) 2-
l .-
l L
l ?
l
-
2
.-
Q
1-
FIG.6.19. The relation
between maximum
Co) o
I I l I l l I capillary rise and depth
10 20 30 40 50 60 70 of water table, for data
6.1.4.3.3 P O N D E D R A I N F A L L INFILTRATION
The term used [lo1J represents the appearance of surface detention, c o m m o n with heavy
rainfalls or imgation.
A method of analysis based on a solution of a difference equation instead of the differ-
ential equation has been developed [30, 1021.This method offers the possibility of com-
puting flow rates for different rainfall intensities. For the solution, a K (y) function or
D (4) function of the soil must be available.
Figure 6.20shows the result of the analyses for various rainfall intensities.Also given
is the flux for the case of ponded infiltration in the same soil. It shows that the flow rate
for rainfall ponding is larger over a greater time range than in the case of regular ponding.
The period of constant flow rates increases with a decreasing intensity of rainfall.
296
Analysis techniques and interpretaiion of research results
\
\
'
, 450
The problem of temporary ponding comes close to the problem of redistribution of

moisture. By redistribution is understood the transmission of water in an unsaturated
soil after infiltration and ponding stop and water has to be redistributed throughout
the soil profile. This problem has been investigated on an experimental basis [137, 1381.
Figure 6.21 shows some of the results and the curves indicate that the greater the volume
of infiltration, the faster the redistribution.
For a numerical solution of this problem, not only the total hysteresis, but also the
various hysteresis-scanningcurves of the moisture characteristic must be known before-
hand. Furthermore, the hysteresis effect in K (y) must be included.
Moisture content (fractional volume)

O 02 04
Fio. 6.21. Moisture

profiles during redistri-
1 5 0 ~Depth ícml bution following in-
filtration into slate dust
il 381
297
A numerical solution by means of a finite-differencemethod applied to equation (30)

is given in the literature [131]. Such types of solution are possible only when a digital
computer is available. Moreover, the results must be computed separately for each soil
type, for which both y (4) and K (y) must be known.
6.1.4.3.4 PONDING INFILTRATION
W h e n water is applied in excess to the top of a soil column initially having a homoge-
neous, constant moisture content, water will penetrate into the soil with a velocity de-
pendent on its physical properties. The entry of water is demonstrated by Figure 6.22
[71].The rate of infiltration is infinite at zero time and gradually decreases to a constant
value as time proceeds.
Time hours
O 50 100 150
I I I I l I I
FIG.6.22. into
infiltration Massfour
types of soil [71].
O , 1O 0
Time
200 300
min.
400 500
This type of infiltration is the most intensively investigated one.

Some workers [14,201 distinguish a number of zones in the soil. The zones are given
schematically in Figure 6.23. From top to bottom this figure shows:
1. The saturated zone, reaching to a depth of about 1.5 cm.
2. The transition zone, extending to a depth of about 5 c m in which a relatively rapid
decrease of moisture appears.
3. The transmission zone, increasing in length when infiltration proceeds, with a nearly
constant moisture content.
298
4. The wetting zone, a region of fairly rapid change of 4.

~
I S. The wetting front, a region of very steep mosture gradient which represents the vis-
1 ible limit of moisture penetration into the soil.
moisture content
I soturoted zone
I
II transition zone
I
l
l
I
I
l
l
1
I
I
I
l
XI
d
01
51
el
I
I
I
I
I
l
I l
L
c
c
a
r
I
l
I
wetting zone
/ I
l
I
1 I
wet front
FIG.6.23. Different
zones in the soil during
ponded infiltration [14].
A relatively simple analysis [123]can show the dependency of rate and volume of in-
filtration on the moisture suction of the soil. This analysis is based on the constant
moisture content in the transmission zone. Ignoring the ponding of water over the sur-
face and assuming that the flow is caused by capillary forces and gravity only, the ad-
vance of the wetting front can be represented by:
where Kt is the hydraulic conductivity of the transmission zone and p is the effective
porosity of the soil. This equation shows that the wetting front advances more quickly
in dry soils which have large suction.
Using the initial condition that the wetting front at t = O is located at z=O, inte-
gration of equation (40) gives:
Without introducing large errors, the moisture content in the whole profile may be con-
sidered to be constant and equal to that in the transmission zone. Consequently p is
299
Representarive and experitnental busitis
constant and the mass infiltration in the soil is given by F = p.z. By solvingp from equa-
tion (41)and substitutingp = FIZw e obtain:
F=K,t 1--In-
z ':T.
With an increasing moisture content the suction is decreasing and in wet soils y+O.
In this case equation (42)reduces to:
F = K1.t. (43)
Equation (43) means that Flt = K t and therefore the equation expresses Darcy's law
for unit potential gradient. This should be found after a sufficient time.
For the mass infiltration there is yet another analysis. Rearranging equation (41)we
obtain :
Y P
__
z-7~ In-Z + Y = K,.t= t . (44)
W h e n T is plotted against z on log-log paper for constant values of Y, straight lines are
obtained with a slope equal to 0.55:
log z = log a + 0.55 log t (45)
where a represents the intercept on the z axis. From the above equation it follows that:
which, in general holds for O < z < O S y . Substitution of equation (46) into F =p.z
then yields:
which agrees with the infiltration equation of Kostiakov [53]:
where b is assumed to depend on the hydraulic conductivity of the soil and a = approx-
imately 0.5.The infiltration equation of Kostiakov is valid only for the first stage of
infiltration (zt0.5~). For-+t the value of a must approach unity. Therefore the value
of a cannot be a constant and the application of the equation is rather limited.
For rather large values of t it is advisable to follow a numerical solution of equation
(30) as given in the literature [83]. From the analysis based on this equation the infil-
tration equation :
F = Stil2 + At (49)
was derived [84]where the sorptivity S is a function of the initial and saturated moisture
content and A is a parameter that stems from the above-mentioned numerical solution.
For larger values of t the equation:
F =G + k.t
is proposed [U].
300
Analysis techniques and interpretatioti of research results
6.1.4.4 Saturated-flow analysis

6.1.4.4.1 FLOW A N A L Y S E S F R O M G R O U N D - W A T E R C O N T O U R M A P S
The construction of ground-water contour maps has been discussed in section 5.5.2.
From contour maps, a cross-section such as that shown in Figure 6.24may be construct-
ed. If the height difference between the isohypses is given by Ah (= O.l), the flow per
unit length of contour line is given by:
q = K D,+D,
- A h = KD- A-
.- h
2 Ax Ax
where KO represents the transmissibility of the aquifer. The Ax-value can be measured
from the m a p and is therefore known when KO is set beforehand. O n the other hand,
KD may be computed when q is known. The latter circumstance, however, will not
occur so often since normally KB is known from borings, pumping tests or measure-
ments on undisturbed cores (see section 4.4.2).
O8
i I i
l I I
I I l
I I I
I I 1
I I I
FIG.6.24. Schematic
diagram for computation
of flow intensity from a
contour map.
Another possibility for analysing ground-water observations is given in Figure 6.25.

Between two small parallel rivers the hydraulic head of the ground-water is Ah. N o w
the discharge expressed as a surface layer is:
where q is equal to the effective rainfall (rain minus evaporation) on the area. The annual
run-off is often known and, with known Ah, the value of KB can be estimated at least
roughly.
For semi-confinedaquifers still another type of solution is worthy of mention. Sup-
posing that the confining layer consists of a layer of clay or peat with relatively low
301
hydraulic conductivity,then there will be a difference in piezometric head equal to Ah

between the aquifer and the upper layer given by:
Ah
q=- (53)
c
where c represents the leakage factor or vertical resistance of the confining layer, being
equal to D‘lK‘ where K’is the vertical hydraulic conductivity of this layer.
This analysis is often used to determine the seepage into low-lyingriver valleys [25].
For this purpose contour maps for deep and shallow waters are constructed and from
these a m a p showing differences between then can be made. In addition to this, a m a p
with c-values must be available. By applying equation (53) for each class of Ah and c
and multiplying by the area, total q is found.
-140 -120
FIG.6.26. M a p of piezometric pressures [129].
302
An additional advantage of ground-watercontour maps is the possibility of detecting

geological differences. An example of that phenomenon is given in Figure 6.26 [129].
In an area of some 220 km2,observations of piezometric heights give a contour map in
which buried old river beds,filled up with coarse material,could be traced rather simply.
For a more detailed mapping of these beds, which are of great importancefor the trans-
mission of water, a more detailed net of observations would be necessary.
6.1.4.4.2 F L O W A N A L Y S E S B A S E D ON P I E Z O M E T R I C H E A D S N E A R F R E E
W A T E R SURFACES
When observations of piezometric heads (see section 4.4.2)are made near free water
surfaces such as streams,lakes,etc.,such data can be used to compute the hydrological
constants of an aquifer. For this purpose the known boundary conditions along the free
water surface are used.
The first example pertains to a set of observations as given in Figure 6.27[129].Here
the aquifer is a semi-confinedone and consists of sandy material covered by a heavy
clay layer with a thickness of 2-6 m.Rows of observation wells were installed perpen-
dicularly to a stop-bankedriver. When the logarithms of the readings from given rows
are plotted against the distance from the stop-bank,Figure 6.28 is obtained.According
to a theory of Mazure, a straight line must be obtained from the relation:
where x is the distance and KD and c are the properties mentioned above. By taking
the natural logarithm of both sides of equation (54) one obtains:
or :
log h, = log ho --
0.43~
WDc) '
Hence the slope of the obtained line must be equal to 0.43x/I/(KDc).
C-50- piezometric pressure c m + O D
observoilon well
I FIG.6.27. Ground-watercontours along a river [129].

I
l 303
dirtonce from dyke
FIG.6.28. The relation between distance and piezometric head in two cross-sections of
Figure 6.27 [129].
I The lines need not necessarily pass through the zero point of the distance but will
I
I
generally give an intersect log ho. This is due to the fact that the river does not, as the
l
theory assumes,cut through the whole aquifer.In addition to this there will always be
I some foreland left between the river and the stop-bank.Apparently there is some type
l of radial flow,causing an additional resistance equal to a horizontal flow over some
I 220-250m.The same effect will be found in analysing data obtained from,for instance,
tidal movement.
When the hydrological constant I/(KDc)is known, it is simple to find the flow into
the aquifer. Applying Darcy’s law:
q = -KD-dh (57)
dx
and substituting dhldx from equation (53):
is obtained. Expressed in another way, this is:
With the latter equation,in which qo is the total flow per unit length of river, there is
immediately a distribution of the Row through the aquifer for each x.
304
l
Analysis techiques a d interpuetniion of research results
-
120
100- --\. {p,. Riverlevel
p,.
E 80-
S 60-
2 40-
{20- *
\.
'
.'
\. \-i. II* ' . '
.\a'
.-=-..s'
j i .'
i
z o-
.\*'
e
I I
I l
80 - i &Oc Well 1 at 163m
60-
40-
i
-
.
.
.\.'
. a '
. . s '
I
II
*/
/PL l
.'
\.
\. L.../-
i
/-\.
.\'
.'
/-
i
20-
0-
*i..
Pl
I
I Well 2 at 390m
.. .
K
.
'
--.*
60-
40-
20-
.-'-.-.
\.-.
.'.'
l
I
,
I
ICI-
/p.\,. .'.'
L. /-
.'
..-' .'. /*
*/,+\ '
J -.d.
0- 26-2-'59 I 27-2-I59
1 1 1 1 1 ' 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ~ I ~
8 1
0 12 14 l
6 18 20 22 24 2 4 6 8 10 12 14 16 18 x) 22
The movement in free surface water can be represented by:

40= M + U sin nt (60)
where:
$0 = the water level in the river;
M = the mean value of the water level in the river;
U = the amplitude of the tides;
n = 2n/Tis the frequency of the tides (= 2n radians).
According to the theory of Steggewentz [128] the transmission of the waves in the
aquifer is given by:
$z = M + sin (nt + bx) (61)
where x is the distance from the free surface water and a and p are constants dependent
on the hydrological properties of the aquifer.
For a phreatic aquifer (i.e.,an aquifer consisting of one homogeneous layer), the re-
lation between a and ß on the one side and the hydrological constants on the other is
given by :
where p is the effective porosity or drainable pore spam of the aquifer.
305
For a semi-confinedaquifer as described above the relation becomes:
a2 -ß'
= pcn
2aß
where c represents the leakage factor or vertical resistance of the confining layer.
It has been proved [I281 that the compressibilityof the aquifer must often be taken
into account. In that case the above relations become:
Sn
2aß =-
KD
where S represents the storage coefficient (the capacity of releasing or taking up water
when the compressibility of the aquifer is taken into account).
The analysis of data is now reduced to determining the values of OL and B from obser-
vational data. In order to explain this, consider equations (60) and (61). From these
equations it is evident that the amplitude in the aquifer is e-"" times that in the free
surface water. Hence the ratio of the amplitude U, at a distance x and the amplitude
UOat x = O is given by:
or
and therefore a straight line must be obtained when the logarithm of the amplitude
ratio is plotted against the distance x. Results for three rows of observation wells are
shown in Figure 6.30.Here again the obtained lines do not pass through the origin:
this is due to the fact that the river does not penetrate fully into the aquifer.
From equations (60)and (61) it is also easily seen that the phase shift which is the
lag in time from the maxima and minima of the observations,is equal to Px. Therefore
the observed time lag (expressed in radians) plotted against the distancex gives a straight
line having an angle /3x (Fig. 6.30). The free-surface-waterdata need not necessarily be
involved in the analysis.For the zero movement,the well closest to the free surface wa-
ter may be used. Furthermore,the circumstance a = j3 points to a phreatic aquifer and
equation (62) must then be used to compute the hydrological data.The movement need
not show pure sinusoids at all and somewhat distorted ones will generally be dealt with.
The analyses may also be applied to the yearly fluctuation,but a Fourier analysis must
often be applied to resolve the movement into fundamental sinusoids [I IS].
It may happen that piezometer observations are available along free surface water
not showing the typical train of sinusoids in its level.This often occurs along ephemeral
streams. T w o conditions may occur: (a) a sudden rise or fall of the water level in the
channel,the level being constant during the rest of the time; (b) a slower rise or fall
306
'ioglornpliiude ro tio)
Ohose shift in rod.
disionce in rn
FIG.6.30. Relation between distance and logarithm of the amplitude ratio (left-handscales)
and the relation between distance and phase shift (right-handscales) for some wells given in
Figure 6.27.
of the water level in the channel,the lowering or rise being a linear function of the time.
Solutions for these cases are given in the literature [34].
6.1.4.4.3 F L O W A N A L Y S E S IN D R A I N E D A R E A S
When observations in drained representative or experimental basins are carried out, a

combination of discharge and depth of water table may be very useful in analysing
ground-waterbehaviour. For a complete treatment of flow problems in areas with paral-
lel systems,refer to the literature [66].Here only a few principles will be discussed,and
they may be used for analysing the flow.
In the first place time-independentobservations of the so-calledsteady-stateflow can
be used. In its simplest form the steady-stateflow towards a system of parallel drains
can be described by the well-known Hooghoudt equation:
or :
For symbols,see Figure 6.31.The factor d depends on drain spacing,geometry of drain

and depth of the impervious layer. Further, it should be remarked here that the first
term of the right-handside of equation (67)pertains to flow below the drains,the second
to the flow in the region above the drains.
Ground-waterobservations are normally made midway between the drains.When they
are combined with discharge measurements,a plot of q against m is obtained,but care
must be taken that only data pertaining to a falling water table are used. Owing to the
non-steady state behaviour of the system, observations pertaining to a rising water
table will give another relation.
The general slope of the curve will give information about the permeability of the
soil. Changes to a larger slope above a certain m-valuepoint to a larger hydraulic con-
ductivity, as a smaller slope indicates a smaller permeability.
! 307
I I
I ' k2
H I
I d = fíH.L.ul
I '
I I
I '
I
i
~,/,/,/,,,,,//,,11,~/,,,,,,,,~,,/,,//////,,//,/,,,/////,,,//,/,,/~,/~~
FIG.6.31. Schematic representation of flow towards parallel drainage channels.
In order to facilitate the analysisit is worth-whileto divide both members of equations

(68) by m. This gives:
4
m
-=u+bm.
This equation indicates that by plotting q/rn against rn, a straight line must be obtained.
N o w either a straight line parallel to the m-axis is obtained with an intersect 'u' or a
straight line with an intersect 'u' and a slope 'b'. In the former case the flow in the re-
gion above the drains may be neglected.
The analysis of the non-steadystate ground-watertable midway between two parallel
drains can be carried out in several ways,depending on the initial conditions set for the
problem.
Firstly, at t = O an instantaneous recharge R is applied, causing a rise of the water
table equal to R/p.For this case the height of the water level midway between the drains
is given by:
where z is taken as the ground-water level above the water level in the drains and
a = (nZKD)/(pLz) [31].For large values of t the terms with m > 1 may be neglected,and
equation (70)reduces to :
This equation shows a straight line when z is plotted against t on a semi-logscale.

The second and following terms of equation (70)may be neglected when they are
smaller than 1 per cent of the first term [55] and hence:
0.44
-1 e-Bat(- 1 e-at+e-sat(0.03 and t)-.
3 1O0 oc
Equation (71)is often used to determine the magnitude of the ground-waterflow to-
wards river systems [51].Applying Darcy's law,it is easily found that for the discharge
one obtains the general equation:
qt = qt-1 e-" (73)
308
Analysis techiiiqiies and inrerpretatìon of research results
where qt is the discharge expressed as a function of the discharge qt-i of the preceding
day. Pure base flow occurs as soon as a plot of qt against qt-1 shows a straight line.
Further, m is a constant depending on geological and hydrological properties of the re-
gion (see section 6.1.3.2.2).
Another initial condition to be applied to non-steadystate flow is the case of a steady
infiltration rate f starting at t = O. The flow rate towards the drains in this case is given
in the literature [55,681:
The height of the water table midway between the drains may be represented by:
The analysis based on these equations is similar to that described above. Solutions for
non-constant infiltration can be derived simply by means of superposition. Equation
(74)offers the possibility of dividing flow between surface flow and ground-waterflow.
This is done by computing the ground-waterhydrograph for the system under discus-
sion (with known a). The total ground-waterflow over a certain period is obtained by
integrating the flow rate. This amount is subtracted from the measured total flow,
leaving the surface run-offas the difference between them.
6.1.5 Erosion and sedimentation

6.1.5.1 Analysis of study data
When data are gathered from measurements and field observations of a variety of hy-
drological and geomorphological variables in a representative or experimental basin,
it is only the first step to an understanding of the basin as a unit. In past years,the sedi-
ment yield as measured in a reservoir or at some other gauging point has been related to
a variety of basin variables by multiple regression analyses which may delineate the
effect of the characteristicsof basins on sediment yield [6],but there are often too many
interdependent and uncontrollablevariables at work in a natural basin to permit a sound
determination of many sedimentation phenomena [86].Moreover, extrapolation of sed-
iment-yieldrates in reservoirsto a largerarea downstream beyond the limitsof any record
is hazardous because of the land-formchanges which afford added opportunity for sed-
iment deposition [40].
In selectingresearch basins care must be taken that all variables important for accom-
plishing the stated objectives of the study can be measured accurately [86].If conditions
are carefully controlled and basic processes are understood, the relation of sediment
yields to basin characteristics will be more meaningful. Some of the relationships that
have been significantin basin studiesare described.For measurement and data processing,
see sections 4.9and 5.6.
6.1.5.2 Relation of basin area to sediment yield
The size of a basin has an important effect on the total sediment yield [38,401.Figure
6.32shows the relation of mean annual sediment yield to basin area for 99 small reser-
vojrs by size classes.The graph shows that the average unit sediment yield from basins
0.16km2 in area is about six times that of basins i .6km2in area. Similarly [40],records
309
FIG.6.32. Relation of
mean annual sediment
yield to drainage area
I 1 I I
for 99 basins in eastern 0.07 0.1 0.2 0.5 1.0 2.0 I
Wyoming 1401. BASIN AREA ( S Q U A R E KILOMETRES)
rom 1,100measurements in the United States show that basins less than 10 km2in area
have unit sedimentyieldsabout seven times those of unit sedimentyieldsfrom catchments
of 1,000km2.The significance of this relationship is that the larger basins have more
diversified topography and environments for redeposition of eroded material from head-
water areas.
6.1.5.3 Relation of relief and length to sediment yield

A basin characteristic that has been found to be correlated with sediment yield is basin
slope. This has been expressed as a dimensionless parameter called relief ratio [104].
It is defined by the equation R = hill where R is the relief ratio, h is the basin relief
between the maximum and minimum elevations in metres, and 1 is the maximum basin
length in metres. Good correlations have been found between sediment yield and relief
for a group of small basins underlain by a variety of rock types in the Colorado Plateau
Province of the Western United States (Fig.6.33)[1041.Similarly,good correlations be-
tween relief ratioand sedimentyield for several basins in Eastern Wyoming are available
[38].It should be pointed out, however,that some exceptions do occur.In basins with
two distinct types of topography,the reliefratio is not a satisfactorymeasure of geomor-
phological character or sediment yield.
6.1.5.4 Relation of drainage density to sediment yield

Drainage density an is index of topographical texture and can be expressed as the total
length of channels (in km) within a basin divided by the basin area (in kma). Several
3 10
Analysis techniques arid interpretation of research results
FRIABLE SANDSTONE
CONGLOMERATE R SANDSTONE
FIG.6.33. Regression
of sediment loss versus
relief ratio for basins in
the Colorado Plateau 0.1 I I I I
0.2 0.4 0.6
[104]. RELIEF RATIO
investigators have found that there is a good correlation between drainage density and
sediment yield on small catchments.
6.1.5.5 Other basin characteristics affecting sediment yield

There are several variables that have been measured and correlated with sediment yield
with varying degrees of success [64].They are mentioned here because they may prove to
be important in some regions where representativeor experimental basins will be estab-
lished.These are summarized [6] in three groups: flood hydrographs or storms causing
these which are producing and transporting sediment;the basin conditions and land
use variables which are changeable or changing with time;and the basin characteristics
such as soils and rock type which are essentially constant throughout time. The sedi-
ment-measuringdevice should also be considered. If, for example, a reservoir is used
to measure sediment,its trap efficiency or the part of the total sediment it catches is
important.
311
6.1.6 Water qualily

Variation in water quality in representativeand experimental basins may be related to
streamflow and source of supply.Each water-quality observation should be correlated
with the daily mean discharge and the latter compared with the annual average discharge
(see section 5.4.5). A useful additional analysis is to separate the flow into surface and
subsurface-flowcomponents (see section 6.1.3.3.1)and to relate these components to
changes in water quality.
The ultimate aim is to predict long-termcharacteristicsfor a given stream and these may
include the following: (a) average values of the principal ions in water and physical
characteristics for various stages; (b) variations of such values and the correlation of
such variations with stream-flowcharacteristics;(c) relations between water quality of a
stream and source of supply.
On experimental basins additional analyses are possible if observations are made on
micro-channels.The following analyses are recommended in particular : (a) variations
of water quality in micro-channels and the correlation of such variations with flow
characteristics;(b) influence of cultural changes on the water quality; (c) comparison
of water-qualitydata with those observed on representativebasins and analyses to ex-
plain any differences.As a final step water-quality maps should be made. Such maps
can be made for various seasons showing quality characteristics observed on basins.
Given a sufficiently dense network of observational stations,isolines of water quality
can be drawn. Such maps are useful for water-qualityprediction on channels where no
observations are made. Hydrochemical maps in particular may be used to estimate
the hydrochemical regimen of planned artificial reservoirs by means of a water-salt
balance [i, 92, 1251.
6.2 Relation between elements of the hydrological

cycle and between these and basin characteristics
in approximately stationary condition
6.2.1 General
Within the purposes of representativeand experimental basins (see section 1.3) a num-
ber of problems arise, such as those of flood prediction,prediction of the effect of a
natural or cultural change on the hydrological regimen, frequency of occurrence of
drought, etc.
The great physical complexityof a basin (even if there is no more fundamentalreason)
makes it virtually impossible to solve problems by the application of basic physical laws.
In the solution of practical problems,therefore,some idealization of the physical system
and of the input to it must be made.
The solution to these problems is attempted, in general, by postulation of model
systems, which bear a restricted equivalence to natural prototypes and of which the
operation in transforming the input data can be described numerically or analytically
for the approximateprediction of the behaviour of the prototypes. These relationships,
because of the extreme complexity of any basin, cannot be established simply on the
basis of what is known at present regarding the interactions of its physical components
and the interveningmechanisms. In seeking these relationships on the basis of fielddata,
three major problems are encountered:
1. The time variability of basins due to natural changes (weathering,erosion, climatic
change,etc.,which determine the geomorphological evolution of basins) and in some
cases due also to cultural changes.
2. The uncertainty with respect to space and time distribution of the inputs and
312
Analyssis techizigires und itzterpretution of research results
outputs of hydrological systems,and with respect to the states and properties of their
interior elements in time.
3. The inherent non-linearity of the processes of mass and energy transfer which con-
stitute the hydrological cycle.
These problems have a strong bearing on the accuracy of any quantitative description
of the hydrological behaviour of natural basins.
6.2.2 T i m e Variability of basins

The time variability of basins can be disregarded at least during some periods of obser-
vation. The hydrologist then relies on the analysis of sets of data on selected inputs and
outputs measured during these periods, when the basins are thought to operate in a
stable manner. Conclusions regarding natural or cultural changes are then based on
comparisonswith the relationships developed in this calibration.
6.2.3 Uncertainty as regards system inputs a n d outputs

It is clear that neither the system input (as represented,for example, by the intensity
of natural precipitation over a basin at every point in time and space) nor the states of
storage of water in every basin element,nor the motion of each fluid particle, can be
measured or even estimated.Hence, any representation of the operations of a hydrolog-
ical system must be either an empirical relationship between inputs and outputs derived
from field data,or a relatively coarseidealizationof thefunctionsof the prototype,based
on these data.In hydrology,field measurements normally yield only indices or estimates
of the prototype inputs and outputs. Stream gauging for instance, frequently lacks
great accuracy and ground-waterflows are seldom,if ever,determined directly. In gen-
eral, the relation between a real input and any adopted index is variable: hence, the
usefulness of the index depends on this degree of variability. Accordingly,a good index
would be one which,when introduced into a fixed expression,gives a consistent meas-
ure of the true input within prescribed limits of accuracy. Unfortunately, objective
methods for the assessment of index reliability have not been developed in representa-
tive and experimental basins ; however, field measurements are sometimes sufficiently
accurate to consider the indices obtained to give such a consistent measure within an
acceptable limit of accuracy.
6.2.4 Non-linearity of hydrological processes

Many hydrologicalprocesses are non-linear.It is known,for example,that the movement
of flood waves in rivers is governed by non-lineardifferential equations,and therefore
the linear assumption of the unit-hydrographconcept (see section 6.2.5.3) cannot be
absolutely correct.This disregard of physical laws may be unwise,but it does not neces-
sarily imply failure of the method to achieve its end.in representativeand experimental
basin research, however, accurate data is generally produced and utilized for wider
prediction purposes and the closest possible equivalence of the model to the prototype
is desirable.This implies a realization of the non-linearityof hydrological processes [3].
6.2.5 Types of model

With few exceptions, the current procedures used in hydrological applications involve
some degree of lumping of the input and the output parameters. This is true, whether
a basin is considered in toto as a ‘blackbox’,whether a model with discrete component
313
subsystems is used, or whether the basin is subdivided into partial areas and successive
routing operations are performed in order to compute the output. Consequently,the
problem consists of devising a model such that, when it operates on a lumped index,it
produces an output that agrees reasonably with the measurable output of a prototype
system which,in general operates non-linearlyon a distributed input. This concept is
illustrated in Figure 6.34.
With reference to the figure,a model W must be chosen such that the result of its
operation on the lumped index xi(t) is an output yl(t) which agrees as closely as possible
with the output y(t) of the prototype, which operates on a distributed input x(t,s)-a
function of time and space.
Distributed
basin input Y (f)
x (i, 5)
+ Minimized
4 H
x,(t) YI (tl
FIG.6.34. Relationship between a prototype hydrological system and a lumped parameter

model.
Various types of model have been proposed to perform this task. They fall generally
within the following categories:
1. Linear normal models (multivariate, correlation and regression analysis, factor and
discriminant function analysis).
2. Graphical analyses.
3. Models with a central linear element (unit hydrograph).
4. Non-linear functional models (general non-linear analysis).
5. Conceptual models (general synthesis).
It must be realized that the development of any relationship,however simple, may be
classified as a model. In normal hydrological problems, relatively simple models are
frequently satisfactory,but for research on representativeand experimental basins the
use of more complex model techniques is warranted. For practical application,models
should be verified with the results of field experiments.
Examples given in the description of the models have been restricted to rainfall-run-
off relationships, but such models are equally applicable to any other relations.Note
that some of these are to be classified as component models only. Some important ones
are:
Run-offfrom snowmelt with climatological factors;
Base flow with climatological characteristics;
Water level of the aquifer with climatologicalcharacteristics;
Natural recharge of the aquifer with climatological characteristics;
Evaporation with climatological characteristics;
Hillslope erosion with climatologicalcharacteristics;
314
Sediment transportation with run-offand other Characteristics;

Water-qualitycharacteristics with climatological and hydrological ones;
Relation between the elements of hydrological cycle and basin characteristics.
The various types of model are discussed briefly below.
6.2.5.1 Linear normal models

In this procedure various combinations of variables are tested to explore the significance
of their effect on the hydrological system. The combination that yields a relationship
most closely approximating the recorded output function in terms of the recorded input
function and other arbitrary parameters is adopted as the best prediction equation.
Precipitation and flow characteristics are frequently the primary input and output
parameters. The input is operated on by a system that has the form of a linear normal
model involving additional arbitrary basin parameters. These parameters are chosen by
judgement for a number of models on the basis of a general knowledge of the hydrology
of the basin. Each individual model yields both a series of regression coefficients,as de-
termined by methods of regression analysis,and a set of error estimates and measures
of correlation as determined by analysis of variance and covariance. By examining
the error estimates of all the models and the measures of correlation,the best model is
selected.
A typical model involving only precipitation and run-off volumes during discrete
periods of constant duration and other variables such as evaporation, might be thus:
n
yc = ao + ai xt + a2 xt-i +a3 xi-a + ... + 2 bjzj + e (76)
i-1
where a’s and b’s are regression coefficients,xt, xt+~,etc., are values of the precipitation
during periods concurrent and antecedent with the run-offyz, the z’s are other variables
(such as evaporation) during concurrent and antecedent periods, and e is an error term.
Ithas been found,in practice,that in prediction equations like equation (76),the resid-
uals are not normally distributed,owing to non-randomeffects. This makes these linear
normal models imperfect in the general case.
A great deal of subjectivityunderlies the use of these models. Not only is it possible to
obtain almostequally good prediction equations on the basis of sets of different parame-
ters, but judgement must be exercised to avoid using physically irrelevant parameters.
The elimjnation of parameters which possess strong interdependence can be attempted
by some of the more recent methods of factor analysis. These procedures, together
with theuse of discriminantfunctions to test for differences,permit insight into the prob-
ability of successful utilization of these equations for prediction purposes [1261.
Other characteristicsof the procedure include: lack of assurance that among the mod-
els investigated the optimum has really been included; the impossibility of generaliz-
ing the prediction equation to other, similar systems; and inverse correlations with
some of the variables when such relationships are illogical from physical considerations.
O n the other hand, the methods of correlation analyses are powerful tools when ap-
plied to the test of well-definedhypotheses in many fields of physical hydrology. They
have also legitimate applicationswhen the purpose of the analysis is to evaluate physical
parameters in the experimental study of systems that are well defined by functional re-
lationshipsobtained independently [731.
6.2.5.2 Graphical analysis

The structure of the linear normal models does not lend itself to the representation of
short hydrological episodes such as,for instance,individual storms and floods,in which
the complexity of the system’s response relationships is more evident. These models
315
generally break down under such circumstances as the often-observed case where a giv-
en storm produces virtually 100 per cent run-offin the middle of the winter and under
summer conditions a very similar storm produces virtually no run-off.Systematic pro-
cedures for taking this variation into account have been reported in the literature [65],
Notable successeswere obtained in relating,by graphical correlations,the observed storm
run-offvalues attributable to individual rainfall events, to the amounts and duration
of rainfall and to the Conditions in the basin at the times of occurrence of the storms.
In application to a given basin, the method may require,for instance,a record of rain-
fall and discharge (see section 6.2.5).
O n the discharge hydrograph,the base-flow separation is made (see section 6.1.3.3)
and the remainder of the hydrograph (i.e.,the direct surface run-off) is divided up be-
tween individualrainfall events.The volumes so determined are computed and graphical
relationships are established between the rainfall amounts and durations, the times of
year, the base-flow discharges at the time of occurrence of the storms and an index of
antecedent rainfall.It should be emphasized that the storms used in such analyses should
not be limited to those producing large amounts of storm run-off;if the selection is so
limited,insufficient notice will be taken of storms containing substantial rainfall but
producing little or no storm run-offbecause of antecedent conditions [so].
Once the relations have been established for a given basin, they are used to predict
the run-offto be expected from any given occurrence of rainfall in any given condition,
and hydrograph prediction becomes possible. See, for example,Figure 6.35 [16,171.
Il
O
20 LO 60 no 100 120
Moss ruinloll for euch storm mm
FIG.6.35. Simple
example of multiple
correlations for rainfall- (b)
run-off relations.
Representative basin of
Mayo Ligan (Chad):
(a) correlation between
ratio of surface ñow-
-
k
LO
20
O
rainfall and mass rainfall
for humidity index
Ih = 26; (b) correlation
between AKu and - 20 J I I I l 1 I
humidity index Ih
(Ih = 1:Pi [Piprevious I I l l
tl -LO
storm; regression for
correction]). Ih
316
As well as its dependence on base-flow separation (which is an arbitrary procedure)

this method, successful though it has been on a number of occasions,is subject to cer-
tain other weaknesses.The graphical correlationis sometimes difficult to apply and the
number of degrees of freedom used up is difficult to determine. This means that the
accuracy of the prediction cannot be readily assessed. Perhaps a greater criticism is
that graphical correlations,unlike linear regression analysis,for example,fail to provide
an objective measure of the statistical significance of each independent variable. This
can be overcome. at least to some extent,by completing the analysis without a suspect
variable and considering whether the result obtained ‘with’is significantly better than
that obtained ‘without’.However,the subjective nature of graphical analysis may not
render successive computations by this procedure entirely comparable [9].
6.2.5.3 Models with a central linear element (unit hydrograph)

6.2.5.3.1 GENERAL
As background for the discussion of this method and other methods to follow, it is
necessary to define the meanings of ‘systemanalysis’and ‘system synthesis’.Jn system
analysis, the relationship between input and output is established by a mathematical
process involving the use of measured input and output data only, without any attempt
to describe the internal mechanisms of the system in explicit form. This relationship
has the form of unique functions,which are made to operate on the input in order to
produce the output.In general these functions are not required to have a physical mean-
ing or to pcssess parameters fulfilling conditions of dimensional consistency.
In system synthesis,on the other hand,the investigator attempts to describe the oper-
ation of the system by a linkage or combinationof components,the presence of which
is presumed to exist in the system and of which functions are known and predictable.
The linkage of components must be made in such a manner that the correct output is
produced whenever a specific input is applied. In general,the process of synthesis does
not yield a unique model of the unknown system.
Pure synthesis or analysis can be performed on a system independently,or a combi-
nation of both can be employed. A method of partial synthesis with linear analysis is
the basis of the classical unit-hydrographprocedure. Since its formal presentation by
Sherman [lo71this procedure has had very extensiveuse throughout the world.Its present
mathematical basis was developed during the last decade [28,781.
6.2.5.3.2 C O N C E P T O F UNIT H Y D R O G R A P H
The unit-hydrographconcept is a deliberately simplified hypothesis of part of the behav-

iour of a basin in transforming rainfall to discharge. This hypothesis is that that portion
of the total recorded rainfall (expressed as a function of time) which becomes ‘effective’
(i.e.,contributes to direct or surface run-off)is as the input to the output of a time in-
variant (stationary) linear system.
In the application of the above concept for transforming a record of rainfall into a
hydrograph,three main operations are involved : (a) determination of the effective rain-
fall (or rainfall excess); (b) determination of the direct or surface run-off;(c) linkage
between effective rainfall and direct or surface run-offby a linear system operation.
In essence,it is assumed that the gross precipitation input is first modified by infil-
tration over the surface of the basin to yield a rainfall-excessfunction.The basin then
operates on the rainfall excess in a manner equivalent to a linear system to produce the
direct or surface run-offwhich,modified in turn by the baseflow (or baseflow plus inter-
flow) becomes the recorded gross river discharge.
This sequence represents a partial synthesis,yielding the structure of a three-element
cascade as illustrated in Figure 6.36.
317
The two end operations are performed first:

1. The rainfall excess is found by empirical procedures consisting of subtracting from
the gross precipitation values an infiltration function.Several indices have been pro-
posed for this purpose,including the +-index,the antecedent precipitation index,etc.
[136].The transformation from gross rainfall to rainfall excess is frequently the most
difficult of the three unit-hydrographoperations and it is important to consider this
in any analysis.
2. The direct or surface run-offis found by flow separation (see section 6.1.3.3.1).
The volume of rainfall excess must be equal to the volume of direct or surface run-off.
Some adjustmentsmust usually be made in the results of 1 and 2 above in order to ob-
tain this equality.
The two functions found by the above procedure are the input and the output of the
central element of the cascade. The function linking them is the unit hydrograph,which
must now be determined.This is done by one of a number of procedures to be discussed.
6.2.5.3.3 DETERMINATION O F UNIT H Y D R O G R A P H
With a known unit hydrograph,the process of reconstruction of past run-offrecords,

or the prediction of floods,is carried out with the three-elementcascade operation per-
formed in the sequence shown in Figure 6.36 startingwith the record of rainfall.
As remarked, the unit hydrograph is equivalent to a stationary,or time-invariant
linear system. This type of system satisfies the principle of superposition,namely the
output of the system due to an input which consists of two or more parts is equal to
the sum of the outputs due to each of these parts acting independently;and the response
to a given input is independent of the time at which it occurs.
If the input and the output of a linear system involvingstorage,or 'memory'elements,
are expressed as functions of time x(t) and y(t) respectively,the relation of superposition
may be represented concisely by the convolution integral or functional:
t
y(t) = j h(t)x(t-t)dT (77)
-W
or, in alternative form:
t
y(t) = j h(t-.t)x(.t)dt. (78)
-m
h(t) is a function of time, known as the kernel of the integral or the impulse response,
or system response function of the system it represents.It is the output which would
be observed in the limit if the system, while remaining linear, received a single input
of very high intensity and a duration approaching zero,such that the product of inten-
sity and duration remained unity. The lower limit of integration is -00, to take into
account the fact that systems of this type have memory. which means that,in general,
the inputs of the past, even if they are remote,may affect present and future outputs.
In hydrological literature the unit hydrograph u(t,T)is defined as a function of time
(t) giving the storm run-offdue to unit volume of effective rainfall generated uniformly
(or in some specified, perhaps typical, non-uniform manner) in space and time on a
basin in a period T.Thus we may speak of six-hour,three-hour,etc.,unit hydrographs,
the stated period indicating the duration of effective rainfall.In particular we may speak
FIG.6.36. The unit hydrograph cascade.
318
of an instantaneous unit hydrograph u(t,O) which is the limit of the finite period unit
hydrograph as the duration of effective rainfall is reduced indefinitely.
Clearly the instantaneous unit hydrograph (IUH) is analogous with the impulse re-
sponse of the general linear system and the general convolution integral has its counter-
part in the definitive equation [54,63, 781.
If adequate data on input and output are available,the function h(t) of equation (77)
can be found by any one of a variety of techniques which have been used in hydrology,
such as those of harmonic analysis [81], Laguerre analysis [29]or other methods [27,
32, 701.The use of a matric solution involving least squares fitting has been developed
in the United States [114].
With the above procedures available, the difficulty remains not in the inversion of
the convolutionintegral but in defining the input and output data. When measurements
are available, an appropriate course of action to reduce error would be to find,in the
records of rainfall and discharge,single,isolated floods caused by storms of short du-
ration and to derive a set of unit hydrographsby reducing the ordinates proportionately
to the volume of direct or surface run-off.If T,the duration of effective rainfall,is small
relative to the period in which the flood rises, the variation of effective rainfall within
the period may be neglected,and the curve so obtained may be taken as the unit hydrv-
graph of duration T.This unit hydrograph may be quite close to the instantaneous unit
hydrograph.
The advantage of the short,intense storm lies in the fact that the shorter the duration
of rainfall the less the effect on the unit hydrograph of uncertainty in defining the losses
and therefore the effective rainfall. Similarly baseflow separation is somewhat less ar-
bitrary when the storm is isolated in time. It is important that the unit hydrographs
should be obtained,if at all possible, from fairly large floods;otherwise they might not
be representativeof the true conditions during high-flowperiods. If,however, the unit
hydrograph concept is used to study the relation between the full discharge hydrograph
and a rainfall record from which losses have been deleted in accordance with some
criterion,the more complex methods of inversion mentioned above would seem to be
advantageous [50,1031.
If significant inconsistencies in unit hydrographs derived from adequate records of a
single basin are observed, an attempt could be made to show the dependence of the
derived unit hydrographs on parameters of the storm and conditions in the basin at
the time of the occurrence of the storm. Such dependence might indicate the necessity
for a more sophisticated non-linearanalysis. It should be remarked,however,that any
conclusions reached with respect to the applicability of the unit hydrograph should be
restricted strictly to the relationships between indices of the prototype input and output
and not to the real input and outputs. The prototypes,in general,unquestionably repre-
sent fully non-linear systems.
It must be emphasized again that the unit-hydrograph hypothesis is nothing more
than a convenient assumption.Indeed,the cascade structure involved in it is quite arbi-
trary. Its merit lies in its simplicity and in the fact that it is the most general assumption
which can be made without recourse to non-linear methods.
6.2.5.3.4 USE O F UNIT H Y D R O G R A P H S
In representativeand experimental basin research,unit hydrographs can be used as an

index of some of the flood-producingcharacteristics of a basin or as a component in a
conceptual model (see section 6.2.5.5).
The linear convolutionoperation has been employed as a linkage between hydrological
elements that do not involve surface flow. In the Netherlands,where the occurrence
of surface flow is rare,this method has been applied to the study of the relation between
percolating rainfall and the discharge of parallel field drains and ditches [SI. Through
319
the use of the linearized Dupuit-Forchheimermodel for two-dimensionalground-water

flow it was found that this relationship is governed by one parameter which is called
the reservoir coefficient.This compound parameter incorporates all hydrological charac-
teristics of the drainage situation studied, but probably cannot be generalized to the
transformation of percolating rainfall into the ground-water outflow from a natural
basin.
Studies of basin characteristicsshould be directed to the elucidation of some of these
relationships with a view to determining the number and combination of characteristics
which are more fundamental or to which the hydrological response might be more
clearly related.A n example of such a study,using factor analysis,is given in the literature
[116].
Finally, as the natural systems involve non-linear operations, the trial of a direct
linear convolution relationship between recorded index inflows and outflows in a basin
might prove to be fruitful in certain cases.This would be equal to eliminating the two
end elements of the usual unit-hydrographcascade,on the assumption that the operation
of a particular basin is approximately equivalent to a linear operation between the input
and the output indices. A possible application of this concept is further discussed in
section 6.3.3.
6.2.5.4 Non-linear functional models (generai non-linear analysis)

Because the operation of the hydrological cycle is manifestly non-linear,even in the com-
paratively simple matter of the movement of floods in rivers,research in representative
and experimental basins should include studies in general non-linearanalysis.
Such relatively complicated techniques may well be too complex for use with the
kind of data to which the unit-hydrographconcept has been applied in the past, but
one important feature of representative and experimental basins is the production of
accurate,detailed data,an essential requirement for non-linearanalysis.
The general case of a non-linear,time-invariantsystem operating on an input x can
be expressed by the summation of a first term identical to equation (77) and a series
of similar terms of progressively higher order, forming a functional series thus [133]:
t t t
y(i)= J hi(ti)x(t-ti)dti-~- j J hz(rl,t2)X(f-tl)X(I-52)dtldZz + . ..
-m -m -00
m
= 2 J.. . J
n=l
where the symbol

t
-00 -M
n
t
hn(ti,. .., t n )
n
II
i= 1
.
~(t-tj)dti.. d t n
1 1 1 represents the product of n factors of the form x(t -

i=
tj).
i (79)
It can be seen that whereas in the representation of a linear system one employs only
the first term of the series,which is the ordinary convolution integral,the general non-
linear case involves other terms which are multidimensional generalizations of the con-
volution integral with generalized impulse responses hn(cl, . . .,tn). This resembles the
expansion of non-linearfunctions in a Taylor series.
The functional series representation of a system is illustrated in Figure 6.37 in which
each box corresponds to a subsystem of progressively higher order. Each one of these
subsystems,denoted by the symbols Ei, . ..E n ...is described by one of the elements of
the sum of equation (79).
Because hydrological systems are non-linear,recent work has been devoted to the
development of procedures for establishing the relationship between input and output
indices by a system with this structure.The problem,in its general form,is to determine
each system response function so that the sum:
Eixi + Ezxi + . . . + Enxi = yi(t) (80)
320
Analvsis techniques and interpretation of research results
matches the output of the prototype with reasonable approximation.Note that the input
component subsystem Ej is an index of the real prototype input.
X I of each
FIG.6.37. Functional
series representationof a
non-linearsystem.
This is analogous to the inversion problem of the unit-hydrographanalysis. General

methods of direct non-linearinversion are not yet available,however,and current work
involves the use of a series of approximations [2,5].
It appears at the present time that for the development of adequately precise input-
output relationships it is necessary in non-linearanalysis to have available a fairly long
set of concurrent records of inflow and outflow containing sufficient information to de-
termine properly the structure and the parameters of the response functions which de-
fine the system. The demands made in this method by length and information require-
ments of the records are obvious from the fact that,whereas in some synthetic models
the informationsupplied is frequentlynot only a main precipitation input but also several
disturbance functions,in the analytic model the lack of explicit data on these modulating
functions must be made up by longer statistical time series representing appropriate
samples of periods under a variety of conditions.
In contrast with some of the other methods discussed,general non-linear analysis is
essentially independentof any qualitativeor quantitativejudgement based on incomplete
knowledge of the physical hydrology of the basins under study. In this sense it is free
of subjective bias. O n the other hand, this method does not yield any information (not
even qualitative) on the role played by any of the generally recognized components of
the hydrological cycle.However,in those cases where abrupt changes in the system have
occurred at a known point of record it is possible to detect corresponding changes in
the system functions.The procedure can be useful therefore in evaluating the time va-
riability of basins.Another potential application of non-linearanalysis is the establish-
ment of functionallinkages between component inputs and outputs in conceptualmodels
6.2.5.5 Conceptual models (generai synthesis)

The process of general synthesis ordinarily begins with the postulation of a model of
which the structure is based on qualitative and semi-quantitative knowledge of the
321
phenomena involved in the hydrological cycle.This model contains elements defined by

explicit functions,which describe operations effected on variousportions of the input and
the storage.A model of this type is shown in Figure 6.38to illustrate one way in which
the process of synthesis can be accomplished. The recorded input is processed through
the model, and the resulting output is compared with the recorded output of the natural
system. If an acceptable agreement is not found, one or more of the functions of the
component subsystems are modified and adjusted,and the process is repeated in a sys-
tematic way until there is adequate correspondence between the synthetic and the re-
corded outputs.
For the purpose of predicting the hydrograph of a particular basin,the problem may
be considered solved if, on optimization of the model parameters, the computed and
observed outputs agree to within an acceptable tolerance.if, however,the model is seen
as a tool to extend understandingof this phase of the hydrologicalcycle (and particularly
if it is hoped to relate the values of the parameters of the model to the basin character-
Y Evaporation
Precipitation
Run-off from
impervious areas
-7
Interception and
depression storage
- Direct run-off -- Surface run-off
L o w e r zone
Interflow
_f_ storage
I 1r 11 i
-
4
Ground-water Translation and Translation and
storage routing storage routing storage
11 v
V L
t
S u b s urfoce
Hourly or daily
ground-water
stream flow
I J I I
FIG.6.38. A conceptual model of a hydrological system [21].
322
istics by application of the model to many basins-either so that the model may be
used for prediction on basins without suitable records,or that it might be used for pre-
diction of the hydrological effect of changes proposed within a basin), then considerably
more than a mere ability to reproduce the observed output is required. These extra
requirements include a quantitative evaluation of the significance of the parametric
values determined for the model.
It is important to note with caution that far-reachingconclusions regarding the pro-
totype on the basis of analyses of the characteristics and the interaction of model com-
ponents are not always warranted and are often very risky.It must be recalled that the
modelling of a complex natural system involves simply recognizing the main functions
of the prototype and devising mathematical or analogue elements which, if properly
linked,can perform approximately these functions.The prototype and the model have,
as mentioned in section 6.2.1,only limited equivalence. This should always be taken
into account in the application of models to ungauged basins.
The significance of the model parameters is analogous to the significanceof regression
coefficients in linear regression analysis. Indeed,this analogy can be taken further so
that if the model has two parts which are very similar in their operation, it becomes
difficult to determine the parametric value of either of these parts reliably (for the same
reason that it is extremely difficult to find, by linear regression analysis, which of two
highly correlated independent variables has produced a given effect). It would therefore
seem justifiable that a conceptual model should be initially as simple as possible and
that additional modifications or components should be adopted only when these are
clearly seen to be necessary on physical grounds or have been shown to improve sig-
nificantly the reproduction of the observed hydrological system. One must also bear in
mind that the mere recognition of the existence of a physical phenomenon does not
necessarily indicate that this phenomenon should be reproduced in a conceptual model.
It'isnecessary that the effect of the phenomenon on the output be substantial.In some
cases the effect on the output may be variable,for instance with increasing precipitation
the effect may increase or decrease. This should be carefully checked and, to test it,
a method is required for the quantitativeevaluation of the efficiency of the model and
of the significance of each individual part. One such method might be the specification
of a measure of error expressing the disagreement between computed and observed
outputs and the comparison of the residual and initial values of this quantity. The
significance of a suspect part of a model of a single basin, or group of basins,might be
judged by the reduction obtained in the residual error on the introduction of this part
and subsequent optimization.
The neglect of this principle by the postulation of too complex a model in the first
instance may render difficult or impossible the extrapolation of results obtained from
one basin too another,or the recognitionof relationshipsbetween physical characteristics
and the parameters of the model-relationships which must be recognized if the pos-
tulation of a suitable conceptual model for ungauged basins is ever to be achieved.
It is as yet too early in the development of the technique to lay down any rules con-
cerning the elements to be incorporated in the actual model, but much can be learned
from the extensive work done by a few workers [21,951.
1 6.2.5.5.1 AUTOMATIC PARAMETER ADJUSTMENT
The successful operation of a digital computer in which the parametric values are ad-
justed by the operator has, until recently,relied to a considerable extent on the exper-
ience and personal judgement of the operator. However, optimization techniques have
been developed which determine the values of system parameters which maximize, or
minimize, some function dependent upon those parameters. These techniques are com-
pletely objective;many useless tests may be made of situations that would be dismissed
323
out of hand by an experienced hyman investigator, but the tremendous speed with
which a computer can make them compensates for such inefficiency.
In the basin-modelcontext, an obvious parameter-dependent function to be opti-
mized (minimized) isthe differencebetween an observed outflow and the outflow computed
by the model when supplied with the corresponding observed input.Other error criteria
could be used (e.g.,magnitude or timing of peak flows) or,in fact,any combinatioii of
such criteria.
One such technique has been developed [24]and applied to synthetic data of output
obtained by running the input data through a model with assigned parametric values.
It was demonstrated that the optimization techniques were capable of recovering the
specific parametric values efficiently and techniques were also developed for demonstra-
ting the sensitivity of the model to variation in each of the parametric values taken
separately.
6.3 Natural and cultural changes

Methods, available in the literature, for finding out the effects of natural or cultural
changes on the hydrological regimen have been strongly oriented towards demonstra-
ting that effects do occur,without providing much quantitativeinformation that can be
used elsewhere. The translation of the research results can therefore be done only by
implication and be applied only to study basins that are closely comparable to the re-
search basin. Such methods,mainly employing graphical analysis and linear normal mod-
els, are of interest in future avenues of research and are outlined briefly in section 6.3.1.
To aid the translation of results to other basins, hydrological characteristics can be
determined [13,36, 72,89,110, 1241.This might help to establish not only that effects
are observable,but also the magnitude of these effects.This is a wide field of study and
some restriction must be imposed upon any research (see section 6.3.2).
More complex mathematical models (described in sections 6.2.5.3to 6.2.5.5)could
be applied to determine the hydrological effects of natural and cultural changes. Some
details of possible approaches are given in section 6.3.3.
6.3.1 Graphical analysis and linear normal models

For a truly statistical analysis of the hydrological effect of natural and cultural changes,
it would be necessary to replicate and to change treatments from one basin to another.
In experimental basin research this procedure is too costly, too time consuming, and
too impractical(see section 2.4). Evaluation procedures are therefore dependent on time
trends of flow after precipitation and other climatic factors have been taken into account.
For the study of the effect on the hydrological regimen of cultural changes several years
of calibration (pre-treatment), a treatment period, and finally a period of years of eval-
uation of the treatment effect are required. This is not always feasible for the study of
the effect of natural changes but alternative successful techniques are not yet available.
Graphical display of the effects of natural and cultural changes in basins in streamflow
is an effective means of showing the approximate magnitude of time trends. Salient fea-
tures of relationships displayed on graphs lead a scientist to select arithmetic or loga-
rithmic form,straight-lineor curved relationships in his analytical analysis. Least squares
or. multivariate statistical techniques may then be used as an objective way of putting
numbers into the relationships,for putting confidence limits on these numbers and for
evaluating the statistical significance of the relationships.
Analytical techniques for studies in which only one basin is observed-before and
after treatment-will differ from those for studies where several comparative basins are
324
Analysis iechtziqiies arid itztevpreiation of research results
observed before treatment, with one remaining untreated throughout the study period
as a control and the others being treated. Surface and subsurface-waterstudies and the
determination of hydrological characteristicsare important for single-basinor compara-
tive basin studies in quantifying the effect of natural or cultural changes,as described
in sections 6.1.3,6.1.4and 6.3.2.
6.3.1.1 Test of representativeness of period of study
It is important that the study period be one in which the climate is representativeof the
long-termclimatic pattern. If the study period is excessively wet or dry. hot or cold,
etc., the results may not be suitable for application in a normal period.
One test of representativenesscompares data from a nearby long-termrecord with the
short-termrecord from the study area as shown in Figure 6.39 [42]. The comparison
may be made for data on run-off,precipitation, air temperature,pan evaporation,or
any other climatic factors which are known to be important and for which a suitable
long-term record can be found.If the study period is found not to be representativeof
the long-term climatic pattern, great care must be exercised in the application of the
study results.
25 , , , , ,
Legend
O = Actual record
__ - Expected values based $ 20
on 1919-48period.
6 15
Y IO
FIG.6.39. Normalcy
test for annual
precipitation. ,949 ,911 ,953 1955 ,957 i913 1951 1963 1965 1967
6.3.1.2 Single-basin technique
In the single-basintechnique, basin flow is related to the factors which influence it,
usually through regression equations which relate monthly, annual or storm run-offto
climatic factors. These equations are then applied to climatic factors in the post-treat-
ment period and the estimated run-offcompared with the actual run-off.Differences are
a measure of the effect of the change in basin characteristics,but must be examined in
relation to the magnitude of differenceswhich could occur within the confidence interval.
Rainfall/run-offrelationships [i 151 (Figs. 6.40and 6.41-mass curve and double mass
curve), show how the data can be graphed for initial study. Comparisons are made of
run-offdata between the 28-monthcalibrationperiod and the 54-monthevaluation period.
In this case only a slight decrease in surface run-offis noted. There will be situations
where the difference in relationship between precipitation and run-offbetween the two
periods will be sizeable and more readily apparent.A regression analysis technique [i151
is also usable to evaluate the effect of changes. Its results are in agreement with the
double mass curve technique (Fig. 6.41). Studies of flow volumes and peak discharges
between the calibration and evaluation periods are not valid unless causative factors
such as rainfall intensities and antecedent moisture for both periods are normal. Ab-
normally wet or dry periods,either before or after the natural or cultural change,could
result in erroneous conclusions.
325
9
CALIBRATION ACTION
E
-
I
6
PRECIPITATION
= 5
O
O
0
- 4
z
0
$ 3
3
$ 2 / RUNOFF-/;)
o
u
I
O
1954 1955 1956 1957 1958 i959 1960 1961 196:
FIG.6.40. Accumulated precipitation and run-off for calibration,

action and evaluation periods.
CALIBRATION 1 ACTION
O
O
0 I .2
I
IL
IL
z
æ
œ
w
O
O
M I
I
I I
2
I I
3
I I
4
I
5
1 I l
6
I I
7
I I
8
I 19
PRECIPITATION (i000M M I
FIG.6.41. Double mass curve: precipitation versus run-off.
326
One pre-treatment calibration regression equation [90]in a single-basintechnique is :
Y = 18.99+O.95Xi- 1.69X2-0.80X~
in which:
Y = annual run-off(mm);
XI = annual precipitation (mm);
XZ = annual evapotranspiration (mm);
X3 = total measured soil and ground-waterstorage change plus 102.4m m .
Other parameters may improve the reliability of prediction, especially those involving
relations between climatological and hydrological characteristics. Standard errors of
estimate in the example given were below the expected minimum practical change in
run-off;namely, about 10per cent of the annual flow. Predicted values for the evalua-
tion period are compared with observed ones and the effect of a natural or cultural
change can thus be evaluated.
The adequacy of the single-basintechnique depends on the ability to predict run-off
from climatic data with a reasonable degree of accuracy and the normality of the data-
collection periods. The research scientist should determine if the streamgauge measures
all of the flow from the basin or if there are any leaks [22](see section 6.3.1.3).
Statistical control can be provided by establishing basins in a nesting arrangement,
as discussed in section 2.4.
6.3.1.3 Comparative-basin technique (for the study of the effect on the hydro-
logical regimen of a cultural change)
The comparative-basin technique employs a control basin which remains untreated

throughout the calibration and evaluation periods.After satisfactory correlationbetween
flow from the control and other basins is attained,treatment is applied to all but the
control basin and observed flow from the treated basins is compared with that predicted
from the original regression equation based on flow data from the control basin. The
difference between predicted and observed values is a measure of the effect on the hydro-
logy of land treatment.
The minimum length of the pre-treatment or calibration period should be such that
when a regression equation is calculated to show the relation of run-offfrom one basin
to that of another,the residual deviations around the regression line are random and
conform reasonably well to a normal frequency distribution [134].Furthermore,it is
important that flow values between the two basins be correlated to a high degree,other-
wise the error band may be wider than the effect of treatment which the analysis tech-
nique is to evaluate.The calibration period should also be long enough to cover a wide
range of climatic conditions,both wet and dry years.
After the calibration period has run for several years, the data are analysed by the
method of covariance for agreement with the conditions listed above. Tables are given
in the literature which may be used to estimate how long the calibration period should
run if the data indicate that the calibration should continue. The method of analysis
for the evaluation period is also given [134].
The control-basinapproach has been used to show by graphical and statistical tech-
niques the magnitude of changes in run-offcaused by reafforestation by pine in a basin
of shallow-rootedscrub vegetation [42,76,771. A double mass curve of flow for the
control basin versus the treated one (Fig. 6.42)revealed a trend of change.Next,annual
flow values for both basins were plotted against time (Fig. 6.43). Variations in annual
amounts from the control basin were caused by climatic variations whereas variations
in the values for the treated basin were the result of both climate and treatment. Both
trends were downward,that for the treated basin being of greater slope. The standard
321
IRUN-OFF C O N T R O L BASIN (IOOOMk)
Fio. 6.42. Double mass curve of run-off from treated and control basins.
1 - (a) Treated, adjustedvalues. (Slope = 7.11 mrn/yr & 1.27;ri= -0.79).
6 O0
J
e
- o
O
I l I I I l ~ I I I I I I I I ~ ~
(b) Treated basin. (Slope ~ 9.40 mm/yr i 4.32;r = -0.49).
600
m 1
:I
a o
1949 1951 19ä3 I955 1957
I 1959 1961 1963 1965 1967
I
(c) Control basin.-(Slope= 2.80mm/yri 4.57;r = -0.15).
FIG.6.43. Time trends of annual run-off from treated and control basins.
328
Analysìs teclmiqries atid iiiterpretatioii of research remlts
error was large for both basins. The effect of year-to-yearvariations in climate was
removed from the treated basin data by the equation:
= (QR+ 12.4873)-0.8573Qc
where:
QR,= adjusted annual flow (reafforested basin) ;
QR= annual flow (reafforested basin);
Q c= annual flow (control basin).
Constants in the equation were determined statistically €oreach set of basin comparisons.
The value of this technique in comparative basin studies is evident in the reduction of
scatter ofthe points(Fig. 6.43~) and the lowered standard deviation from f4.32to &1.27.
A mathematical curve or straight line is fitted to the data according to the best defi-
nition. The data on Figure 6.43 were fitted with a straight line,since there was no evi-
dence in the plotting that the trend was beginning to level off. A curve can be fitted
where the data points so indicate [76,771.
A technique for evaluating the effect of land-usechanges on peak-dischargerates
has been described in the literature [42].Flood-peak discharges (mm/ha/hr)from the
reafforested basin (17 ha) as a percentage of the causative-rainfallrate in the same units
were compared with those of a control basin (120 ha) for the calibration and evaluation
periods (Fig. 6.44). Although these two basins are about 1,400 m apart the difference
in their rainfall rates for the same storm was great enough to invalidate a comparison
of peak-dischargevalues alone. In small-basin comparative studies,precipitation rates
0 01967
i960
@1960-6701949
01956 0 i966
P e a k discharge on control basin (percentage of 30 min. rain rate)
@Averages 0 Calibration period 0 Evaluation period
FIG. 6.44. Comparison of flood-peak discharges of treated and control basins.
329
for the basins may be so large as to invalidate the comparison of storm-flowamounts

or peak rates. For monthly and annual flow values,the precipitation difference exercises
lessinfluence.
Another technique has been described [22]for evaluating the effect of changes in land
use on basin hydrology to meet urgent needs that cannot wait for three separate periods
of years of data collection,calibration and treatment.The water balance for compara-
tive basins is given by:
E P-QfAMa f A G - Q d p (83)
where, over a specified period:
E = totalevapotranspiration;
P = totalprecipitation;
Q = totalstreadow;
AMs=change in soil-waterstorage in root system;
AG = change in ground-waterstorage below root system;
Q d p = total unmeasured flow out of the basin (deep percolation).
Total streamflow (Q)is usually the most accurate measurement. Precipitation-input(P)
evaluation requires an intensive network of gauges to make it of comparable accuracy.
Changes in ground-waterstorage (AG)are assumed to be zero,since the water-balance
studies are made over annual periods beginning and ending at times of minimum flow.
Root-zonesoil-waterstorage (AM,) for shallow-rootedcrops will be small compared
with other items in the annual balance study. However, for deep-rootedcrops, AMs
values may be large-as much as 500 mm in a depth of 6 m. AM,must be measured
where its value is expected to be large;this is practicable only where there is a high degree
of uniformity in soils and vegetative covers.
The estimation of leaks or deep percolation ( e d p ) is done by independent checks [22J
A comparison of E derived from the water-balanceequation with the potential evapo-
transpiration (Et),completed by the modified Penman method [69](see section 5.3.4)
provides a ratio EIE6 which can be used as an indicator of the reasonableness of the
results. This is a crude method and can reveal only large leaks.Where reliable measure-
ments of AMsand AG are made and where soil-watersuppliesare sufficient throughout
the year to permit E to approach the potential evapotranspiration(Et)making the ratio
E/Et constant, AM,values can be computed over short periods and compared with
observed A M , values. If there is a progressivedivergence-the predicted values exceeding
the obsemed soil-waterdeficit-this is evidence of a sizeable leak.
The best solution of the leak problem is to select basins where there are no sloping
strata of impervious material which would cause large unmeasurable errors (see also
section 4.3). In Kenya, for instance, it was found that the geological conditions were
such that in basins of 600 ha the surface and subsurface basins were essentially identical,
but for basins of less than 30 ha deep percolation was large.
These techniques were applied to forested (control) and partly-cleared (treated)
basins and produced reasonable results in a seven-yearperiod. The EIEt value for both
the control and treated basins was 0.93 at the start,and remained high for the control
basin for the period of study. Large areas of the treated basin were developed for tea
production and the EIEt ratio decreased to less than 0.80.For a controlbasin in bamboo,
E/Etwas 0.75 and for vegetables and pine seedlingsit was 0.66.In both cases the analy-
sis technique provided a means of evaluating the effect of treatment on evapotranspi-
ration.
The control versus treated-basintechnique can be used over a calibration and evalua-
tion period to quantify the effect of treatment on flow-durationcurves (Fig. 6.45)[91].
Mathematical equations (statistically sound) developed for predicting the run-offfrom
one basin from that of the control basin for the calibration period are used to predict
run-offfrom the treated basin in the post-treatmentperiod. Predicted values compared
with observed values reveal the effect of the change in land use.
330
Graphicaltreatment of basin data,statistical techniques [1081 and combined graphical

and statistical methodology [35]will be of considerable help to the research hydrologist.
$ 0.05 - \ -
L \
FIG.6.45. Flow-
4
I
o
- \ -
\
duration curves for ?!a 0.01 I I
6.3.2 Hydrological characteristics

Fundamental hydrological research implies the study of physical phenomena related
directly or indirectly to the hydrological cycle.Such research is carried out in represent-
ative and experimental basins by determining the hydrological characteristics(see section
6.1.3) and by studying the relation between them and the climate, geomorphological,
vegetational,pedological and geological characteristics (see sections 6.1.1 and 6.1.2).
Many investigations of this nature have been carried out on a piecemeal basis [26,42,
1321 and it appears that more logic should be applied before such analyses are begun,
since the number of causalrelations in any hydrological system is infinite and it is obvious
that a full descriptionof all causalrelations in the hydrologicalcycle is beyond the bounds
ofreality.
Hydrological characteristics have components of variability and randomness because
of varying influences of climatic and basin characteristics and when natural or cultural
changes occur in a basin, the non-stationarityof such hydrological systems will cause an
increase in the random component of the hydrological characteristics.
Hydrological characteristics are generally related to each other and to basin charac-
teristics in a non-linearfashion,and most of the effort expended in hydrological analy-
sis has been directed towards the formulation of procedures for the linear approximation
of such systems [4].
i For these reasons,a study of the hydrological characteristicswhich are likely to have
the smallest random component during a period when the basin is in a relatively station-
ary condition will be the most desirable. In this context it is important to realize that
apparently static characteristicsmay become dynamic during a hydrological event. For
I instance,studies in the U.S.A.[116], have indicated that during a rainstorm a variable
~
part of the basin area may contribute to surface run-offbecause lower portions of a
I 331
basin may have higher soil-moisturelevels at the beginning of run-ofthan relatively

elevated portions.
The hydrological cycle, being a sequential dynamic system,is in a non-stationary
equilibrium and this impliesthat a change in one characteristiccauses a change in one or
more other characteristics.
The setting-upof a working hypothesis as to which characteristics are expxted to
change as a result of the natural or cultural change under study will assist in limiting
the objective of the research.
6.3.2.1 Example of working hypothesis
The Puketurua experimental basin in New ZeaIand,approximately 300 ha in size,has

sub-basinson the main stream of 55 and 2 ha respxtivaly as statistical controls (see
section 2.4). The climate is humid with an average rainfall of 1,250mm,and high rainfall
intensities (50 mm with durations of one hour or less) are common.
The land ir m3derately steep with many small depressions, a shrubby vegetation,
approximately 2 m high with little or no understorey,and a thin litter.The soil is a pod-
zolized, heavy, im2ermeable clay and the ground-water table is at a relatively great
depth,providing base flow in all basins. In due course the land will be developed into
grass and it is anticipated that a dense,vigorous grass cover will develop and can be
maintained in the basin.
After some initial observations had been taken (simple soil and vegetation maps,
streamflow,precipitation, climate and soil-moisturemeasurements) and some geomor-
phological features had been studied,it was postulated that a number of importantbasin
and hydrological characteristics would undergo changes as set out in Table 6.4(the
postulate considers only the relatively stationary period before development (calibration
period) and the relatively stationary period commencing some time after development
(evaluation period) .)
As a result,it is possible to narrow the investigation,initially,to those characteristics
which are likely to be subjected to great changes and from this example it appears that,
in particular, phytomorphological characteristics and their relation to interception,in-
filtration,surface detention,soil moisture and evapotranspiration should be studied.Al-
though relatively large changes in the interflow regimen are expected,the over-allrates
of run-offdistribution with time are not expected to change greatly.
Work is in progress to determine the characteristics which are expected to undergo
great changes on a quantitative basis and to review the working hypothesis regularly
in the light of new information obtained [52,60,67,11 i].
6.3.3 Other mathematical models

The analysis of past changes in a basin can be attempted with the aid of mathematical
models based on the techniques described in sections6.2.5.3to 6.2.5.5on the assumption
that if a particular model is developed, which portrays with relative faithfulness and
consistency the operation of the system during some period in the past, any change in
the basin will produce a deterioration of the predictive ability of this model as time
goes by.
Some investigators have used the unit hydrograph in estimating changes in the ñood
regimen due to cultural changes. Such predictions can be attempted if the effect of the
proposed changes on the unit hydrograph can be anticipatedby experience or by analogy
with other basins. Caution, however,should be exercised in this regard.
As stated in section 6.2, the application of the unit-hydrograph method entails a
three-stageoper-.-m.The first and last operations of this cascade are based on largely
332
Analysis iechtziques and interpreiation of research results
I I I VIIAA AIA IIAAV
R 1 AV I
arbitrary numerical and geometrical constructions,which do not lend themselves to a

truly systematic application,because their inevitable lack of consistency precludes the
establishment of explicit mathematical relations.Hence this procedure does not furnish
completely reliable grounds to forecast future basin behaviour after a change, or to
prove or disprove the stationarity of a particular hydrological unit during the recorded
past,because any deteriorationof its predictive capacity may, depending on the situation,
be ascribed either to the inapplicability of the model within certain ranges of input,
or to an incorrect manipulation of the data in the first and last elements of the cascade.
O n the other hand,there is a possible alternative for the symptomatic assessment of
a system’s invariance by the methods of synthesis and pure analysis. The sequence of
operationsmight be as follows:
(a) The relation between gross input (precipitation) and gross output (streamflow)
would first be assumed to be a linear operation,as represented by the convolution in-
tegral.Now,even if the impulse response function were known,it could not be expected
to yield a faithful reconstruction of the flow record from its beginning, because of the
effect of previous inputs.However,these effects of remote events become more and more
attenuated,so that the reconstruction would improve uniformly towards the end of the
sequence.Therefore it would be reasonable to assume that,if the effect of remote inputs
is separated at the beginning of the flow record by an approximation such as the pro-
cedure of flow separation (see section 6.1.3.3.1), a first approximation of the system
response function could be obtained which would contain only the error due to this
initial correction. The estimate of the response function could then be improved by
systematic adjustment of the initial subsurface flow only.
It must be emphasized that the so-called subsurface flow separation should not be
interpreted as a true segregation of a real-flow component. It would simply have the
character of a correction for the vanishing effect of previous unrecorded inputs.
If the above trial did not yield a uniformly satisfactory fit between computed and
recorded streamflows (subject to some arbitrary criterion for acceptance), it would be
concluded that the linkage between gross input and output is not linear;or,if a gradual
worsening of thepredictionisobserved after tryinga reasonablycomplete setofinitialsub-
surfaceflow-separationoperations,one would suspectthat the system is,in fact,changing.
(b) If no systematic trend towards improvement or worsening of the prediction is
observed by step (a) above, the procedures of general non-linear analysis (see section
6.2.5.4)or general synthesis (see section 6.2.5.5)would be tried. A process similar to
the one just described for the case of the linear assumption would lead to indications
regarding the stationarity of the system.
For any conclusions on system variability to be valid,evidence must have been reached
that the model is competent in describing the operation of the system for a period of
sufficient length, and covering an adequately wide range of input variations, to sup-
port the assumption that it was faithful before the change took place.
If the recorded hydrograph appears to show a gradual trend towardsmore pronounced
fluctuationsthan the computed hydrograph,the basin may be assumed to be evolving
in the direction of a diminishing damping power. The converse conclusion would be
reached if the record gradually becomes smoother than the computed flow sequence.
If, on the other hand, no systematic trend is observed with the passage of time but
unexplained inconsistencies appear between the recorded and computed hydrographs,
a more profound alteration of the regimen of the basin may be suspected.
334
6.4 Water-balance studies

In any representative or experimental basin the calculation of a water balance is of
considerable help before detailed data analysis is undertaken.
The general equation for the water balance is :
P = Q+EfAMs&AG+dV+e (84)
where, over a specified period:
P = totalprecipitation;
I
I
Q = totalstreamflow;
E = totalevapotranspiration;
A M , = change in soil-moisturestorage;
AG = change in ground-waterstorage;
A V = change in storage of liquid and solid precipitation in endoreic depressions;
e = an error term which includes not only deep percolation ( e d p ) but also errors
associated with other elements of the water balance.
All the water-balanceelements are calculated as mean values for a basin and expressed
in depths (mm).
In exceptional cases,representative or experimental basins have to be selected when
economic activities affect the water balance. In such cases additional terms should be
introduced,e.g. (Qcx), amount of water used for economic activity (irrigation,water
supply,etc.); (QB)amount of water returned from the economic activity (effluent dis-
charge,etc.).
In basins containing one or more natural or artificial reservoirs which affect the ñow
regimen significantly,additional parameters which characterizethe increase or decrease
of storage should be introduced.
Below are given some details of the main elements of the water balance. For further
details, refer to the relevant sections of chapters 4 and 5 [7,56, 57, 1391.
(a) Precipitation. Data obtained at precipitation stations within a basin are used for
the determination of the mean basin precipitation.Sometimes it is expedient to use data
obtained at precipitation stations outside the basin. If the analysis shows that data are
doubtful at any one station,they should be either excluded or corrected by correlation
with data from adjacent stations.In winter,when the error in solid precipitation measure-
ments is especially great, the amount of precipitation is estimated, in case of lack of
thaw, by the difference between the snow pack at the end and at the beginning of the
period, taking into account evaporation from the snow surface.In regions where thaw
takes place regularly the amount of precipitation for the winter period is determined
solely from readings from precipitation gauges, and corresponding corrections are
introduced.
(b) Streamflow.Flow data should,if possible,be separated into surface and subsurface
components.Flow-separationmethods are given in section 6.1.3.3.1.
(c) Evapotranspiration. This should be estimated by using any of the methods given
in section 5.3.4.If possible,evaporimeters (see section 4.2.4.2.1) or the energy-balance
method (see section 5.3.5)should be used. For the periods when the basin is covered
with snow,evaporation values may be determined from observed data of snow evapo-
ration.These values are usually small and vary slightly from year to year (e.g.25-30mm
per year in the greater part of the U.S.S.R). In spite of the relatively small variability
of the evaporation from snow,its role in the water balance of some arid and semi-arid
regions may be rather significant.
(d) Change in soil-moisture storage. Soil-moisture variations occur throughout the
year in most regions and may affect the water balance for any given period. Mean soil-
moisture values for a basin observed at the beginning and end of a water-balanceperiod
335
are used to determine the change in storage and whether this change is positive or neg-
ative.
(e) Ground-water storage (provided that the data is representative of the research
basin). Ground-water levels are obtained from wells located within the basin or from
similar,adjacent basins. The change in ground-waterstorage for any one period is cal-
culated by multiplying the difference in mean basin ground-waterlevels at the beginning
and end of the period by the specific yield.Values of specific yield are found in the lit-
erature or may be calculated approximately as follows:
su= p-Fe
where Suis the specific yield,p the effective porosity, and F, the field capacity of the
aquifer under consideration.
In some cases the change in ground-waterstorage may be calculated by relating base-
flow to the mean ground-waterlevel of a basin. This relation is similar to a storage/
discharge rating curve in that baseflows,observed during relatively stable low-flow per-
iods,are plotted against mean basin ground-waterlevels and a curve of relation drawn
through the points.
(f) Change in storage of liquid and solid precipitation. The solid precipitation accumu-
lated on the land surface during the winter period in the form of snow and water is par-
ticularly important.In some basins depression storage caused by rainfall may also be
important.
Data for the various components of the water balance may be tabulated as shown
in Figure 6.46.Graphical methods for water balance studiesare given in the literature [12].
6.5 Translation of results to other basins

The translation of research results from one basin or group of basins to others is a pre-
diction problem.
This prediction problem applies to all components of the hydrological cycle-evapo-
ration,surface run-off,baseflow,recharge of aquifers,interception,erosion,etc.,but so
far the main work has concentrated on the prediction of surface flow.
As an example of translation of results,some details are given of streamflow predic-
tion which may be classified into three types, as follows: (a) flood prediction for the
design of protective structures such as spillways,flood channels, culverts, stop banks,
etc; (b) prediction of streamflows for long-range operations and basin-yield studies;
(c) short-termprediction for flood-warningservices and for basin-controlsystem opera-
tions.
The use of a particular type of model for the solution of these problems in the basin
for which it was originally developed determines to a large extent the success it is likely
to have in predicting the behaviour of any ungauged basin to which it may be extended,
and it is important to realize that for a sound prediction two conditions are necessary:
first,good fitting of the models; second,a wise choice of the physical parameters of the
basin and the precise quantitative determination of such parameters.
6.5.1 Flood prediction

Once a model is established for a particular basin, this application is probably the most
straightforward use of any model. The experience gained so far with the unit hydro-
graph (see section 6.2.5.3)and with conceptualmodels (see section 6.2.5.5)has indicated
that the prediction of flood peaks is not as reliable as might be desired,particularly in
336
x
YI
o
.-
Y
.e
Y
a
.e
c
u
337
Represeritative arid experimental basins
the case of very high floods.Because this is associated with imperfections of the models
as a whole, or of their individual components,it can certainly be hoped that the greater
number of more refined structures being perfected will result in a correspondingimprove
ment in the quality of the results. Some of the subjective factors which are currently
involved in building the models and in optimizing their parameters may be gradually
eliminated with further research [i i].
The non-linear functional models (see section 6.2.5.4)are likely to be the best pre-
dictors of high floods if the necessary equations can be developed from sufficiently long
records of rainfall and run-off.Because the mathematical operations involved are com-
pletely non-subjective,many of the sources of bias which are present in synthesis do not
exist in this process. However, it must be indicated that, at present, no experience is
available in the standardized use of non-linear analysis methods.
6.5.2 Long-range prediction of streamflows

Statistical methods, among which may be mentioned the use of Monte Carlo proce-
duresfor the generation of simulatedrainfall sequences,may yield synthetic time seriesof
input values which, used in connexion with conceptual models, (see section 6.2.5.5)or
with non-linearfunctional models (see section 6.2.5.4)permit the generation of the cor-
responding flow sequences. Long synthetic series obtained in this manner could then be
used in connexionwith basin-yieldstudiesor generalwater-resourcesplanning and system
operations.This theoreticalpossibility is limited by the fact that its predictive reliability
is quite sensitive to the faithfulness of the models. It can be shown statistically that
the models must be capable of yielding very close matching between the statistical prop-
erties of the observed-flowrecord and the computed outputsin order to permit attaching
validity to any conclusions drawn from this type of study.
6.5.3 Short-term prediction of streamflows

The application of conceptual (see section 6.2.5.5)or non-linear functionalmodels (see
section 6.2.5.4)is also straightforward in this case, and it can be said to improve with
the multiplicity of the input data.
Conceptual models with distinct components representing tributary basins of large,
controlled river systems have permitted the use of analogue techniques of considerable
complexity.These techniques permit the rapid manipulation of component operations
to obtain quickly many alternative forecasts.
A very promising technique which is analogous to the process of system analysis is
currently under investigation [lo]. This approach attempts the determination of the im-
pulse-responsefunction of a system on the basis of the stochastic properties of the input
and output sequences, rather than on their absolute values. Predictions are made of
ranges of river flows that can be expected with a certain level of probability. It is believed
that this approach is a realistic one in that the procedure recognizes the fact that models
and prototypes have only a limited equivalence and that errors in matching can always
be expected.
6.5.4 Summary of prediction techniques

The ability of the various methods to predict the behaviour of ungauged basins can be
deduced from the above remarks and from the discussionscontained in sections 6.2 and
6.3.3.They can be summarized as follows.
338
Analysis techniques and inferpre?u?ioii o,f research resuits
(a) The unit hydrograph inerhod, because it involves the determination of a specific
response function on the basis of records from a specific basin, does not furnish a direct
means of extrapolating in space. However, as indicated in section 6.2.5.3.4,several in-
vestigators have sometimesattempted to derive empirical relationships between the char-
acteristics of a basin and the unit hydrograph parameters,so that synthetic unit hydro-
graphs can be derived for the ungauged basins when only certain geomorphological or
other physical characteristics are known [log].
Various investigators [33,39,79,11 31 have used regression analysis in a process where-
by unit hydrographs have been derived for a number of streams in a region, and the
numerical measures of the basin’s characteristicshave been related to descriptive meas-
ures of the various unit hydrographs. The problem of the form of the relationships
which might be expected is one that has not been satisfactorily solved. In the absence
of guidance,most investigators have used characteristics such as area,basin and channel
slope,stream length, drainage density,etc. (see section 6.1.1.2) and have attempted to
select the relationships which are statistically more significant. It has been observed that
very different combinations of basin characteristics often serve almost equally well,
thus indicating high mutual dependence between these variables. Studies of basin char-
acteristics should be directed to the elucidation of some of these relationships with a
view to determining the number and combination of characteristics which are more
fundamental or to which the hydrological response might be more clearly related. An
example of such a study,using factor analysis,is given in the literature [117].
Graphical analyses have been used in West Africa for sixty basins using storm run-off
coefficients and elements of unit hydrographs for the prediction of ten-yearfloods [99].
The study was hampered by difficulties in standardizing the determination of the ele-
ments of the unit hydrograph and the basin characteristics (see Figs. 6.47 and 6.48).
A R E A (KU’)
FIG.6.47. Example of graphical analysis (run-off-rainfallagainst slope and area) for various
basins in West Africa. Sahelian and subdesertic areas (annual precipitation 150-800 mm);
permeability of soil: PI -P2 (impervious).
339
f
T
FIG.6.48. Example of graphical analysis (rise time-slope and area) for various basins in
West Africa.
340
The unit graph is limited to very small basins and for larger basins the isochrone
method may be more useful [136].
The results of the above methods have had varying degrees of success,but they do
not permit generalizations.Much additional work is required before standard procedures
can be recommended.
(b) The procedures of general non-linear analysis, when applied to the description of
the behaviour of a basin,have the same limitations as the unit hydrograph in the sense
that the impulse response functions of all orders apply strictly only to the basin for
which they were derived. Therefore,they cannot be used as a rule to predict directly
the behaviour of ungauged basins.
O n the other hand, if non-linear analysis is used in conjunction with a conceptual
model to describe the operation of one or more of its components,it may have the same
range of applicability as these models.
(c) Conceptualmodels appear promising for the prediction of the behaviour of ungauged
basins, provided structures of sufficient generality are developed. Their limitations
depend on how well each mechanism can be described in the representative or experi-
mental basins,as well as in the ungauged basins, and on how demanding they are with
respect to input data. The words of caution given in section 6.2regarding the physical
interpretation of model parameters vis-a-visthe characteristics of the prototype should
always be kept in mind.
References
O.A. J 953.Osnovy gidrokhiniii [Principles of hydrochemistry], 2nd ed.Leningrad,
1. ALEKIN,
Gidrometeoizdat.
2. AMOROCHO, J. 1961.Predicting storm run-offon small experimental watersheds.Discussion.
PP-OC.ASCE, 82 (HYZ): 185-91.
3. __- . 1963. Measures of the linearity of hydrologic systems. J. geophys. Res., 68(8) :
2237-49.
4. ____ . 1966. The non-linear prediction problem in the stud-v of the riin-off’ cycle.
5. __- ; ORLOB, G.T. 1961.Non-linearanalysis of hydrologic systems. Water Resources
Centre contrib., 40 : 1-47. Berkeley, University of California.
6. ANDERSON, H.W. 1957. Relating sediment yields to watershed variables. Trans. Amer.
Geop-s. Un.,38(6) :921-4.
7. ANDREJANOV, V. G.1960. Vnutrigodovoe raspredelenie rechnogo stoka [Annual distribution
of streamflow]. Leningrad, ci idrometeoizdat.
8. BALEK, J. 1965.The importance of beta-radioactivitymeasurements in representative and
research areas. ZASH publ. m. 66 (2). (Symposium of Budapest,)
9. BECKER, A. 1966. Ergebnisse einer Untersuchung über die Struktur der koaxralen gra-
phischen Storkregen-Abfluss-Beziehungen.Wasserwirtschaft-Wassertechiiik, 16(3) :90.
10. BERNIER, J. 1962. Théorie stochastique des réservoirs. L a houille blanche, 20(5) :434-44.
Il. ___ ; VERON,R. 1964. Sur quelques difficultés rencontrées dans l’estimation d’un
débit de crue de probabilité donnée. Revue de statisfique appliquée, Paris, 12(1).
12. BLOEMEN, 1967.Determination of deep recharge on discharge at a constant rate studying
fluctuations in ground-water level. J. Hydrol. (Netherlands).
13. BOCHKOV,A. P. 1965. O vlijanii agrotekhnicheskikh i lesomeliorativnykh meroprijativ
na stok rek lesostepnykh i stepnykh raionov [Onthe influence of agrotechnical and forest
reclamation measures on river run-off in forest/steppe and steppe zones]. T d y GGI,
vyp. 127, p. 10-81.
14. BODMAN, G.B.;COLMAN, E.A.1943. Moisture and energy conditions during downward
entry of water into soils. Proc. Soil Sci. Soc. Amer., 8 : 116-22.
15. BOUCHARDEAU, A. 1959. Méthode d‘extrapolation du coefficient de ruissellement sur les
bassins expérimentaux de la zone sahélienne du Tchad.Annuaire hydrologique de l’Orstom,
1957. Paris, Orstom.
341
16. BRUNET-MORET, Y.1965. Influencedu corps de l'averse sur le ruissellement d'un petit bassin.
Paris. (Cahier Orstom d'hydrologie no. 3.)
17. CAPPUS, P.1960. Bassin expérimental d'Alrance. Étude des lois d'écoulement. Application
au calcul et à la prévision des débits. Mémoires et travaux de la Société hydrotechnique de
France, no. 1. Paris.
18. CHAPMAN, C. A. 1952. A new quantitative method of topographical analysis. Amer. J.
Sci., 250 :428-52.
19. CHILDS, E. C.;COLLIS-GEORGE, N.C. 1950. The permeability of porous materials. Proc.
Roy. Soc., A201 :392-405.
20. COLMAN, E.A.;BODMAN, G . B. 1944. Moisture and energy conditions during downward
entry of water into moist and layered soils. Proc. Soil Sci. Soc. Amer., 9 :3-11.
21. CRAWFORD, N.H.;LINSLEY, R. K. 1963. A conceptual model of the hydrologic cycle.
IASH publ. no. 63 :573-87,
22. DAGG, M.; BLACKIE, J. R.1965. Studies of the effects of changes in land use on the hydro-
logical cycle in East Africa by means of experimental watersheds. IASH publ. no. 1q4):
63-75.
23. DARCY, H . 1856. Les fontaines publiques de la Ville de Dijon. Paris, Victor Daimont.
24. DAWDY, D . R.;O'DONNELL, T. 1965. Mathematical models of catchment behaviour.
P ~ O C .ASCE, 8(HY4) :123-37.
25. DERIDDER, N. A.;WIT, K.E. 1966. Evaluation of seepage and salt intrusion in the
south-western part of the Netherlands. J. Hydrol. (Neth.).
26. DILS, R. E. 1953. Influence of forest cutting and mountain farming on some vegetation,
surface soil and surface run-off characteristics. (US.Forest Serv. paper no. 24.)
27. DISKIN, M.H.1964.A basic study of the linearity of the rainfallIriIn-offprocessin watersheds.
Ph.D. thesis, University of Illinois.
28. DOOGE, J. C.I. 1959.A general theory of the unit hydrograph.J.geophys.Res.,62(2) :241-56.
29. ___ . 1965. Analysis of linear systems by means of Laguerre functions.J. Soc. indust.
und appl. Math. (control) series A. 2(3).
30. DIJEIRELJIL, P. 1966. Les caractères physiques er morphologiques des bassins versants. Leur
détermination avec une précision acceptable. Paris. (Cahier Orstom d'hydrologie no. 5.)
31. DUMM, L. D. 1954. A new drain spacing formula. Agric Eng., 35 :726-30.
32. EAGLESON, P. S. et al. 1965. The computation of the optimum realizable unit hydrographs
from raiirfall and run-off datu. MIT. Hydromechanics Laboratory. (Report no. 84.)
33. EDSON, C.G.1951. Parameters for relating unit hydrographs to watershed characteristics.
Trans. Amer. Geophys. Un.,32 :951-9.
34. EDELMAN, J. H.1947. On the computation of groundwater flow. Thesis, Tech. University,
Delft, The Netherlands.
35. EZEKEL, M.;Fox,K.A.1961. Methods of correlation and regression analyses. N e w York,
Wiley.
36. FEDOROV, S. F. 1962. Vlijanie lesa na vodny balans malykh vodosborov/po materialam
Valdaiskoy gidrologicheskoy laboratorii [The influence of forest on the water balance
of small basins/from data obtained at the Valdai Hydrological Research Laboratory].
Trudy GGZ,vyp. 95, p. 55-100.
37. GARDNER, W. R. 1958. Some steady state solutions of the unsaturated moisture flow
equations with application to evaporation from a water table. Soil Sci., 85 :228-33.
38. GOITSCHALK, L. C. 1964. Reservoir sedimentation. In: Ven Te Chow (ed.). Handbook
of applied hydrology, p. 17-1 to 17-33. N e w York, McGraw-Hill.
39. GRAY, D.M.1961. Synthetic hydrographs for small watersheds. Proc. ASCE, 87 (HY4):
33-54.
40. HADLEY, R. F.; SCHUMM, S. A. 1961. Hydrology of the upper Cheyenne River basin:
Sediment sources and drainage basin characteristics. USGS Water Supply Paper 1531-B:
137-97.
41. HARR, M. E. 1962. Groundwater and seepage. New York, McGraw-Hill.
42. HARROLD, L.L.et al. 1962. Influence of land use and treatment on the hydrology of small
watersheds in Coshocton, Ohio, 1938-57.( U S D A Tech. bull. 1256.)
43. HOLTAN, H.N. 1945.Time condensation in hydrograph analysis. Trans. Amer. Geophys.
Un.,407-13.
44. ____ ; KIRKPATRICK,M. H.,Jr. 1950. Rainfall, infiltration, and hydraulics of flow
in run-off computation. Trans. Amer. Geoph.ys. Un.,31 :771-9.
342
Analysis techniques and interpretation of resezirch results
45. -- ; OVERTON,
D. E. 1964. Storage flow hysteresis in hydrograph synthesis. J.
Hydrol. (Neth.),2 :309-23.
46. HORNER,
W.W.;LLOYD,
C. L.1940. Infiltration capacity values as determined from a
study of an eighteen month record at Edwardville,Illinois. Trans. Amer. Geophys. Un.,
pt. 2: 52241.
47. HORTON, R.E. 1938. The interpretation and application of run-offplot experiment with
reference to soil erosion problem.Proc. Soil Sci. Soc. Amer., 3: 340-9.
48. INTERAGENCYCOMMITTEE ON WATER RESOURCES.
1966.Methods of flow frequency analysis.
Notes on hydrological activities, Bull. 13.
49. JOHNSON, E.A.; DILS, R.E.1956. Oufliiie for compifing precipitation,riin-ofland ground-
water data bom srriall watersheds. US. Forest Serv. (Tech. note no. 34.)
50. KALININ, G. P.;MILYUKOV, P.I. 1958. Approximate calculations of the unsteady jlow
of water masses. (Trudy Ts. IP, iss. 66.)
51. KNISEL, W.G.1963. Base flow recession analysis for comparison of drainage basins and
geology. J. geophys. Res.,68 : 3649-53.
52. KORZOUN, V. I. 1968. Stok i poteri talykh vod na sklonakh polevykh vodoshorov [Run-off
and snowmeltlosses on the slopes of field catchments]. Leningrad,Gidrometeoizdat, 168 p.
53. KOSTIAKOV, A.N.1932. O dnamike koefficienta prosachivania vody v pochvogruntakh
i neobkhodimosti dinamicheskogo podkhoda k eye izucheniu v meliorativnykh tseliakh
[On the dynamics of the coefficient of water percolation in soils and the necessity for
studying it from a dynamic point of view for purposes of amelioration].Pochvovcdenie (3),
p. 293-7.
54. KOVZEL, A.G.1951.Opyt proektirovania gidrogafavesennego stoka dla malogo vodosbora
(Some experience of the computation of a spring run-offhydrograph for small basins).
Trudy GGZ, vyp. 31, p. 54-74.
55. KRAYENHOFF VAN DE LEUR, D.A. 1958. A study of non-steadyflow with special reference
to a reservoir coefficient. De Ingenieur, Delft,70B :70-84.
56. KUDELIN, B. I. (ed.). 1966. Podzenniy stok na territorii SSSR (Subsurface flow in the
territory of the U.S.S.R.). Moscow University.
57. KUZIN, P.S. 1960. Klassifikatsia rek i gidrologicheskoe raionirovanie SSSR [River classi-
fication and territorial division of the U.S.S.R.into hydrological regions]. Leningrad,
Gidrometeoizdat.
58. KUZMIN, P.P.1961. Protsess-v tajariia snezhiiogo pokrova [Snowpack melting processes].
Leningrad,Gidrometeoizdat.
59. KUZNETSOV, V. I. 1964. Isparenie so sezhnogo pokrova [Evaporation from snow cover].
Trudy GGI, vyp. 109, p. 3-56.
60. KUZNIK, I. A. 1962. Agrolesonieliorativnye nieropriatia, vessenniy siok i erozia pochv
[Forestreclamationmeasures,spring run-offand soil erosion]. Leningrad,Gidrometeoizdat.
61. LANGBEIN, W.B. 1940. Some channel storage and unit hydrograph studies. Trans. Amer.
Geophys. Un.,21 :620-7.
62. - - et al. 1947.Topographic characteristics of drainage basins. 157 p. (USGS Water
supply paper 968-C.)
63. LARRIEV, J. 1954. Contribution à l’étude des crues. L a honille blanche, 8:125.
64. LEOPOLD, L.B.; MADDOCK, T.1953. The hydraulic geometry of’stream channels and some
physiographic implications. (USGS Prof.paper 252.)
65. LINSLEY, R. K. et al. 1949. Applied hydrology. New York, McGraw-Hill.
66. LUTHIN, J. N.(ed.). 1957. Drainage of agricrrltural lands. Amer. Soc. Agron.
67. LVOVICH, M. 1. 1963. Chelovek i vody [Man and water]. Moscow, Geografgiz.
68. MAASLAND, M.1959.Water table fluctuationsinduced by intermittentrecharge.J.geophys.
Res., 64 :59.
69. MCCULLOCH, J. S. G. 1965.Tables for the rapid computation of the Penman estimate of
evaporation. J. E.Afr. Agric. and Forestr-v,30(3) :286-95.
70. MEHMETCIK, B. 1966. Instantaneous unit hydrograph derivation by spectral analysis and
its numerica! application. Cento Svmp. on Hydrol. arid Water Resources Development.
Ankara.
71. MILLER, R.D.;RICHARDS, F. 1952. Hydraulic gradients during infiltration in soils.Proc.
Soil Sci. Soc. Amer., 16 :33-8.
72. MOLCHANOV, A.A. 1960.Gidrologicheskaya rol lesa [Hydrologicalrole of forests], 2nd ed.
Moscow, Academy of Sciences of the U.S.S.R.
343
Representative and experimental busins
73. MORISAWA, M. E. 1962. Quantitative geomorphology of some watersheds in the Appa-

lachian plateau. Bull. Geol. Soc. Amer., 73 :1025-46.
74. MUSGRAVE, G.W.;HOLTAN, H.N.1964.Infiltration.In: Ven Te Chow (ed.). Handbook
of applied hydrology, p. 12-1 to 12-30, section 12. New York, McGraw-Hill.
75. MUSKAT, M.1937. The flow of homogeneous fluids through porous media. Ann Arbor,
Michigan, J. W.Edwards.
76. NAKANO, H.; KIKUYA, A. 1963. Effects of changes in forest conditions,especially cutting,
on run-o#'(I). Tokyo, Govt. Forest Expt. Stn. (ßull. no. 156.)
77. ___ __-__ . 1964. Effects of changes in forest conditions, especially cutting, on
run-of (II). Tokyo, Govt. Forest Expt. Stn. (Buli. no. 170.)
78. NASH, J. E. 1958. Determining run-offfrom rainfall.Proc. Inst. Civil Eng.,London, 10.
79. . 1960.A unit hydrograph study with particular reference to British catchments.
Proc. Inst. Civil Eng.,London, 17 :249-82.
80. NEMEC, J.; MOUDHY, M.1965.Some considerationson the use of experimentalwatersheds
for modelling of hydrological process of surface run-off. IASH publ. no 66, 1 :84-8.
81. O'DONNELL, T. 1960. Instantaneous unit hydrograph derivation by harmonic analysis.
IASH publ. no. 51 :546-57.
82. OGNEVA, T.A.1966.Raschet isparenia za konkretnyeperiody vremeni metodom teplovogo
balansa [Computing of evaporation for definite time intervals by heat budget method].
Materialy mezhduvedomstvennogo soveschania PO Probleme izuchenia i obosnovania metodov
rascheta isparenia s vodnoy poverkhnosti i sushi, p. 23247. Valdai, GGI.
83. PHILIP, J. R.1957. Numerical solution of equations of the diffusion type with diffusivity
concentration dependent. Australian J. Physics,10.
84. . 1957. Sorptive and algebraic infiltration equations. Soil Sci.,84 :257-64.
85. . 1966. A linearization technique for the study of infiltration. Unesco symp. on
water in the unsaturated zone, paper IIa-9.Wageningen.
86. PEST,R.F.;HEINEMANN, H.G.1965. Experimental watersheds contribute useful sedi-
mentation facts. IASH publ. no. 66 : 372-8.
87. POL~ARINOVA-KOCHINA, P. 1962.Theory of groiindwater movement. Transl.from Russian
by R.J. de Wiest. Princeton,Princeton University Press.
88. POPOV, O.V. 1964. Metodika izuchenia i rascheta podzemnogo pitania rek [Methods of
study and computation of subsurface alimentation of rivers]. Trudy GGI,vyp. 114,p. 5-86.
89. RAKHMANOV, V. V. 1962. Vodo okhrannaya rol lesa [Water-conservingrole of forest].
Moscow, Goslesbumizdat.
90. REIGNER, I. C. 1964. Calibrating a watershed by using climatic data. U.S.Forest Serv.
(Res.paper NE 15 :45.)
91. REINHART, K.G.et al. 1963. Effectof streamflow on four forestpractices in the mountains
of West Virginia. U.S.Forest Serv. (Res. paper NE I.)
92. REZNIKOV, A.A.et al. 1963.Metody analizaprir odnykh vod [Methodsof water-resources
analysis]. Moscow, Gosgeoltekhizdat.
93. RICHARDS, L.A. 1931. Capillary conduction of liquids through porous media. Physicr,
1 :318-33.
94. RIJTEMA, P. E. 1965. An analysis of actual evapotranspiration. Wageningen, Pudoc.
95. ROCHE, M.1965.Point de vue matriciel sur un opérateur linéaire de transformation pluies-
débits. Paris. (Cahier Orstom d'hydrologie, no. 2.)
96. . 1966.Recherche d'un hydrogramme standard.Paris.(Cahier Orstom d'hydrologie,
no. 6.)
97. RODIER, J. A. 1959. Résultats obtenus sur les bassins expérimentaux de la France d'Outre
Mer. Paris, Orstom. (Mémoires et travaux de la SHF,no. 11.)
98. . 1966. Méthodes utilisées pour le calcul du coeficient de ruissellement sur les
bassins représentatifs et expérimentaux. Paris. (Cahier Orstom d'hydrologie, no. 4.)
99. ; AUVKAY, C. 1965. Estimation des débits de crues décennales pour les bassins
versants de superficie inférieure à 200 km2 en Afrique occidentale. Paris,Orstom-Cieh.
100. ROMANOV, V. V. 1962. Isparenie s bolot Evropeyskoy territorii SSSR [Evaporation from
swamps of the European part of the U.S.S.R.]. Leningrad,Gidrometeoizdat.
101. RUBIN, J. 1964.Soil/waterrelations during rain infiltration.III:Water uptake at incipient
ponding. Proc. Soil Sci. Soc Amer., 28 :614-9.
102. . 1966.Numerical analysis of ponded rainfall infiltration. Unesco symp. on water
in the unsaturated zone, paper IIa-2.Wageningen.
344
Analysis techniques and interpretarion of research results
103. SATO,S.;MIKKAWA, H.1956.A method of estimating run-offfrom rainfall.Proc. Regional

Tech. ConJ on Water Resources Development in Asia and the Far East. United Nations.
(Flood control series, no. 9.)
104. SCHUUM,S. A. 1954. The relation of drainage basin relief to sediment loss.Tenth general
assembly, IASH, 1 :2169. Rome.
105. ___ . 1956. Evolution of drainage basins and slopes in badlands at Perth Amboy,
N e w Jersey. Bull. Geoì. Soc. Amer., 67 :597-646.
106. SHARP,A.L.;HOLTAN, H.N.1942.Extension of graphic methods of analysis of sprinkled
plot hydrographs of control plots and small homogeneous watersheds. Trans. Amer.
Geophys. Un.,vol. 23 :578-93.
107. SHERMAN,L. K. 1932. Streamflow from rainfall by the unitgraph method. Eng. News-
Record, 108 :501-5.
108. SNEDECOR,G.W. 1946. Statistical methods. Ames, Iowa, Iowa State College Press.
109. SNYDER, F.F. 1938. Synthetic unit-graphs.Trans. Amer. Geophys. Un.,19 :447-54.
110. SOKOLOV,A. A. 1962. O vlijanii lesa na maksimalny stok vesennego polovodja [On the
forest effect on the maximum spring flood]. Trudy GGZ, vyp. 99, p. 79-140.
111. SOUBBOTIN, A.I. 1966.Stok talykh i dozhdevykh vod/poexperimentalnym dannym [Run-off
of snowmelt and rainfall waters/from experimental data]. Moscow, Gidrometeoizdat.
112. STRAHLER,A.N.1957. Quantitative analysis of watershed geomorphology. Trans. Amer.
Geophys. Un.,36(6) :913-20.
113. TAYLOR, A. B.; SCHWARZ, H.E. 1952. Unit-hydrograph lag and peak flow related to
basin characteristics. Trans. Amer. Geophys. Un.,33(2) :235-46.
114. TENNESSEE VALLEY AUTHORITY. 1961. Matrix operations in hydrograph computations.
(TVA Res. paper no. 1.)
115. ____ . 1963.Parker Branch research watershed. North Carolina State College. (Proj.
authorization no. 700.)
116. 1964.Bradshaiv Creek-Elk River. ‘4 pilor sftidy in area-streamjQctor correlation.
Office of Tributary Area Development. (Res.paper no. 7.)
117. . 1966. Design of a hydrologic condition snrvey rising factor analvsis. TVA, Div.
of Water Control Plannjng. (Res. paper no. 5.)
118. TIMMERS, H.J. 1955. Determination of soil permeability in situ. Neth. J. agric. Sci., 3 :
11 9-26.
119. TODD,D.K. 1959. Groundwater hydrology. N e w York, Wiley.
120. TOEBES, C. 1964.Itfiltration analysis.Wellington,SC and RCC.(Handbook of hydrological
procedures, prov. procs. 21, 22, 23, 24, 25.)
121. ___ ; STRANG,D.D.1964. On recession curves. 1. Recession equations. J. Hydr-01.
(NZ), 3(2) :2-15.
122. UNITED STATESDEPARTMENT OF AGRICULTURE. SOILCONSERVATION SERVICE. 1956.Hydro-
logy. In : Engineering handbook, section 4.
123. V A N DUIN, R. H.A. 1955.Tillage in relation to rainfall intensity and infiltration capacity
in soils. Neth. J. agric. Sci.,3 : 182-91.
124. VOLFTSUN,I. B. 1962. Izmenenie usloviy formirovania poverkhnostogo stoka talykh vod
pod viijaniem lesonasazhdeniy [Change of formation conditions of surface run-off of
snowmelt water under the influence of afforestation]. Trudy GGI, vyp. 95, p. 29-54.
125. VORONKOV,P. P. 1963. Gidrokhimicheskie obosnovania vydelenia mestnogo stoka i
sposob raschlenenia ego gidrografa [Hydrochemicalbasis of local run-offseparation and
methods of its hydrograph analysis]. J. Meteorologia i Gidrologia (8):21-8.
126. WALLIS, J. R. 1965. Multivariate statistical methods in hydrology-a comparjson using
data of known functional relationships. Water Resniarces Res., 1í4):447-61.
127. WERNER, P. W.; SUNDQUIST, K. J. 1951. On the groundwater recession curve for large
I
watersheds. IASH publ. no. 32. Brussels.
I
128. WESSELING, J. 1959. The transmission of tidal waves in elastic artesian basins. Neth.J.
Agric., 7 :22-32.
I 129. -__. 1?64.Review cf meth~drused in the hydrological analysis of an agricultural
I area. Landboiiwk. Tijdschr.,76 :82741.
130. WHIPKEY, R. Z. 1965. Subsurface stormfow from forested slopes. IASH Bull.,no.
lO(2).
131. WHISLER, F. D.; KLUTE, A. 1966. Analysis of infiltration into stratified soil columns.
Unesco symp. on water in the unsaturated zone, paper IIa-3.Wageningen.
345
Representairve and experimental basins
132. WICHT, C. L. 1965. The validity of conclusions from South African multiple watershed
experiments. Znt. symp. on forest hydrol. Pennsylvania State University.
133. WIENER, N. 1958. Nonlinear problems in random theory. New York, Wiley.
134. WILM, H.G.1949. H o w long should experimental watersheds be calibrated? Trans. Amer.
Geophys. Un.,30(2) :272-8.
135. WISLER, C. O.;BRATER,E. E. 1949. Hyduology. New York, Wiley.
136. WORLD METEOROLOGICAL ORGANIZATION. 1965. Guide to hydrometeorological practices.
(WMO no. 168, TP. 82.)
137. YOUNGS, E. G.1958. Redistribution of moisture in porous materials after infiltration, I.
Soil Sci., 86: 117-25.
138. ___ . 1958. Redistribution of moisture in porous materials after infiltration, II.
Soil Sci., 86 :202-7.
139. ANON. 1967. Vodnye resursy i vodny balans territorii Sovetskogo Sojuzu [Water resources
and water balance of the U.S.S.R.territory]. Leningrad, Gidrometeoizdat.
346
Bibliography
O.A. 1953. Osnovy gidrokhimii [Principles of hydrochemistry], 2nd ed. Leningrad,

ALEKIN,
Gidrometeoizdat.
BEAR,J. et al. 1968. Physical principles of water percolation and seepage. Paris, Unesco.
(Arid zone research, XXIX.)
BUDYKO,M.I. 1959. Teplovoi balans zemnoi poverkhnosti [Heat balance of the earth's surface].
A.I. 1960.Obschaja gidrologia/vodysushi [General hydrology/continentalwaters].
CHEMOTAREV,
CHOW,VEN TE(ed.). 1964. Handbook of applied hydrology. N e w York, McGraw-Hill.
DAVIS, S. N.;DEWIEST, R. J. M. 1966. Hydrogeology. N e w York, Wiley.
DEWIEST, R. J. M. 1965. Geohydrology. N e w York, Wiley.
INTERNATIONALATOMIC ENERGY AGENCY. 1963.Radioisotopes in hydrology.Proc.Symp.Tokyo,
5-9 March. Vienna.
IVANOV, K.E.1957. Osnovy gidrologii bold lesnoi zony [Principles of swamp hydrology in the
forest zone]. Leningrad, Gidrometeoizdat.
KONSTANTINOV, A. R. 1963. Isparenie v prirode [Evaporation in nature]. Leningrad,Gidro-
meteoizdat.
KUDELIN, B. I. (ed.). 1966.Podzemny stok na ferritorii SSSR [Subsurface flow in the U.S.S.R.].
Moscow, Moscow University.
KUZMIN, P.P. 1961.Protsessy iajania snezhnogopokrova [Snow-packmelt processes]. Leningrad,
Gidrometeoizdat.
LEBEDEV, A.V. 1963. Melody izirchenia balansa grunfovykh voà [Methods of studying ground-
1 water balance]. Moscow, Gosgeolizdat.
LJNSLEY, R. K.,et al. 1949. Applied hydrology. N e w York, McGraw-Hill.
MOLCHANOV, A. A. 1960. Gidrologicheskaja rol lesa [Hydrological role of forests], 2nd ed.
Moscow, Academy of Sciences of the U.S.S.R. (Translated into English by Israel Program
for Scientific Translations.)
OURYVAEV, V.A. 1953. Experimenialnye gidrologicheskie issledovania na Valdae [Experimental
hydrological investigations in Valdai]. Leningrad, Gidrometeoizdat.
RESEARCH INSTITUTE FOR WATER RESOURCES. 1966. Hydrological methods for developing water
resources management. Series of manuals for International Post-Graduate Course, ed. by
K.Ubell, Budapest.
Subject 1, Hydrological bases of water nianagemerit.
Subject 2, Sirrface water liuvdrology.
Subject 3, H J ~ v ufo groiiiidwc?cT.
~o~~
Subject 4, Soil physics.
Subject 5, Qualitative characteristics of water resources.
Subject 6, Mathematical staiistics as a method for hydrological investigations.
Subject 7, The hydrologic cycle and the water balance in nature.
Subject 8, Hydrological .forecasting.
347
Subject 9,Estimation of surface-water resources.

Subject 10,Estimation of ground-writer resources.
Subject 11, Hydrology of water storage.
Subject 12, Sediment transportation in alluvial streams.
Subject 13, Effect of man’s influence on hydrological phenomena.
RODE, A. A. 1965.Osnovy uchenia o pochvennei vlage.T o m I, Vodnye svoistva pochv: pered-
vizhenia pochvennoi vlage [Principles of soil-moisturestudy. Vol. 1, Soil physical properties
and soil-moisture transfer]. Moscow, Academy of Sciences of the U.S.S.R.
SHAMNOV, G.I. 1954.Rechnye nanesy: rezhim, raschety i metody izmereniy [River sediments:
regimen, computation and methods of measurement]. Leningrad, Gidrometeoizdat.
SILIN-BEKCHURIN, A.T. 1965.Dinamika podzemnykh vodjs osnovami gidravliki [Underground
water dynamics/with hydraulic principles], 2nd ed. Moscow, Moscow University.
SOILCONSERVATION AND RIVERS CONTROL COUNCIL. 1963-.Handbook of hydrologicalprocedures.
Procedures 1- . N e w Zealand.
____ . 1967-.Annual hydrological research reports. For IHD Experimental Basin Project,
N e w Zealand.
SOKOLOVSKY, D.L. 1959.Rechnoy stok [River run-off], 2nd ed. Leningrad, Gidrometeoizdat.
SOPPER,W.E.;LULL,H.W.(ed.). 1967.International symposium on ,foresthydrology. Proc.
Nat. Sci. Foundation Advanced Sci. Seminar, Pennsylvania State Univ.,Aug. 29-Sept. 10,
1965.
UNITED STATESDEPARTMENT OF AGRICULTURE. 1962.Field manual for research in agricultural
hydrology. Agric. Res. Serv. (Agric. handbook no. 224.)
. 1963.Proceedings of the federal interagency sedimentation conference. Agric. Res.
Serv. (Misc. publ. no. 970.)
____ . 1963.North Appalachian experimental watershed, Coshocton, Ohio. Annual reports.
Agric. Res. Serv.
VELIKANOV, M.A.1964.Gidrologia sushi [Land hydrology],5th ed.Leningrad,Gidrometeoizdat.
VOSKRESENSKY, K.P. 1962.Norma i izmenchivost godovogo stoka rek Sovetskogo sojuza [Normal
annual discharge and its variations for U.S.S.R. rivers]. Leningrad, Gidrometeoizdat.
WECHMANN, A. 1963.Hydrologie. Berlin, VEB Verlag für Bauwesen.
WUST,G.1951.Die Kreislaufe des Wassers auf der Erde. Kiel, Karl-Gripp-Festschr.
ANON.1957.Rukovodstvo PO proizvodst nabbdenig nad ìspareniem s selskohoziajstvennykhPolei.
Chast 1, Nabludenija nad ispareniem metodom pochvennykh ispariteley. Chast 2, Nablu-
denija nad ispareniem gradientnym metodom. [Guide for evaporation practices on agricultural
fields.Part 1, Evaporation measurements by tank evaporimeters.Part 2,Evaporation measure-
ments by the gradient method]. Leningrad, Gidrometeoizdat.
. 1954. Rukovodstvo stokovym stantsijam [Guide for run-off stations]. Leningrad,
Gidrometeoizdat.
_-___ .1967.Vodoye resursy i vodny balans territoriiSovetskogo Sojusa [Water resources and
water balance of the U.S.S.R.]. Leningrad, Gidrometeoizdat.
-~ . 1965.Voprosy experimentalnogo izuchenia stoka: vodnogo balansa rechnykh vodos-
borov [Problems in the experimental study of run-off and water balance of river basins].
Materialy soveschania (4-7 August 1964). Valdai, GGI.
348

Representative and Experimental Basins.: Unesco

Uploaded by

Copyright:

Available Formats

Representative and Experimental Basins.: Unesco

Uploaded by

Document Information

Original Description:

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

Representative and Experimental Basins.: Unesco

Uploaded by

Copyright:

Available Formats

Representative and

Edited by C.Toebes and V. Ouryvaev

A contribution to the International Hydrological Decade

Printed by Henkes-Holland, Haarlem

The International Hydrological Decade (IHD)1965-74 was launched by the General

2 Selection and organization of basin networks

2.2.2.I Example in the U.S.S.R. 48

3 Planning of observations according to the

I 4 Methods of observation and instrumentation

4.2.4.2.2.4 The non-weighable lysimeter 108

4.3.8 Run-off plots 137

4.9.3.1 Techniques of sediment measurement 172

5 Data processing and publication

5.1.2.1.2 Features 200

5.5 Subsurface water 229

6 Analysis techniques and interpretation of research results

6.1.2.2.1 Precipitation 273

6.2.5.4 Non-linear functional models (general non-linear analysis) 320

ALLENJr., W.H.,United States Department of Agriculture, Agricultural Research Service,

OURYVAEV,V. A., State Hydrological Institute, U.S.S.R.

1.1 Scope and purpose

1.2 Definitionof representative and experimental basins

Representative basins are basins which are selected as representative of a hydrological

I 1.2.2 Experimental basins

1.3 Purposes of representative and experimental basins

1.3.2 Experimental basins

1.4 Survey of research needs

1.4.1 Stafing for basins

1.4.2 Basic recommendations on standardization of methods

Minimum 1 hydrological 1 hydrological 2 hydrological

1.4.3 Data-reporting methods

1.4.4 Research co-operation and research observation

Where variousprojects are involved,these should be correlatedto ensure that maximum

tame of region Yamagata

ieology : hasement complex ( t d f , schistose tuff)

1 raingauge and 1 V-not& weir equipped vi:

FIG.1.1. Basin card for Shione Experimental Basin, Japan [13].

1939 - basin established

Dota col1ecte.d:daily precipitation anü f'iovr

Research into methods of' maiiagei-neiit of head-

Report on observations made at experirnontal

Hydrological model. Mathematical formulation to simulate natural hydrological phenomena

Albedo. Theproportion of radiation fallingupon a non-luminousbody which is diffusely reflected

SURFACE WATER TERMS

SUBSURFACE WATER TERMS (see Figs. 1.2, 1.3 and 1.4)

SUSPENDED INTER- ZONE

GROUND .'.:.I..::'.. .. ZONE OF

FIG.1.3. Subsurface water zones.

FIG.1.4. Relationship between moisture potential and moisture content.

of the land surface as in soil creep,landslide or slip.Mass movement may be classified in a

1.6 Measurement units and symbols

30. POPOV, O.V.; KUZNETSOV, V. I. 1960. Experimentalnye issledovania elementov vodnogo

2.1 Use of maps

2.2 Selection of hydrological regions

aspects. It should be noted, however, that principles of selection of physiographical

I 2.2.1 Selection of hydrological regions in areas where no

2.2.1.1 Example in Brazil

2.2.1.2 Example in New Zealand