WORLD METEOROLOGICAL ORGANIZATION

LECTURE NOTES FOR

TRAINING AGRICULTURAL
METEOROLOGICAL
PERSONNEL

Prepared by
Jon WIERINGA
Jakob LOMAS

SECOND EDITION

WMO-No. 551
Secretariat of the World Meteorological Organization
Geneva - Switzerland
2001
© 200 l , World Meteordogical Organization

ISBN 978-92-63- 10551-6

NOTE
The designations employed and the presentation of material in this publication do not imply
the expression of any opinion whatsoever on the part of the Secretariat of the World
Meteorological Organization concerning the leg al status of any country, territory, city or
area, or of its authorities, or concerning the delimitotion of its frontiers or boundaries.
TABLE OF CONTENTS

FOREWORD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
PREFACE.......................................................... ix

CHAPTER 1: AGRICULTURAL METEOROLOGY - ITS SCOPE AND AIMS . . . . . . . 1

1.1 Aims . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Range of subject matter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2.1 Soil and water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2.2 Plants and crop microclimate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2.3 Farm animals (farm livestock) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2.4 Diseases and pests of crops and animals . . . . . . . . . . . . . . . . . . . . . . . _2
1.2.5 Farm buildings, equipment and operations . . . . . . . . . . . . . . . . . . . . 3·
1.2.6 Artificial modification of meteorological regimes . . . . . . . . . . . . . . . . 3
1.2.7 Climate change . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 Use and provision of agrometeorological information . . . . . . . . . . . . . . . . . . 4
Review questions on Chapter 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

CHAPTER 2: RADIATION AND THE SURFACE ENERGY BALANCE . . . . . . . . . . . . . 7

2.1 Solar energy ('short-wave' energy) from Sun and sky . . . . . . . . . . . . . . . . . . 7
2.1.1 Direct and diffuse components of solar short-wave radiation . . . . . . . 7
2.1.2 Estimation of global radiation on a horizontal surface . . . . . . . . . . . . 9
2.1.3 Emission and reflection of radiation . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.1.4 Energy in the visible spectrum - light . . . . . . . . . . . . . . . . . . . . . . . 11
2.2 The energy balance and its components . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.2.1 The long-wave budget . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.2.2 Surface radiation temperatures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.2.3 Total radiation budget and complete surface energy balance . . . . . . . 14
2.3 Special aspects of radiation and temperature in agriculture . . . . . . . . . . . . . . 16
Review questions on Chapter 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

CHAPTER 3: THE SOIL AND ITS HEAT BALANCE.......................... 21

3.1 What is 'soil'? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.2 Transmission of heat in the soil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.3 Soil freezing, and the role of snow cover . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.4 Diurnal and annual variations of soil temperature and moisture. . . . . . . . . . 25
3.5 Amodel of soil temperature diurnal course at different depths . . . . . . . . . . . 27
Review questions on Chapter 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

CHAPTER 4: WATER AND THE HYDROLOGICAL CYCLE IN AGRICULTURE . . . . . 31

4.1 Water and vegetation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
4.2 Moisture characteristics of soils . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
4.3 Determination of water loss from land surfaces . . . . . . . . . . . . . . . . . . . . . . . 35
4.3.1 Fundamentals of the evaporation process . . . . . . . . . . . . . . . . . . . . . 35
4.3.2 Existing methods to determine evaporation . . . . . . . . . . . . . . . . . . . . 37
4.3.3 Energy balance estimation of evaporation . . . . . . . . . . . . . . . . . . . . . 38
4.3.4 Aerodynamic estimation of evaporation . . . . . . . . . . . . . . . . . . . . . . 40
4.4 'Combination' methods of Penman and others . . . . . . . . . . . . . . . . . . . . . . . 42
4.4.1 Development of the original Penman equation . . . . . . . . . . . . . . . . . 42
4.4.2 Evaporation formulae of Priestley-Taylor and Penman-Monteith . . . . 44
4.5 Special forms of precipitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
4.5.l Dew . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
4.5.2 Snow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
4.6 Soil moisture budgets - irrigation need . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
Review questions on Chapter 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

iii
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

CHAPTER 5: SMALL-SCALE CLIMATE, REPRESENTATIVITY, AND THEIR

DEPENDENCE ON TOPOGRAPHY . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
5.1 Micro-, topo- and meso-climatology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
5.2 Observation representativity, exposure and sampling . . . . . . . . . . . . . . . . . . 55
5.3 Wind behaviour in common inhomogeneous terrain . . . . . . . . . . . . . . . . . . 57
5.3.1 Wind around barriers of varying porosity . . . . . . . . . . . . . . . . . . . . . 57
5.3.2 Wind reduction by shelterbelt arrays . . . . . . . . . . . . . . . . . . . . . . . . . 60
5.3.3 Wind representativity at toposcale . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
5.4 Toposcale representativity of meteorological observations . . . . . . . . . . . . . . . 66
5.5 Topoclimatologi(al effects a'rising from landscape variations . . . . . . . . . . . . . 70
5.5.1 Effects of slopes on incoming solar radiation . . . . . . . . . . . . . . . . . . . 70
5.5.2 Soil temperatures on slopes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
5.5.3 Effects of slopes and hills.on airflow . . . . . . . . . . . . . . . . . . . . . . . . . 72
5.5.4 Local mesoscale circulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
Review questions on Chapter 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

Chapter 6: AGROMETEOROLOGICAL MANAGEMENT AT MICROSCALE AND

TOPOSCALE ... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
6.2 Soil cultivation and·treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
.6.2.1 Effects of surface-colouron soil temperature . . . . . . . . . . . . . . . . . . . -80
·6.2.2 Mulching ._. ... ... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... . . . . . •8.1
6.2.3 Surface geometry effects on temperatures . . . . . . . . . . . . . . . . . . . . . ·83
,6.3 ·, Crop management and layout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
~6.3.l Spacing of.crop rows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .84
6.3.2 Shading,-cover crops and weeding . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
6.4 r-Wind shelter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -88
6.4.1 Effect of wind-shelter on microclimate . . . . . . . . . . . . . . . . . . . . . . . . 88
6.4.2 Windbreaks.against damage and erosion . . . . . . . . . . . . . . . . . . . . . . ·89
6.5 Jrrigation and drainage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
6;6 Frost, and protectioffagainst frost damage . . . . . . . . . . . . . . . . . . . . . . . . . . -93
6.6.1 Passive methods of protection against frost . . . . . . . . . . . . . . . . . . . . 93
6.6.2 Active methods of .protection against frost . . . . . . . . . . . . . . . . . . . . . 95
6.6.3 Short-term frost forecasting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
6. 7 Artificial climate in _glasshouses and stables_. . . . . . . . . . . . . . . . . . . . . . . . . . 98
·Review questions on Chapter 6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102

·.CHAPTER 7: WEATHER·HJ\ZARDS.ADVERSELY AFFECTING AGRICULTURAL

O.UTPUT ... ... .... ... ... : . . . . . . . . .. . . . . . . . . . . . . . . . . ... . . . . . . . .107
.7.-1 Drought . . . . . . ... . ..... . . . . . . . . . . . .. . . . . . . . . . ... . . . . . . . . . . . . . . . . . . . . 107
7.2 Artificial stimulation-of.precipitation . . . . ... . . . . . . . . . . . . . . . . . . . . . . . . 110
·7:3 ·ttail ........... .• ..,. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
7.3.1 Distribution- of hail in space and time . . . . . . . . . . . . . . . . . . . . . . . . 112
7.3.2 Active suppression of hail . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
7.4 Fire in vegetation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
7.5 Atmospheric.transports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
7.5.1 Elementary aspects of transport over mesoscale distances . . . . . . . . . 116
7.5.2 Point sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
_ 7.5.3 Line sources, e.g. sea-salt tran~port . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
·Review questions on Chapter 7 . ... : .. . ....... . .. .. . . . . . . . . . . . . . . . . . . . . . 120

CHAPTER"S: OPERATIONAL AGROMETEOROLOGY . . . . . . . . . . . . . . . . . . . . . . . 123

8.1 Alternative forms of agrometeorological decision-supporting activity . . . . . . 123
8.2 Operational modelling for tactica_lagrometeorology . . . . . . . . . . . . . . . . . . . 126
8.3 Protectidn of crops against pes~s and diseases . . . . . . . . . . . . . . . . . . . . . . . . 128
8.4 Agroclimatological surveys ....·....... . ·. . . . . . . . . . . . . . . . . . . . . . . . . . . 130

iv
TABLE OF CONTENTS

8.4.1 Agroclimatology of the Sahel - an example of presentation . . . . . . . 131

8.4.2 Irrigation need - a climatological case study . . . . . . . . . . . . . . . . . . 132
8.4.3 Agroclimatological analysis of a rainfed semi-arid situation . . . . . . . . 134
8.5 Computer weather modelling for agriculture . . . . . . . . . . . . . . . . . . . . . . . . . 134
8.6 Modelling of heat stress for avocado - a case study . . . . . . . . . . . . . . . . . . . 136
8. 7 Agrometeorological weather and climate information . . . . . . . . . . . . . . . . . . 138
Review questions on Chapter 8 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139

CHAPTER 9: AGROMETEOROLOGICAL INSTRUMENTS AND OBSERVATION . . . . 143

9.1 Basic observation rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
9.1.1 Agrometeorological networks and documentation . . . . . . . . . . . . . . . 143
9.1.2 Dynamic responses of meteorological instruments . . . . . . . . . . . . . . . 144
9.2 Agrometeorological instruments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
9.2.l Air temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
9.2.2 Grass-minimum temperature and radiative surface temperature . . . . 147
9.2.3 Soil temperature and soil heat flux . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
9.2.4 Wind . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
9.2.5 Radiation and sunshine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
9.2.6 Humidity, dew and leaf wetness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
9.2.7 Evaporation and evapotranspiration . . . . . . . . . . . . . . . . . . . . . . . . . 154
9.3 Observations of 'state of the ground' and soil moisture.. . . . . . . . . . . . . . . . 156
9.3.l 'State of the ground' . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
9.3.2 Soil moisture ..... , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
9.4 Biological observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
9.4.1 Observations for research and for operational use . . . . . . . . . . . . . . . 165
9.4.2 Observations of natural phenomena for agroclimatological use . . . . . 166
9.4.3 Specific examples of biological/phenological observations . . . . . . . . . 166
9.4.3.1 Wheat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
9.4.3.2 Maize . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
9.4.3.3 Avocado . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
9.5 Remarks on experimental procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
Review questions on Chapter 9 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170

Appendix I: NOTES ON THE STATISTICAL ANALYSIS OF METEOROLOGICAL

DATA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
Appendix II: UNITS, CONVERSION FACTORS, USEFUL CONSTANTS . . . . . . . . . . . 177
Appendix III: ESTIMATION OF SOLAR ELEVATION AND DAY LENGTH . . . . . . . . . 179
Appendix IV: RELEVANT WMO PUBLICATIONS, AND AGROMETEOROLOGICAL
CORE LIBRARY.. .. ...................................... 181

REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185

V
FOREWORD

The Commission for Agricultural Meteorology has, on several occasions, stated

that training and education in agricultural meteorology should be a matter of
great importance to Members, particularly in developing countries. In 1980, the
Lecture Notes for Training Class II and Class III Agricultural Meteorology Personnel
(WMO-No. 551), prepared by R. W. Gloyne (United Kingdom) and J. Lomas
(Israel) were published in the WMO Training Series, translated into Spanish, and
used in national and regional training centres.
The need arose to update such useful material and Professor J. Wieringa of
Wageningen Agricultural University in the Netherlands agreed to review and
update the previous edition of the lecture notes with the assistance of Mr Lomas.
I wish to express to both authors my sincere appreciation and gratitude for
their excellent work. The title of the publication has been adjusted to take into
account the new WMO classification, which was approved by the Executive
Council at its fiftieth session in June 1998, and came into effect on 1 January
2001.
I feel confident that the material contained in this publication will prove to
be invaluable not only to students but also to instructors.

(G.O.P. Obasi)
Secretary-General

vii
PREFACE

This Compendium of Lecture Notes on Agricultural Meteorology has been developed

from an earlier WMO training publication (Blue series), which was based on a draft
by J. Lomas and completed and finalized by R. W. Gloyne in 1980. Because of exten-
sive development of small-scale meteorology since the 1970s, it proved necessary to
rewrite and rearrange most of the text. Only the sections in Chapter 9 on observa-
tions of evaporation, soil water and biological phenomena are largely unaltered.
The 1980 edition (WMO-No. 551) sought to be accessible to WMO Class III
and Class II personnel with an assumed full secondary school education in mathe-
matics and basic sciences, who were to be trained in meteorology in order to
manage and develop weather-related operations and to dispense appropriate infor-
mation to users. These revised notes aim at providing operational meteorological
personnel (Class technicians under the newly-adopted WMO system of classifica-
tion) with basic knowledge about aspects of meteorology which are important to
agriculture, so that they can provide meteorological support to farmers and agron-
omists. For specialization in agrometeorology, additional studies in agricultural and
biological sciences would be required, as well as advanced study of small-scale
meteorological subjects like evaporation and soil energy balance.
The assumed meteorological knowledge level of users of the Compendium is
summarized in the three WMO compendia listed below. As they are frequently
referred to in these notes, it was judged practical to refer to them by abbreviations.
In addition, one non-WMO publication contains so much basic agrometeorology
information that an abbreviation was also necessary. These four basic publications
are:
Abbreviation:
(a) Compendium of Lecture Notes for Training Class IV
Meteorological Personnel, B. J. Retallack, Vol. II -
Meteorology (WMO-No. 266, 2nct edition, 1984) GM-II
(b) Compendium of Lecture Notes in Climatology for Class III
and Class IV Personnel, B. J. Garnier (WMO-No. 726, 1992) CL-G
(c) Compendium of Lecture Notes for Training Class IV
Agrometeorological Personnel, A. V. Todorov
(WMO-No. 593, 1980) AG-T
(d) The Climate near the Ground, R. Geiger. This "bible of
micrometeorology" has been republished in entirely revised
German and English editions on several occasions since 1927.
The 5th edition, 1995 (Vieweg), revised by R. H. Aron and
P. Todhunter, maintains the section sequence of the 4th
edition (Vieweg, 1961 and Harvard University Press, 1965), but
omits Section 47 of the 4th edition so that section numbers
4 7 and subsequent sections differ by one. To make these two
editions equally usable, they are here both referred to
(without their year of publication) as: Geiger (section ##)

In references to Geiger's figures and tables, the notation of the 1995 edition
is used, and the section number is given first (e.g. Figure/Table ##-1).
Many section headings refer to relevant WMO Technical Notes, Handbooks
or other review publications which should be accessible without excessive trouble.
A major difference between agrometeorology and most other applied
branches of meteorology is the importance of small-scale exchange of various
forms of energy. Physical climatology deals with radiation and evaporation (see
CL-G), but usually only as long-term regional averages. Agrometeorology has a
strong practical interest in all boundary-layer processes from mesoscale to
microscale (see section 5.1).
ix
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

Therefore boundary-layer physics and surface-layer energy balance are dealt

with in Chapters 2 to 4 in the discussion on solar radiation, various transports in
the upper soil layer, and the water balance at the surface. A fourth subject, atmos-
pheric motion, is introduced in GM-II and CL-G only in a large scale context. As
wind provides transport for surface energy exchanges, it is necessary to discuss in
Chapter 5 small scale wind dependence on local terrain, and also terrain-depend-
ence and representativity of other properties at various scales in various types of
terrain, e.g. mountainous country.
The range of possibilities to manage the small scale agricultural climate is
presented in Chapter 6 which deals with treatments of the soil, and with handling
and sheltering crops to optimize crop development in given weather. In addition,
Chapter 7 reviews the weather hazards threatening crops, such as drought, hail,
fire and pollution. Chapter 8 discusses general aspects of agrometeorological oper-
ations, such as forecasting, and gives examples of agrometeorological modelling
approaches. Case studies of local crop weather problems can be better dealt with
by local teachers.
GM-II, and most other basic meteorology books, cover meteorological instru-
mentation and observation only by reviewing classical station instruments.
Operational agrometeorology needs observations which are often not available at
weather stations, and these are discussed in Chapter 9. A list of references and
recommended publications, unit conversion factors, etc. are given in the
Appendices.
In the earlier 1980 edition the authors had managed to locate or devise ques-
tions in the style of W. P. Lowry, which are more than mere exercises on the
preceding text. Many of them were drawn from, or at least utilize data from, actual
investigations from a wide range of climatic regimes. Many of these questions
have been used in this publication, with the only change being that all quantita-
tive data in the questions, as well as in the text, were changed into decimal units.
On subjects which were not covered in the earlier edition, some new questions
have been added. The questions are grouped into three categories:
• Relatively straightforward questions based on the preceding text;
• Those which are more difficult and/or involve wider consideration;
• Questions which are really small research projects, requiring field work, or
access to and use of local climatological data.
The authors are grateful for the cooperation of the following colleagues in
reviewing compendium texts and suggesting amendments: Herman Boyen, Hans
Friesland, Per-Erik.Jansson, Kees Stigter, Bert Wartena and Gaetano Zipoli. The support of
the COST-711 Action of the European Community, of Wageningen Agricultural
_University and of the IATA Institute in Firenze is gratefully acknowledged.

Jon Wieringa Jakob Lomas

Wageningen Agricultural University WMO/RMTC for Applied Meteorology
The Netherlands Bet Dagan, Israel
CHAPTER 1
AGRICULTURAL METEOROLOGY - ITS SCOPE AND AIMS

1.1 AIMS The primary aim of agricultural meteorology is to extend and fully utilize our
knowledge of atmospheric and related processes in order to optimize sustainable
agricultural production with maximum use of weather resources and with
minimal damage to the environment. This entails improving the quantity and
quality of agricultural crops, timber and other forest products (e.g. natural rubber),
vegetable fibres (e.g. cotton, flax, sisal), and animal products and by-products
(e.g. hides).
A secondary aim concerns the conservation of natural resources. The climate
may place constraints upon a particular form of land-use at a given place and time.
An agricultural meteorologist should be (but often will not be) consulted when
questions are being examined of land-use, of the exploitation of resources and of
the deployment of technological processes. For example, the short-term benefit
from the cultivation of semi-arid grassland has often (e.g. in the former Soviet
Union) been gained at the expense of long-term damage from erosion by wind
and water. Meteorologists have the advantage of being able to take into consider-
ation processes on very different time-scales.

1.2 RANGE OF The subject includes:

SUBJECT MATTER (a) The Earth (physical) sciences - specifically the physics of the atmosphere,
i.e. meteorology and climatology, but also soil physics and hydrology;
(b) Certain biological sciences - specifically physiology, ecology and pathology
of plants and animals, and associated 'technologies' of agriculture.
A more detailed statement is given in the opening paragraphs of the WMO
Guide to Agricultural Meteorological Practices (WMO-No. 134):
"Agricultural meteorology is concerned with interaction between meteoro-
logical and hydrological factors, on the one hand, and agriculture in the
widest sense, including horticulture, animal husbandry and forestry, on the
other hand. Its object is to discover and define such effects, and thus to apply
knowledge of the atmosphere to practical agricultural use. Its field of inter-
est extends from the soil layer of deepest plant and tree roots, through the
air layer near the ground in which crops and woods grow and animals live,
to the highest levels of interest to aerobiology, the latter with particular refer-
ence to the effective transport of seeds, spores, pollen and insects.
In addition to natural climate, and its local variations, agricultural meteor-
ology is also concerned with artificial modifications in environment (as
brought about, for example, by windbreaks, soil management, irrigation,
glasshouses, etc.); in climatic conditions of storage, whether indoors or in
field heaps; in environmental conditions, in animal shelters and farm build-
ings; and during the transport of agricultural produce by land, sea or air."
This compendium considers the relationship between weather and agricul-
ture, and touches upon subjects that may be conveniently classified under seven
main headings: soil and water, plants, farm animals, diseases and pests of crops
and animals, farm buildings and equipment, farm operations, and artificial modi-
fication of the meteorological and hydrological regime. Each of these will be
briefly discussed in this chapter.

1.2.1 Soil and water Weathering is an important factor in creating and then determining the nature of
a soil. Actual and historical weather affect the mechanical, physical and chemical
properties of the soil, the organisms it contains and its capacity for retaining and
releasing heat, nutrients and moisture. Rainfall not only adds chemical compo-
nents to the soil, but it also washes out ('leaches') soil nutrients.
The mechanical state of the soil - as it affects the cultivation, pest control
and harvesting of crop plants, management of pastures (stocking density, etc.) is
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

greatly influenced by local weather conditions. In these matters the cooperation

of environmental scientists is vitally important: meteorologists must not simplify
too much the surface aspects of their work, while on the other hand hydrologists
should not assume full knowledge of all weather and climate factors. All in all,
cooperation between meteorologists, hydrologists and soil scientists needs to be
improved.
The water reservoir for plant growth and development is contained in the
soil. The amount of water available in soil depends on the effectiveness of precip-
itation or irrigation, on the one hand,· and on soil's physical properties and depth
on the other hand. The rate of water loss from the soil depends on the climate,
and on the physical properties of the soil and the root system of the plant
community. Efficient utilization of soil water and cultivation techniques for
preservation of soil moisture are of major concern to agrometeorology, especially
in semi-arid areas.
Erosion by wind and water depends on regional and local weather factors.
The extent to which a given tract of land suffers from erosive agencies is, more-
over, determined by the presence and vigour of the vegetative cover. Excessive
concentration of salt at the soil surface may occur in areas especially where there
is significant evaporation and drainage is not adequate, particularly when irriga-
tion water is of poor quality.
In all regions with a .marked seasonal variation in weather, seasonal changes
influence soil conditions and hence the farrrii:qg1programme ('calendar') - e.g. the
beginning and end of rainy seasons in (sub)tropical countries; in wet climates, the
period during which the water content of the upper soil layers exceeds the 'field
capacity' (see section 4.2); or the duration of frozen ground in high latitudes.

1.2.2 Plants and crop -Development and growth of plants depend on environmental conditions at every
microclimate stage (Woodward, 1987). An understanding of the interrelation between the struc-
ture of the environment (ground cover, surface slope, degree of shelter, etc.) and
the local microclimate, in the crop and around the crop, may result in actions
aimed at the long-term improvement of the growth situation.
Even before planting the influence of the weather should be considered. The
quality of the seed sown depends on meteorological conditions during the year in
which it was produced. The productivity of long-rotation crops, e.g. vines, fruit
and forest trees, can also be affected by weather experienced over many previous
seasons.
Post-harvesting operations, such as drying grass and other crops, and the
capacity to maintain the quality of stored farm crops are affected by seasonal
weather. Weather and dimate are important in the occurrence of forest, bush and
·grass fires, and knowledge of them is important for the defence against such
·hazards.

1.2.3 Farm animals Weather affects animals in all stag~s of growth and condition, regardless of
(farm livestock) whether they are well-fed and healthy, or ill-fed and diseased. Primarily it acts
directly on their feeding, growth, health and offspring production, e.g. by drought
or through the effect of extremely high or low temperatures (Mount, 1979). In
addition, weather influences livestock through their food supply (grass, crops) and
the state of soils on which they are kept. Climate determines geographical distri-
butions of animals. Yield and quality of animal products, and their processing and
storage and transport are also weather-dependent. Keane (1986) deals at length
with many aspects of weather influences on animals.

1.2.4 Diseases and pests of Weather influences the degree to which plants and animals (i.e. the 'victim'
crops and animals and/or -'host') are attacked by pests and diseases, or harbour them. It also affects
the biology of insects and disease organisms, and determines the nature, numbers
and activity of pests (and of the predators on pests), and extent and virulence of
diseases. In crop and livestock protection, the spread and aerial transport of
pests and diseases, and the effectiveness of applied control or eradication
methods depend upon atmospheric agencies. Examples of this include: the

2
CHAPTER 1 - AGRICULTURAL METEOROLOGY - ITS SCOPE AND AIMS

wind-dependent distribution of chemical sprays; the effects of temperature on the

chemical activity of pesticides, or on the vulnerability of pests to such treatments;
the dependence of fungi on the presence or absence of liquid water is climate-
sensitive, be it rainfall or dew.

1.2.5 Farm buildings, equip- Climatic conditions must be taken into account in the planning of farm buildings
ment and operations and, particularly, in the design of animal housing and stores for agricultural
produce. The choice of farm machinery, harvesting technique and the deploy-
ment of labour is also climate sensitive (Dalton, 1974). One important aspect is
the climatological probability that the soil will be sufficiently strong to support
machinery at the time it is needed (see Keane, 1986 and section 9.3.1).
The farmer's best choice of work priorities depends on relevant forecasts
sufficiently far ahead - even plans for a single day call for forecasts for several
days; and those forecasts must include some parameters not normally included in
public forecasts.

1.2. 6 Artificial modification Irrigation, windbreaks, shelterbelts, mulches, storage and conservation of snow
of meteorological regimes and water, as well as soil cultural practices, have an important influence on certain
aspects of local environment, such as soil moisture, wind velocity and atmos-
pheric humidity.
The greatest degree of control of environmental conditions is exercized in
the use of glasshouses and in intensive animal housing. Varying degrees of inde-
pendence from external environmental conditions can be achieved (at times to a
very high degree). However, the cost of the materials and the energy needed to
achieve this independence tend to increase as the difference increases between
desired internal climate conditions and actual weather and climate outside the
structure. Meteorological support in the design and management of such artificial
climates plays an important part in keeping their cost-benefit ratios economically
acceptable.

1.2.7 Climate change There is considerable research on the agricultural effects - both qualitative and
quantitative - of climate change aided by or triggered by man's activity (e.g.
changes in carbon dioxide content of the atmosphere due to the consumption of
fossil fuels, such as coal and oil, or the production of methane by animals and
vegetation). However, this compendium will not discuss the question of climate
change at any length for several reasons. The first is, that climate change is likely
to be a rather slow process, so agriculture might often manage to adapt to envi-
ronmental changes known in advance. The second is, that processes related to
climate change are still subject to much research, so that explicit statements on
their consequences might have to be reappraised within a few years. In particular,
global temperature changes can lead to climate changes of varying sign and
magnitude at regional or local scales.
The major reason for not tackling the issue of 'climate change' as a separate
item here is that, in an agrometeorological context, it is a complex issue. Many
investigators deal with the subject exclusively as the influence of increase in
average temperature (daytime, night-time or 24 hours? seasonal or annual?) on
crop growth; they often seem not to be aware that such an increase could involve
many other weather changes, which are not considered in their model.
For example - higher temperature could mean increased evaporation (not
always, see section 4.3.1), with possible increase in aridity or increase in cloudi-
ness - and the latter would decrease incoming radiation and might (or might not)
increase precipitation. Higher temperatures combined with sufficient radiation
would increase photosynthesis, but the simultaneous increase of metabolism
might shorten growth seasons, leading maybe to a net decrease in production. In
temperate latitudes in spring, earlier high temperatures might mean early flower-
ing, at a time when the risk of frost is still very high. There are many more such
meteorological side effects (Fajer and Bazzaz, 1992).
Briefly, the complexity of climate change makes it an undesirable subject for
an introductory agrometeorological text. For this very reason, climate change

3
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

problems should be investigated by agronomists and professional meteorologists

jointly in order to prevent inadmissible simplifications, either of the weather or of
biological processes - and variability should be considered as important a subject
as possible change.
Figure 1.1 illustrates some factors that can have adverse consequences on
agricultural production. Disasters always arouse more comment than successes -
but many potential disasters can, to a significant extent, be avoided, or at least
their impacts reduced by human action. It is interesting to realize that all factors
shown are weather-sensitive! Much agrometeorological action is directed to fight-
ing such negative weather effects. However, insight into the effects of weather and
climate on vegetation and livestock can also have purely positive effects, thereby
increasing the success of agricultural activities in unexpected ways.
Figure 1.1
The effect of various factors
limiting crop production (after
American Potash Institute).

t
Diseases

~.,t.--:«11
,,
~~:•:•

'\.
Crop
climate
Increasing inputs ~

1.3 USE AND PROVISION To illustrate the variety of information requirements in agrometeorology, we will
OF AGROMETEORO- list a few of the categories of users that meteorologists are expected to support,
LOGICAL INFORMATION showing how their needs vary. These user groups are not listed in any order of
importance and their characteristics, situation and requests may differ from
country to country.
Owners of large farms of one km 2 or more are often faced with large invest-
ments to be made, in mechanization and sometimes also in (part-time) labour.
Generally speaking, they grow a limited number of crops, progressing in some
·cases to monoculture. In order to run such agricultural factories professionally,
they may require extensive weather and climate information for planning, some-
times in a computer-compatible form.
A horticulturist, dealing with high-value crops grown over a very small area,
usually controls the climatic conditions as much as possible in a greenhouse or in
plastic tunnels. In order to make the necessary adjustments to regulate the radiation
passing through his roof and the ventilation air entering below it, he must have
quantitative information on e.g. local humidity and cloudiness as a routine matter.
Intermediate in . scale between the two above-mentioned market-directed
types of farming are family farms. Their main products are usually animals or
regional staples - viticulture, cereal, grain or rice, grasses, beans or beets, fruits,
sugarcane, vegetables - and some supporting minor crops or animals besides.
These farmers are usually avid customers of agro-weather broadcasts (particularly
on hazards such as frosts) which guide their short-term operational decisions.
Cattle farmers need weather information to provide winter forage for their
herds, e.g. to find out when they can make hay. In latitudes without winter,
farmers are interested in early warning about impending disasters such as floods
or storms.
Government planners (say, from the Ministry of Agriculture) may have the
task of forecasting the yield of the next national staple crop harvest, as an 'early

4
CHAPTER 1 - AGRICULTURAL METEOROLOGY - ITS SCOPE AND AIMS

warning' of crop failures is necessary information for economic measures. This

means that the meteorologist must know about the reliability of existing
weather-related crop development models. Another subject for government
advice is the climatological suitability of various regions for growing new types of
crops.
Agricultural product factories, and other related business involved in
processing or transporting food for the market, can use long-term forecasts and
climatic information for planning purposes - provided that they are able to use
predictions phrased in terms of probability. They prefer statistical climate studies
to climate models (Changnon, 1992).
Investigators specialized in related subjects like hydrology and agronomy
often ask for meteorological datasets. This should be used to start cooperation.

Agrometeorological service In order to cope adequately with such varied requests, it is essential to establish
an agrometeorological unit - preferably in a National Meteorological Service
where, together with forecasting services and a national database, the necessary
infrastructure for data collection and analysis is already in existence. In 1980 a
WMO CAgM working group recommended three subsequent development stages
after the establishment of such a unit:
(1) A 'passive' pre-operational stage using available infrastructure, supplying
simple climatic information for agriculture more efficiently by assigning this
task specifically to a few members of a climatological service. At this stage a
national agrometeorological committee, supported by WMO and FAO,
should identify local weather-sensitive problems;
(2) A more 'active' stage, in which derived information (e.g. growing degree
days, leaf-wetness duration) is provided by a specialized unit. Some countries
might stop here;
(3) An operational stage, where advisory services and forecasts are provided, and
the necessary research is done to support such problem-oriented services.

Agrometeorology and agroclimatology can assist in two categories of problems

which could be designated as the tactics and the strategy of food production.
Strategic considerations cover the choice of crop, selection of farm animals, design
and extent of farm machinery, and methods of soil cultivation. Strategic plan-
ning, in particular, includes the correct choice of land use, production facilities
and genetic issues. Likewise, a statistical study of past climates, their variability,
the range of possibilities - particularly extremes and sequences which the atmos-
phere can generate - can prove useful in risk assessment of future proposals.
Tactical items relate to short-term timing of farm operations, and all that
goes to make up good husbandry. Weather forecasts can play a tactical role, always
provided that such predictions are reliable and accessible, that they do not omit
farming-related weather parameters - and that agricultural effects of approaching
weather are either understood by weatherwise farmers, or stated clearly in the
forecast (see section 8.1).
A forecast can only be useful if there is some action that the recipient can
undertake, either to take advantage of favourable circumstances, or to reduce
adverse effects. All too often farmers are powerless to act - an aircraft can change
course to avoid a hailstorm, but farmers cannot shelter their crops.
Common to all types of problem is the need to understand the effects of
physical conditions. Full understanding demands much more knowledge of the
fundamental biological processes. Accordingly, basic research must be carried out
by teams of scientists. Meteorologists are usually well aware that they are not qual-
ified to comprehend fully the intricacies of biology. On the other hand, as
biologists and agronomists are exposed to the weather on a daily basis, they may
often have the illusion that they have sufficient weather knowledge to deal with
it. Resulting irrealistic simplifications in biological 'weather' models must be
approached with wisdom and patience.
Progress in the realms of interdisciplinary science is slow, but it is an encour-
aging virtue of agrometeorology, that valuable progress can be made to obtain a

5
WMO LECTURE NOTFS - AGRICULTURAL METEOROLOGY

useful and valuable interim degree of accuracy, without delaying application until
results of fundamental research are fully known.

Review questions on Chapter 1

A. Questions on basic understanding:

1.1 Describe briefly the scope of agricultural meteorology. Give an approxi-
mate list of services offered in your country or region by the national
Meteorological Service to farmers, growers, forecasters, agricultural
scientists and advisers, etc.
1.2 How do 'artificial climates' differ from 'natural climates'?

B. Questions on extended application:

1.3 The following table gives the long-period monthly means of mean daily
temperature for stations in two contrasting climates.

Monthly means of mean daily temperature ( C) 0

(Eskdalemuir: 242 m above mean sea level (M.S.L.); Winnipeg: 240 m above M.S.L.)

Station l F M A M l I A s 0 N D
Eskdalemuir 1.4 1.8 3.6 5.8 8.9 11.7 13.3 13.0 10.8 7.6 4.6 2.8
Winnipeg -17.4 -15.3 -4.3 3.4 11.3 16.5 20.3 18.6 12.7 6.0 -4.9 -13.4

(a) Plot the annual course of temperature 1;

(b) Determine the length of period when the mean temperature is 6°C or
above (the 'vegetative season');
(c) Assuming that the mean daily temperature falls off with height at the
rate of 0.6°C per 100 m, then by displacing the curves downwards (or by
re-labelling the axes) estimate the length of the growing season at incre-
ments of 100 m, 200 m, 400 m, 1 000 m, 2 000 m above station height.
What are the agricultural implications of the change with height in these
two contrasting climates?

C. Questions involving practical work or local data:

1.4 (a) List several (e.g. four) of the main categories of agriculture 2 enterprises
(including processing industries if any) in your country or region;
(b) From ea·c h type select a particular case (e.g. for animal husbandry -
cattle or sheep or poultry; for crops - wheat or maize or sugar-beet or
potatoes; for horticulture - tree or bush fruit or citrus or some vegetable
crop).
Consider to what extent the di.mate in your country/region is (or is not)
particularly suitable for the selected enterprises and look at the special weather
hazards (i.e. the general 'weather-sensitivity' of the particular activity).

1 When attempting to estimate duration in intervals of a few days, the time axis should be
sufficiently open to allow us to identify a single day. Allowance should be made for
variation in the number of days in a month.
Monthly means should be plotted against the day that falls at the middle of each
month, i.e. in terms of the number of_the day in the year, which are given below:

I F M A M I I A S 0 N D
16 45 75 105 136 166 197 228 258 289 319 350
2 In this compendium, 'agriculture' always includes horticulture and forestry.

6
CHAPTER 2
RADIATION AND THE SURFACE ENERGY BALANCE

2.1 SOLAR ENERGY Although processes within the Earth's crust contribute some of the energy which
('SHORT-WAVE' ENERGY) reaches the surface of the Earth, the overwhelming source of energy available for
FROM SUN AND SKY all terrestrial processes comes ultimately from the Sun. When the Sun is directly
overhead, its energy flux density is about 2 500 times as great as its flowing up
from the interior of the Earth.
The flux of solar energy reaching a unit horizontal surface of the Earth in the
absence of any atmosphere - or alternatively, impinging on the outer limits of
the atmosphere - is termed the extra-terrestrial solar radiation; often it is called
'Angot' radiation, after the first investigator of its distribution across the Earth. It
is the basic energy input of the atmosphere and its variation with latitude and
time of year is shown in Figure 2.1. The Earth radiates in various ways an equal
amount back to the universe.

Figure 2.1
Daily variation of the solar
radiation at the top of the
atmosphere as a function of
latitude.
Units are 106 J m-2 day-I
(Wallace and Hobbs, 1977).

F M A M A S O N D
Every object emits radiation. Wavelength and intensity of emitted radiation
are specified by Planck's universal law, usually summarized in the laws of Wien
and Stefan-Boltzmann. Wien's laws specify that an object with absolute tempera-
ture T emits maximally at a wavelength of (2897 /T) µm, with a peak value
proportional to T5 . So peak emission is at 0.5 µm for Sun temperatures (6000 K),
and at 10 µm for earth temperatures (including clouds). The Stefan-Boltzmann
law (section 2.1.3) gives the total emission.
This means that of the Sun's radiation about 99 per cent lies in the wave-
length range 0.15 µm-4.0 µm. Of this, some 9 per cent is emitted in the ultra-violet
range (<0.4 µm) and is largely absorbed by ozone in the upper atmosphere; about
45 per cent is in the visible spectrum ('PAR', 0.4 µm to 0.7 µm), and the remaining
46 per cent in the near infrared ('NIR', 0.7µm-4.0 µm). For surfaces at earthly
temperatures, about 99 per cent is emitted at wavelengths larger than 4.0 µm. This
is the reason that solar radiation is called 'short-wave' and terrestrial radiation 'long-
wave'; there is virtually no overlap between these two radiation bands.

2.1.1 Direct and diffuse The energy from Sun and sky is utilized both directly and indirectly in terrestrial
components of solar processes: directly by illuminating and heating surfaces exposed to it; and indi-
short-wave radiation rectly after various transformations in processes such as the heating and

7
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

circulation of the atmosphere and of surface water, evaporation and precipitation

- in essence the external environment in which man, animals and plant life
operate. A capsule in outer space only receives energy from a direct solar beam,
whereas on Earth our environment is powered partly by direct and partly by indi-
rect (transformed) solar energy.
As the solar beam passes through the atmosphere, it is subject to scattering
by molecules, drops, small particles and dust (that gives the sky its colour), selec-
tive absorption of certain wavelengths by some of the atmospheric molecular
constituents (notably 0 3 , CO 2 and water vapour), and is reflected by and within
cloud masses. In all, about 45 per cent of the energy is absorbed at a horizontal
surface on the Earth. This arrives in three ways, divided approximately:
17 per cent from the direct solar beam;
21 per cent as diffuse radiation from the clouds;
7 per cent as diffuse radiation from the sky.
A further 22 per cent is absorbed by atmospheric components, including cloud
drops and dust; part of this reaches the surface eventually after re-emission as long-
wave radiation. The remaining 33 per cent of the solar beam is reflected directly
back to the universe. Figure 2.2 shows a global average (Mcllveen, 1992); other
books may give different figures, e.g. for the Northern Hemisphere only.
Figure 2.2 100% 33%
iolar radiation -reflected back ~o univer.se
Global budget of incoming (4% back-scattered
solar radiation (short-wave) for
the Earth/atmosphere system. \ abs
the a
by atmosphere)

in h"

(28% diffuse) (17% direct)

45%
absorbed by surfaf:e

The energy transmitted along the solar beam is strongly directional, but the
'diffuse radiation from sky and clouds is emitted from all regions of the hemi-
sphere. The intensity of diffuse radiation from the sky canopy is not uniform: the
radiance from the cloudless sky is greater at the horizon than at the zenith, and
that of overcast skies greater at the zenith than the horizon - approximately in
the ratio of three to one. However, in many practical applications such variations
can be ignored. For example, Kondratyev and Manolova (1960) found for the
interception of short-wave radiation on inclined surfaces that, if only rough esti-
mates of daily totals were. required, diffuse flux could be considered isotropic, i.e.
uniform over the whole sky canopy.
The geometry of solar radiation can be important in agriculture, specifically
whenever we are concerned with radiation on a non-horizontal surface. Such cases
arise with the techniques of inducing early growth, e.g. by planting on sloping
surfaces facing the Sun or by ridging the soil for the same aim, in the orientation
and design of greenhouses, and in the energy balance of livestock.
The diffuse radiation from sky and clouds gives an input of Don a horizon-
tal surface; the radiation from the direct solar beam has an intensity I normal to
the beam. To denote inclination of the beam with respect to the horizontal, it is
in practice more convenient to use angle h, the elevation of the Sun above the
horizon, than the zenith angle z = 90° - h (see Figure 2.3). The derivation of h
(and of day length S) from latitude, longitude, time of year and time of day is

8
CHAPTER 2 - RADIATION AND THE SURFACE ENERGY BALANCE

Figure 2.3
Diffuse (D) and direct
components of total global
solar 'short-wave' radiation ZENITH
(K-L) on a horizontal surface (in
the Northern Hemisphere).

given in Appendix III, and in more detail in CL-G and by Oke (1987). The global
irradiation of horizontal surfaces is thus given by:
Kt= I sin h + D = I cos z + D
If we are required to estimate total irradiation of a sloping surface, it is neces-
sary to know:
(a) I and D (from observations or from climate data) and h;
(b) The azimuth angles of the Sun and the surface, and of course the angle of
the slope to the horizontal, in order to derive how much of the diffuse radi-
ation from the sky canopy is cut off by the surface (alternatively, how much
of the sky canopy can be 'seen' by the surface);
(c) The intensity of radiation reflected from the ground and from other
surrounding surfaces.
In precise calculations, 'cosine corrections' are sometimes necessary for the limited
validity of the assumption that the surface accepts and absorbs radiation uniformly
from all directions. Actual computation of the angle at which the Sun's beam hits a
sloping surface is beyond the scope of the current text. Reference is made to CL-G
(Annex 4), WMO-No. 467 (TN No. 152) (1977), to Oke (1987, Appendix 1), and to the
European model (Page, 1986). A discussion of the effects of radiation on sloping
surfaces follows in section 5.5.

2.1.2 Estimation of global The direct measurement of global radiation K-L is becoming increasingly practica-
radiation on a horizontal ble with the development of reliable and economic solarimeters (see Chapter 9).
surface However, because until now such measurements have seldom been made avail-
able from station networks, there is a practical need for methods to estimate Kt
from data which are available. There are two alternative empirical approaches,
using either synoptic observations of cloud amount or sunshine duration meas-
urements. Both methods are based on quantification of the decrease, which the
extra-terrestrial 'Angot' radiation [K] undergoes during its passage through the
atmosphere towards the surface. The methods have at least one uncertainty in
common, namely the local variability in 'turbidity', i.e. the extra absorption due
to dust (aerosol) and water vapour. Turbidity is seldom directly measured because
its observation requires a pyrheliometer, and the relation between turbidity and,
to take one example, available synoptic visibility observations have not yet been
developed for this purpose.
A working example of use of cloud cover observations is the model by Kasten
and Czeplak (1980). If relative cloud amount N is expressed as a fraction of
1.0 = overcast (slightly different from the synoptic notation OHS, but simply
translatable), then:
Kt= [K] (a 1sin h + a2) (1 + b 1Nb2}
where a 1 and a2 are local turbidity coefficients. Holtslag and Van Ulden (1983) got
good average results in temperate latitudes, using a 1 = 990 W m-2, a2 = -30 W m-2,
b 1 = --0.75, b 2 = 3.4. The necessary local value of N is an uncertainty source in this
method: a far-off weather station may have a different cloud climate, formal station

9
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

methods for cloud cover observation are not optimal for solar radiation flux esti-
mation (Stigter, 1983), and satellite data may refer to a different type of cloudiness.
More generally known is the evaluation of sunshine duration measurements by
way of an Angstrom regression formula (see e.g. Coulson, 1975), usually of the form:
Kl- = [K] (a+ b (s/S))

where a and b are turbidity coefficients, and s and S are the actual and possible
durations of bright sunshine. Rietveld (1978) and Martinez, et al. (1984) have
reviewed the known worldwide variations of a and b . In view of typical ±10 per
cent variations found in the values of both a and b, it is simplest to say that a =0.2,
b =0.6, and to leave precise determinations of a and b (e.g. season-dependent) to
local users, maybe based on local data.
Stanhill (1983) gives a rough formula for annual average K-l- at any world
location between latitudes 20° and 55°, in which he also includes the (non-negli-
gible) effect of terrain elevation H (m). Summary of annual data from 221 stations
gives at latitude q>:
Northern Hemisphere: K-l- = 9.3i-0.1033 <I>+ 0.00030 H GJ m-2 yrl
Southern Hemisphere: K-l- = 10.34-0.1253 <I> + 0.00035 H GJ m-2 yrl
Elevation effects relate to transmission path differences for extraterrestrial radiation.
The diurnal course uf global radiation on a horizontal surface suggests that,
either for cloudless days, or for monthly averages, the daily course of total (global)
radiation on a horizontal surface can be represented by a sine curve, given the
mean daily integral [Kl-], as follows (Monteith, 1975):
K-l- (t) = K-l- (max) x sin (nt/S)
Here K-l- (max) = max. irradiation at solar noon, S = day length in hours, t =
time in hours after sunrise, and the daily integral [K-1-) = (2S/n) Kt (max). See
Gloyne (1972) for alternative expressions.

2.1.3 Emission and When radiation reaches part of the surface - such as a leaf, rock or river - it can
reflection of radiation be absorbed, reflected or transmitted. When absorption occurs, by Kirchhoff's law,
emission of radiation also occurs with equal efficiency. If unity short-wave energy
References: Geiger sections
of wavelength A arrives, conservation of energy for the fractions which are
2-4; Oke (1987), Iqbal (1983),
absorbed@, reflected (p), or transmitted (\JI), requires that:
WMO TN No. 152.
l;+p+\j/=l
Also, Kirchhoff requires equality of absorptivity and emissivity (l; = e) at a given A.
For example, if pure air (i.e. nitrogen and oxygen) is transparent to visible light,

Table 2.1
Radiative properties of natural materials
Surface State Albedo a Emissivity E

Soils Dark, wet 0.05- 0.90-

Light, dry -0.40 -0.98
Grass Long (1 m) 0.16- 0.90-
Short -0.26 -0.95
Agricultural crops, tundra 0.18-0.25 0.90-0.99
Orchards 0.15-0.20
Forests Deciduous, bare 0.15- 0.97-
Deciduous, leaved -0.20 -0.98
Coniferous 0.05-0.15 0.97-0.99
Desert 0.20-0.45 0.84-0.91
Water High sun 0.03-0.10 0.92-0.99
Low sun 0.10-1.00 0.92-0.99
Snow Old 0.40- 0.82-
Fresh -0.95 -0.99
Ice Sea 0.30-0.45 0.92-0.97
Glacier 0.20-0.40
Source: Oke (1987). See Geiger (Table 4-2) for specific agricultural surfaces.
CHAPTER 2 - RADIATION AND THE SURFACE ENERGY BALANCE

i.e. visible wavelengths are not absorbed, so air will also not emit energy at those
wavelengths.
Every body emits radiative energy; its wavelength distribution is given by
Planck's law. The complete all-wave emitted energy is almost solely dependent on
the surface temperature, according to the law of Stefan-Boltzmann:
Energy emitted= nTT4 W m-2 , with cr = 5.67 x 10-s W m-2 K-4.
Temperature in radiation analysis must always be expressed in Kelvin _
°C + 273.16. The emissivity factor e for surfaces in nature is usually so close to 1
that it is negligible. While 'reflectivity' refers to the reflected fraction of incident
radiation at some particular wavelength, the term albedo (a) denotes the reflection
over the whole solar spectrum.
Most natural surfaces show wavelength variation in reflectivity - e.g. snow
is very reflective to visible and fully opaque to infrared radiation. In particular,
surfaces of leaves show a marked variation in reflectivity according to wavelength.
In the visible spectrum (0.4 µm-0.7 µm), 70 to 80 per cent of the radiation reach-
ing the leaf surface is absorbed; these wavelengths are known as Photosynthetically
Active Radiation (PAR). For green light, absorption is less intensive and reflection
greater than for blue or red light, causing the predominantly green visible appear-
ance of young healthy leaves.
Between about 0. 7 µm and 1.0 µm, many leaves are rather translucent,
reflecting and absorbing only half of the incident radiation. At wavelengths of
more than 1.0 µm, the leaf absorbs an increasing proportion of the solar radiation,
which becomes almost complete at about 2.5 µm. Integrated over the whole spec-
tral range, the coefficients of reflection and transmission are both approximately
0.25 for leaves.
Most natural surfaces act as diffuse reflectors: incident radiation is scattered
in all directions. However, for open water, shiny leaves and other smooth
surfaces, as the incidence angle h becomes more acute, then the reflection
becomes more and more mirror-like (specular) and there is a corresponding
increase in reflectivity.

2.1.4 Energy in the visible Visible radiation plays a fundamental role in biological processes such as photo-
spectrum - light synthesis (PAR) and photomorphogenesis in plants, vision in animals and man and
photoperiod effects both in plants and animals. About 45 to 50 per cent (depend-
ing on the wavelength limits assumed) of the energy reaching Earth from the Sun
is in the visible range.
When evaluating published data on the relationship between irradiation and
plant dry-matter production, a source of difficulty can be that incoming solar
energy is sometimes expressed in terms of the total solar spectrum, and sometimes
in terms of the energy containing the photosynthetic bands, i.e. the visible spec-
trum. Careful reading of such texts is needed, because the units used are not
simply interchangeable.
The measurement of light (or illumination) is understandably linked to the
response of the human eye to certain wavelengths. The normal human eye
responds to energy within the band, but more intensely to some wave bands than
to others, and particularly to the blue-green wave bands (about 0.55 µm). The
linkage between energy and illumination is given by a variable 'luminous effi-
ciency', expressed in lumens per Watt. The Watt-lumen conversion curve has by
definition a peak at 0.555 µm, where by definition 1 Watt = 680 lumen. However,
the curve falls off rapidly towards its ends at 0.4 µm and 0. 7 µm, where 1 Watt =
0 lumen. For instance, at 0.50 µm already we only get 272 lumen for 1 Watt.
Conclusion, never use a photometer just to measure solar radiation.
For global (whole-sky) radiation, we usually assume an average conversion
figure of 1 Watt = 110 lumens for the conversion of energy flux to equivalent
luminous flux. The luminous efficiency of direct sunlight is less, and that from sky
and clouds more than the average figure, since radiation from the sky has a higher
proportion of the shorter wave lengths than is found in illumination by the direct
solar beam.

11
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

A light source emitting luminous energy at the rate of 1 lumen per unit
solid angle (measured in radians) has the power of 1 candela. At a distance of
1 metre from this source, the intensity of illumination of a perpendicular surface
is 1 lux. As 1 W m-2 = 103 lux x luminous efficiency, this means, on a bright day
(KJ, = 1 000 W m-2) we have approximately 110 000 lux.

2.2 THE ENERGY The short-wave (solar) energy budget having been discussed, the terrestrial (long-
BALANCE AND ITS wave) budget is now examined.
COMPONENTS Planet Earth has an external source of energy, the radiant energy of the Sun,
so it must radiate energy to its surroundings at the same rate to remain in ener-
References: Geiger section 5;
getic balance. In latitudes 0°-35° more energy is being gained than lost; from
GM-II, CL-G, Oke (1987).
latitudes 35° to 90° more is lost to outer space than gained from the Sun. The
global atmospheric and oceanic circulation then equalizes the differences.
2.2.1 The long-wave budget
As we have seen, a substantial proportion (about 33 per cent) of solar radia-
tion is reflected - unchanged in wavelength - back into space from the
atmosphere, clouds and the planet's surface (earth and water). The remainder
passes through the atmosphere being only slightly depleted by absorption in the
atmosphere. Some minor atmospheric constituents, however, readily absorb long-
wave radiation. The same constituents simultaneously emit radiation within the
same wavebands, but with an intensity governed by their own temperature.
Because the temperature of· the ground surface is higher than that of atmospheric
constituents, the surface loses long-wave radiant energy at a greater rate than can
be compensated by clouds.
On the basis of a standard absorption spectrum in cloudless skies, the water
vapour and carbon dioxide absorption may be divided into two regions:
(a) Complete absorption:
(i) By a small amount of water vapour from 5 µm to 8 µm;
(ii) By a small amount of carbon dioxide from 13 µm to 17 µm;
(iii) By a moderate amount of water vapour from 17 µm to 20 µm;
(iv) By a small amount of water vapour from >20 µm.
(b) Almost complete transmission:
By any amount of water vapour or carbon dioxide from 8 µm to 13 µm. In other
words, there is no absorption of radiation by the atmosphere in this 'window',
and radiation proceeds undiminished to outer space. However, a sufficiently thick
cloud sheet will effectively block the 'window', making its transmission zero.
A 'small' amount of water vapour or carbon dioxide may be taken as that
normally present in a vertical column of air of unit cross-section near the ground
and about 50 m high; a 'moderate' amount contained in a similar column one
kilometre high; a 'large' amount of a constituent is that in columns extending
· throughout the whole atmosphere. Thus a column of air one kilometre high will
completely absorb energy between 5 µm and 8 µm and greater than 20 µm in the
upward beam of terrestrial radiation.
Hewson and Longley (1944) show how an approximate estimate of nocturnal
cooling under clear skies can be estimated from a graph, in which radiant intensity
is plotted against wavelengths. The above assumptions were slightly modified by
replacing the band 8-13 µm by 8-14 µm to allow for the fact that throughout the
one kilometre layer, required for 17-20 µm absorption, the temperature of the air
will be rather lower than the surface temperature (Figure 2.4).

Figure 2.4
Energy transmitted (hatched
area) and absorbed by the C:
atmosphere (Hewson and -~!ti0
Longley, 1944). ~
ra
ex:

0 5 10 15 20 25 30 35 40
Wavelength in µm

1Z
CHAPTER 2 - RADIATION AND THE SURFACE ENERGY BALANCE

If we assume that the mean temperature of the first SO m of atmosphere is

the same as that of the Earth's surface, then the upward and downward streams of
radiation outside the hatched area will be the same and thus the only way for any
loss is the wave band covered by the hatched area. Stefan's law states that the area
under the entire curve is proportional to the radiant emission from a black body.
As surface temperatures change, so will the area under the curve and the
wavelength of maximum intensity "-max · Wien's law say that "-max T = 2897 K
(with ).., in µm). When the temperature falls, the area under the curve will be
reduced and the peak is moved to longer wavelengths.
If y = (hatched area)/(total area), then from Planck's law it follows that:
T (K) 200 225 250 273 300 325
y 0.22 0.27 0.32 0.35 0.37 0.39
Taking a mean value of y = 0.35 for the 250-300 K range experienced at the
ground, the nocturnal loss of radiant heat occurs at the rate yoT4, where cr is
Stefan's constant. When T = 300 K, the planet's heat loss is about 160 W m-2.
The same authors investigate nocturnal radiation in the presence of cloud,
showing that the loss through the atmospheric 'window' increases with the height
of the cloud base because the temperature of cloud base decreases with height.
They conclude that for a given temperature difference £\T(= T(surface) - T(z = h)), the
loss on overcast nights is proportional to T3 (surface), while if cloudless it is propor-
tional to T\surface) (Tin K).
Long-wave radiation estimation We consider here the long-wave budget of the surface, L = Li-L-!., where:
L = effective outgoing radiation from the Earth's surface (positive; L0 for clear
skies and Le for cloudy skies will be used sometimes as notation);
Li= the terrestrial radiation (depends on surface temperature; upward= positive);
Lt = the long-wave downward counterradiation from the atmosphere and
clouds (a function of air temperature, water vapour and cloud cover; down-
ward = positive).
Lt , although a result of processes taking place in the troposphere, can be usefully
evaluated in terms of observations of air temperature and vapour pressure at 1.5 m
height. The reason for this is that the major portion of downward-radiating water
vapour is usually found in the lowest 1 to 2 km of the atmosphere, within a plan-
etary boundary layer which is well mixed. Therefore surface station observations
give fair information of the downward radiation properties of this vapour.
The oldest formula which appears in many handbooks is by Brunt (1932):
L-!.0 = (0.605 + 0.048 ✓e) crT4 (W m-2)
with surface humidity e in hPa (= mb). Formally Brunt's approach is good physics,
since long-wave emission is largely from water (vapour). However, operationally it
is quite risky because observations of humidity are usually much poorer than
those of temperature.
A more recent, and more reliable, estimation formula is by Swinbank (1963):
L-J., = (5.3 x I0-13) T6 + 60 N (W m-2)
where T is the local temperature measured at standard station screen height, i.e.
about 1.5 m. Swinbank's formula is based mainly on data from Australia and the
Indian Ocean, and Arnfield (1979) shows that its reliability is about ±5 per cent.
Note that Swinbank's formula, as given here, contains a correction for the fraction
of cloud cover, N (with 0:$;N:$;l). Paltridge and Platt (1976) give working formulae
for Lt, which allow for effects of different cloud types and heights of cloud base,
but these effects seem to be marginal.
A third L.j, formula, which uses temperature like Swinbank's formula, was
developed by Idso and Jackson (1969), and used additional field data at freezing
temperatures:
1J,0 = {l - 0.261 exp[-7.77 x I0-4 (273-T)2]} cff4 (W m-2)
It has no correction for cloudiness; nevertheless it seems to perform reason-
ably well.

13
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

The choice between these formulas might be made from what is known
about the available data. If you are sure that the humidity observations are reli-
able - that the psychrometer is ventilated, and that the wet-bulb sleeve is
replaced once a week, not just once a year - it is defensible to use Brunt. But if
you are uncertain (or even negative) about the observations of humidity,
Swinbank or Idso and Jackson are much safer to use. FAO recommends ldso &
Jackson (Allen, et al., 1994).

2.2.2 Surface radiation For estimating Li, we base ourselves on the Stefan-Boltzmann emission law. This
temperatures requires as input the radiation temperature of the ground, T0 • Often large differ-
ences between T0 and the Stevenson screen temperature T occur, e.g. with clear
skies and light winds, associated with relatively high radiative exchange (Gloyne,
1953; Beljaars and Holtslag, 1991). Indeed, IT-T0 I can vary from 4 K to more than
10 K, depending on the amount of net radiation received (or emitted) by the
surface, Q*, and.on the heat conduction properties of the soil. If only Tis known,
Holtslag and Van Ulden (1983) got useful radiation estimates by using a soil-
dependent emission correction:

Li = cff4 + c 5Q* (W m-2)

where ~oil heating coefficient c5 = 0.1 over wet grassland and = 0.4 over dry bare
soil.
In frost studies, readings of a 'grass minimum' thermometer (also called
'radiation minimum' thermometer) are often used. This instrument (GM II, Ch.II-
4; HMSO, 1982) has to be placed, with its bulb fully exposed, above a standard
short grass cover just touching the tips of the blades of grass; or it may be placed
several centimetres from the surface, or even in an array at several different levels
to screen height; or it may be placed over bare ground; or the surface may not be
bare ground or grass. Representativity of grass-mimimum data for e.g. orchard
temperatures, to take one example, is very, very doubtful (Smith and Gloyne,
1951). Also, grass heights are seldom kept constant.
However, the instrument remains in use because of the practical value of the
observations as an indicator for traditional classification criteria for radiation-
dominated temperatures. The basic reason not to use the standard Stevenson
screen temperature for frost warning is that screen thermometers are sheltered
from direct radiation - on purpose, since their aim is to measure air temperature
- and that frost is a situation where the role of radiation is important. We can
understand the survival of the device from the fact that at present, all
meteorological stations are not routinely equipped with real radiometers. In the
.meantime, extensive empirical use of grass-minimum data has been developed
(section 9.2.2), e.g. the French ;indice actinothermique' (Perarnaud and Rayna!,
1991), which we might as well use, as .long as nothing better has been made
available for use. The situation is best summed up in the saying: "Don't throw out
your old shoes before you have new ones".

2.2.3 Total radiation budget Of the net energy received at ground surface (Q*), the short-wave input consists
and complete surface of the non-reflected part of glob.al radiation. The long-wave output is the outward
. energy balance terrestrial radiation, diminished by the do~nward atmospheric long-wave radia-
tion. The formula, over a terrain wit1:1 albedo a, is as follows:

Q* = Kl (1-a) + Ll - Li
This net input is utilized in a number of processes, mainly heating the air,
heating the ground and evaporating water. Non-meteorologists are often
surprised how large and diurnally steady the long-wave components are, quite
comparable with short-wave solar radiation. Kl does not always dominate the
radiation balance, because of its strong diurnal and seasonal variation.
Since direct measurements of Q* are seldom available from stations, it is
interesting to find out how reliable the estimation methods for the various
radiation balance components are. This was investigated by Holtslag and

14
CHAPTER 2 - RADIATION AND THE SURFACE ENERGY BALANCE

Van Ulden (1983), who used the methods of Swinbank (for Ll-), and of Stefan-
Boltzmann with soil heating correction (for Li) to estimate long-wave radiation
fluxes. The overall working formula is then:
Q* (1 + cs) = (1 - a) Kl-+ (5.3 x 10-13) T6 + 60N - cff4
with symbols as defined in earlier parts of this chapter. Basic input is therefore
station screen temperature T, cloud cover N, and surface characteristics (cs and a).
For Kl- information, either an estimate was made from cloud cover (Kasten-Czeplak)
or, alternatively, direct local global radiation measurements were used as input.
Results of comparing these alternatives with full Q* measurements are shown below.
Figure 2.5 shows three facts. First, the modelling of short wave radiation
from cloud cover introduces no bias, but gives a very large increase in uncertainty.
Second, the modelling of the long wave radiation by way of Swinbank's LL-model
and use of the soil-dependent T-T0 -correction performs very nicely. Third, it
follows from the two preceding points that for a good determination of surface
energy balances it is necessary and often sufficient to measure global radiation.
Figure 2.5 Q*obs Q*obs
600~------------- 600-r--------------
Comparison of measured half-
hour averages of net radiation
at Cabauw (Q*obs) with
estimated values (Q*est). In the 400 400
left-hand graph, measured
global radiation is used, while .6
.6
in the right-hand graph the .6 .6

incoming solar radiation is 200 200 .6

.6 .6

.6 .6
estimated by use of cloud cover .6
.6
.6.6.6

data and the Kasten-Czeplak r = 0.982 ~'.6.6 .6 r = 0.857

cr = 25 Wm-2 .6
CT= 63 Wm-2
model (Holtslag and Van
Ulden, 1983).
0 200 400 600 0 200 400 600
Q*est Q*est

Physical shading (agricultural sites are rarely unobstructed) affects the

incoming short-waves and, if the solar beam is cut off by obstructions, the
short-wave input is restricted to diffuse solar radiation from that portion of the sky
canopy 'seen' by the ground, and to reflected short wave radiation from ground
and surroundings. Equally, the long-wave exchange is often only partly between
ground and sky, the rest being between ground and local obstructions which
screen the sky, such as nearby tree lines.
The energy equation is balanced when the net incoming energy is split into
the various 'sinks' or users of the energy. In physical climatology the equation at
the Earth's surface is:
K!. (1 - a) + (L.!. - Li) = H + LE + G (+.!\Q)

where H = energy used in heating the air ('sensible' heat flux);

LE = energy used for evaporation (E given by the amount expressed as the
depth of water; L is the latent heat of evaporation);
G = energy used in heating the ground (see Chapter 3);
L\Q = energy storage, production or use within the system .
.!\Q enters only if the system has a finite size and is not just a two-dimensional
plane; and then it may take many forms. For limited periods various energy 'sinks'
or 'sources' will arise, e.g. a term M for melting of snow and ice, or an extraneous
contribution (a negative sink) when modelling events like forest fires, or a term for
exchange of energy in plant respiration and photosynthesis; and very often,
advection.
The terms H and LE are usually of the same order of magnitude, and each
usually larger than G or .!\Q (see section 4.3.3). A positive value of LE implies that
there is water available and that evaporation is taking place; negative LE-values
occur during dewfall. Without evaporation, as in the desert, much net energy is

15
WMO LECTURE NOTF.S - AGRICULTURAL METEOROLOGY

devoted to heating the ground to a temperature sufficiently high that Li, together
with the air-heating energy, can close the budget. Over the ocean, about 90 per
cent of the energy is used in evaporation.
The partitioning of energy between Hand LE forms a central process in the
'energy balance' approach to the estimation of evaporation and transpiration. This
will be dealt with further in Chapter 4.

2.3 SPECIAL ASPECTS As soon as a green plant has some above-ground parts, photosynthesis begins;
OF RADIATION AND this is the process by which the organic materials of the plant, predominantly
TEMPERATURE IN carbohydrates (glucose, starch, etc.) are formed through the action of photo-
AGRICULTURE synthetically active radiation, PAR (see section 2 . 1 .3). This is expressed
schematically as:
References: Lowry (1967 /1989),
Griffiths (1994), McNaughton CO 2 (carbon dioxide)+ H2 0 (water)+ radiation ➔ carbohydrates+ 0 2 (oxygen)
and Jarvis (1991).
The reverse process for the purpose of maintenance, by which carbohydrates
are 'burnt' to give CO 2 and H2 0, is called respiration and goes on permanently
with a temperature-dependent intensity.
Photosynthesis takes place within the inner leaf tissues and requires access
to atmospheric CO 2 and to light. In most circumstances, leaves receive more than
enough sunlight - i.e. they are 'light saturated' - and other factors, especially
access to CO 2, limit photosynthesis. The outer surfaces of most plant tissues are
relatively impermeable to atmospheric gases and to water vapour. CO 2, 0 2 and
H20 are exchanged through tiny pores called stomata (see Kramer and Boyer,
1995) found largely on the underside of leaves. The length of a stoma, when
opened, is less than 0.03 mm. Depending on the species, there may be up to
several hundred stomata on every square millimetre of leaf surface, though their
total area seldom exceeds one hundredth part of the surface.
In most plants, stomata open in response to a combination of light, CO 2
concentration and humidity. The 'guard cells' that open the pores do so by absorb-
ing water and swelling. Like the surrounding tissues, they contain green
'chloroplasts' which, when PAR falls on them, enable the photosynthetic produc-
tion of sugars and other products that increase the cell's attraction for water. If
conditions are moist enough, the cells absorb enough water to open the stomata.
Although the amount of water actually consumed in photosynthesis is quite
small, the process can usually only take place when accompanied by considerable
water loss in daytime through simultaneous evaporation from the pores. In many
plants the stomata are closed at night, although some species reduce water loss by
accumulating CO2 at night and dosing the stomata by day.
The close coupling of photosynthesis with evaporation makes it possible to
· predict crop yield from measurement or modelling of the meteorological process
of evaporation - assuming that water and nutrients are available in the right
quantities. If adequate water is not available to the plant roots (section 4.2), the
stomata will close, water circulation through the plant and the movement of
plant nutrients and products will decrease, and leaves will wilt.
Evaporation of water from leaves (called 'transpiration') is not simply an
unwanted effect qf the need to allow CO 2 to enter leaves. It plays a key, though
not fully researched, ~ole in the transport of water and nutrients up from the roots.
It also requires energy, which is drawn mainly from the incoming solar energy.
This cools the leaves, which otherwise have a tendency to overheat because most
leaves have little capacity for storing energy. Evaluation of the leaf energy balance
(Oke, 1987) shows that:

T1eaf = Tair + (rb/(pcp)) (Q*ieaf- LE1eaf)

Here rb is the resistance due to the build-up of a steady laminar air layer around
the leaf. This resistance is great for large leaves in weak wind - then (Tieaf - Tair)
can easily exceed 10 K on a warm day, unless the leaf cools itself by evaporation.
Similarly on a cool still night the leaf surfaces can be very much colder than Tair
as measured in the screen. Biological action depends on the real Tieaf, not on easily
available Tair-values.

16
CHAPTER 2 - RADIATION AND THE SURFACE ENERGY BALANCE

Figure 2.6 In a crop or forest the solar energy has to penetrate the canopy to the lower
Growth of a crop during the leaves. The obstruction by leaf arrangement to the penetration of radiation is
period of an experiment with expressed by the leaf-area index (LAI), i.e. the total leaf area per unit of ground area.
difference or similarity in The denser the canopy, the greater the irradiation needed to ensure continued
night-time ('nycto-') photosynthesis by the lower, partially shaded, foliage. In forest canopies, penetra-
temperature Tn and daytime tion of radiation to the forest floor suffers special difficulties, because of the
temperature TP. (Lowry 1967, arrangement of trunk, branches and leaves. In model terms, the forest consists of
from data by Went). a collection of vertical cylinders (trunks) with a dense growth of horizontal
elements (branches) some distance from the ground.
Because the effective radiating temperature of the atmosphere is usually
lowest directly overhead, the net outgoing radiation is greatest in this direction
and decreases with increasing zenith distance to zero on the horizon. Hence a
.c
canopy of a large tree with many branches provides an effective frost protection
!61--+---+---+---+---+-----1
for the underlying surface by shielding it from the coldest portions of the sky
°'
iii
~41---+----+------+---+----l
towards which the radiated losses are greatest, while radiating itself downwards at
ambient temperature (Bohren, 1987). Sellers (1965) provides formulae for evaluat-
ing radiation within a circular rim of given height above the horizon; on slope and
hill effects, see section 5.5 for further information.
0

Every plant has its own sequence of development phases (phenology),

0 5 10 15 20 25 30 marked by changes in aspect: formation of sets of leaves, roots or buds, flowering,
Nyctotemperature (°C) and various stages of ripening. Examples of such sequences are given for grains in
section 9 .4.3.
Growth denotes increase in size, which is not the same thing. Also, desired
growth depends on the progress of development stages. In springtime, warm
weather in the early development stage increases the yield of beets because their
roots only start to grow well when the leaf formation phase has ended. For grain,
warm weather in the early development stage reduces its yield: too early flowering
causes insufficient formation of leaves and roots, giving small ears. So models
must be crop-specific.
Estimation of photosynthesis depends on the estimation of radiation, while
temperature data are needed to monitor growth rates and respiration losses. Night-
time and daytime therefore should really be separated in order to obtain reliable
agrometeorological modelling results.
Relevant in this context is the phenomenon of thermo- or photo-periodicity,
illustrated in Figure 2.6. In this example, plants growing at constant temperatures
(night-time Tn = daytime Tp) developed only moderately, while a clear optimum
of growth is indicated for TP = 25° and T0 = 15°. Various counteracting processes,
e.g. cell elongation versus respiration, may be claimed to be responsible for this
(Lowry 1967) - but the important fact is that such effects exist, and that 24-hour
averages of parameters are therefore often too coarse to describe adequately devel-
opment. In temperate zones, another factor which should not be forgotten is the
length of the day.

Only a fraction of solar input is captured by green plants. When Penman (1948)
developed his well-known method of estimating evaporation from a crop, he
concluded that over an annual analysis period, only about 1 per cent of incident
solar energy was used in photosynthesis - and therefore omitted this item from
his energy balance sheet.
On a bright day about half of the sunlight is photosynthetically active, and
the leaves reflect about 15 per cent of this and absorb 5 per cent inactively. The
basic chemical process of photosynthesis processes PAR at an efficiency rate of
25 per cent. Thus the maximum possible efficiency of photosynthesis is about
10 per cent (Monteith, 1966). In full sunlight, the utilization rate is more likely to
be around 3-5 per cent because of respiration losses. Then for annual crops, a high
proportion of solar radiation falls on a bare soil, either in the spring before germi-
nation, or during the early weeks of plant growth before the crop forms a complete
ground cover, or in the fall after harvest. Even before harvest, when the leaves of
the mature crop start to discolour, less PAR is absorbed for photosynthesis. So the
net annual efficiency of photosynthesis is low.

17
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

But because the solar input is large, the net result in the production of plant
material is still worthwhile. An enormous amount of energy is locked up in all forms
of vegetation over the year, as well as in previous vegetation (i.e. in fossil fuels).

Review questions on Chapter 2

A. Questions on basic understanding:

2.1 Assuming that the Sun radiates approximately 2.3 x 1028 Joule of energy
every minute, estimate the flux on a unit area of 1 m 2 held at right
angles to the solar beam at the mean distance of the Earth from the Sun
(i.e. estimate the 'solar constant'). Express the results in S.I. (Standard
International) units, i.e. W m-2, in cal cm-2 min-I, and in British
Thermal Units, i.e. Btu ft-2 b-I.
2.2 The total energy intercepted by the Earth in unit time is:
n (radius of Earth) 2 (solar constant)
What is this in terms of MJ day-I, cal day-I and kWh day-I?
Show that the solar energy reaching the outer limits of the atmosphere,
if averaged over the whole globe, equals one-quarter of that intercepted
by the Earth.
2.3 Table 2.2 gives period averages of mean daily global (short-wave) radia-
tion on a horizontal surface at places stated. (Sources: World Survey of
Climatology, Volume 5, Elsevier Press, and Solar Radiation National Atlas,
Table 2.2 United States.)

(Wm-2)

Month J F M A M I J A s 0 N D
Place
Barrow 71.3°N, 156.8°W 0 18 87 184 249 256 208 124 75 20 0 0
Troms0 69.6°N, 18.9°E 0 10 so 105 142 194 192 115 58 20 2 0
Reykjavik 64.l°N, 21.9°W 6 26 79 140 209 200 212 170 87 39 10 2
Lerwick 60.1 °N, 1.2°w 11 33 70 129 166 193 173 134 87 42 16 7
Stockholm 59.3°N, 18.1 °E 13 39 87 101 181 214 198 159 102 49 20 9
Copenhagen 55. 7°N, 12.5°E 24 52 99 157 212 245 217 168 129 63 23 16
Kew 5 l.8°N, 0.3°W 25 44 91 130 180 201 182 155 118 68 32 18
; Seattle 47.6°N, 122.3°W 32 61 119 176 .216 223 240 211 145 82 45 29
Salt Lake City 40.?°N; 111.9°W 79 124 172 232 276 301 300 267 216 153 99 71 .
· Lincoln 40.8°N, 96. 7°W 91 126- 170 202 239 264 275 235 192 143 96 77
Washington 38.9°N, 77.0°W 86 120 166 212 249 269 248 221 189 142 98 76
Los Angeles 34.0°N, 118.3°W 120 160 228 250 277 289 311 282 244 1'81 140 117
· Phoenix 33.4°N, 110.2°W 146 198 255 309 351 358 319 297 274 218 167 136
Miami 26.8°N, 80.2°W 169 201 237 262 268 258 258 245 213 186 171 153
Honolulu 21.3°N, 151.2°W 176 205 250 271 299 298 298 297 278 246 206 180
Canton Island 3°S, 1n°w 285 303 307 293 272 266 267 289 310 315 291 277

(a) Make a graph of maxima against latitude, and discuss latitudinal

changes. Why does the Canton Island record have two maxima?
(b) · Compare and contrast the data.for stations at roughly the same lati-
tude but different geographical settings (e.g., inland, coastal,
island);
'(c) Give climatological and/or synoptic reasons for observed
differences in the pattern of radiation received in contrasting
regions;
(d) For half a dozen selected stations, read off from Figure 2.1 the mid-
month figures of mean daily extra-terrestrial radiation for each
month.

18
CHAPTER 2 - RADIATION AND THE SURFACE ENERGY BALANCE

2.4 The energy emitted by a surface of emissivity E at temperature T (K) is

equal to:
Ex er x T4 W m-2 (er= Stefan's constant= 5.67 x lQ-8 W m-2 K-4)

(E = 1 for a 'black body', or more appropriately, a 'full radiator'). Assume

that the Sun is a 'full radiator' and a sphere of radius 6.96 x 10s m at
1.50 x 1011 m distance. Then, using data from 2.2, show that the surface
of the Sun is = 5 800 K.
2.5 The mean temperature of the Earth/atmosphere system has remained
sensibly constant for a long period, so we assume that there is equilib-
rium between the streams of solar and infra-red radiation absorbed
and emitted by the Earth/atmosphere system. If E = Earth's radius,
albedo a = 0.36, er = Stefan's constant and SC = solar constant, then
confirm that:
4nE2 x er T4 = 1tE2 (1-a) SC
and for SC= 1.37 kW m-2 show that T = 250 K.
The observed mean surface temperature is about 288 K. Explain the
difference (see McIntosh and Thom, 1969).
2.6 Assuming that the Earth is a perfect sphere, show by a simple geometri-
cal argument that at local noon the angle of incidence of the direct solar
beam on a surface sloping upwards with angle 4>0 towards the south is
the same as on a horizontal surface situated at a latitude 4> further south.
0

Let such a surface be called the 'corresponding' plane.

Examine the situation with slopes of different angles facing both north
and south and consider east-and west-facing slopes. To what extent is it
possible to use data on clear sky irradiation on a horizontal surface for
estimates of irradiation on a 'corresponding' surface?
2.7 If T(n) = minimum screen temperature during the night,
T(g) = 'grass minimum' temperature during the night,
then investigate the relationship between the difference (T(n)-T(g)), and
other climatic elements - in particular degree (and type) of cloud and
the wind speed. To what extent is it possible to deduce the conditions
during the night from values of (T(n)-T(g))?

B. Questions on extended application:

2.8 On a clear day, the hourly input on 1 cm 2 of horizontal surface at
Toronto was given as below (in J):
Time 0600 0900 1200 1500 1800
Total (global) 92 289 377 268 88
Diffuse 17 29 46 59 17

Compute the intensity of direct radiation at normal incidence (i.e. at

right angles to the direction of the solar beam). From geometrical
considerations the values should all be equal. Explain why differences
do, in fact, arise.
2.9 Using the data of 2.8, compute the intensity of illumination on a hori-
zontal surface at the hours quoted. What kind of hypothesis is needed
to do so?
2.10 Assume the relationship K.!. = [K] (0.135 + 0.68 s/S)
(with s = hours of bright sunshine per day, and S = hours of daylight for
the middle day of each month) applies to Kew Observatory (52°N), and
the observed radiation is:
January April July October

KJ. (W m-2) 25 130 182 68

s (hours) 1.7 4.6 6.2 3.5

19
WMO LECTURE NOTE.5 - AGRICULTURAL METEOROLOGY

Estimate [K] from Figure 2.1 and calculate S using Appendix III. Compare
the calculated values of K with the observed values given above.
2.11 An alternative approach to that of 2.10 can be useful for the study of
global radiation at individual places.
If KJ. 0 = global radiation on days with clear skies, then:
Kl= Kl 0 (c + d (s/S)). Angstrom suggests c = 0.235, d = 0.765.
Use the data found in question 2.10 to compute the mean daily clear sky
irradiation for the place and months suggested. In this question c+d = 1;
in 2.10, a+b<l. Why is there a distinction? Explain the physical signifi-
cance of the coefficients a, b, c, d.
2.12 An approximate relationship exists between mean daily values of Kl and
hours of bright sunshine (s) for a given place. For Poona (19°N) are quoted
(Ramdas, 1961, average values):

l F M A M l l A s 0 N D
Kl (W m-2) 234 260 291 302 332 264 193 201 252 260 244 210
s (hours) 9.3 10.1 10.1 9.8 10.7 5.9 2.8 4.1 6.2 8.6 9.5 8.7

Plot these values and draw the 'best straight line' by eye or calculate it
by 'least squares'. Estimate from it the values of KJ. for s = 4, 10 hours.
2.13 Show that, on a night with a heavy overcast sky, the net outward radia-
tion is proportional to T(s) 3, where T(s) =surface temperature in K, and
to ~T, where: ~T = T(s)-T(b); (T(b) = temperature of cloud base).

20
CHAPTER 3
THE SOIL AND ITS HEAT BALANCE

3.1 WHAT IS 'SOIL'? Soil is not a true solid, but consists of individual particles and aggregates of
mineral and organic materials, separated by spaces or pores which are occupied by
water and air. The relative proportions of these last two constituents can vary
widely according to the geometry of soil spaces and the total water content. The
relative amount of pore space decreases with increasing soil grain size - intu-
itively one would expect the opposite.

3.2 TRANSMISSION OF The energy balance at the atmosphere-earth interface was earlier discussed in
HEAT IN THE SOIL section 2.2.3, and will be discussed again in section 4.3.3. In those discussions we
deal mainly with partitioning atmospheric energy fluxes and above-ground
References: Lowry (1967/1989),
energy storage, and summarize flux into the group simply as G. Here we deal with
Geiger (Sections 6, 19),
partitioning of G, and we can better write here the central part of the energy
Oke (1987).
balance equation as:
Q*-H-LE = G
i.e.: (net radiation) minus (upward atmospheric heat fluxes) = (heat flux into the
ground). Now examine the heat flow and heat storage components of G
separately.
In the context of the energy balance we are only interested in heat storage
changes, not in the total amount of heat stored in the soil. Formally we define the
specific heat of materials, C, as the amount of energy needed to increase the
temperature of a unit weight by 1 K. Because in soil heat flux we are thinking in
terms of distances and volumes, we have to multiply C with the soil density
p (kg m-3) in order to obtain the heat capacity p C (unit: J m-3 K- 1), the amount of
heat which is needed to raise the temperature of a unit volume by 1 K.
If therefore a quantity of heat is added over a unit surface area and is stored
in a layer of thickness dz, the temperature of this layer will increase by ~T, and
then it is valid by definition of heat capacity that:
increase in soil-stored heat ~Q = (density x area unit x dz) C~T = (pC) ~T dz
Essentially this relation says, that for a given heat addition the temperature
increase in the layer is less if the layer is thicker, and that it is proportional to the
heat capacity, which is characteristic of the soil type. (Freezing or evaporation are
complications which will be examined later.)
If a soil surface is heated, energy is transferred downwards through the soil,
mainly by conduction, but also to a lesser extent by liquid flow and by vapour
convection in the soil pores. Per unit time, the flux of energy through a soil layer
of unit horizontal area and thickness dz are proportional to the vertical tempera-
ture gradient over dz: if temperature changes rapidly with depth, the soil heat flux:
G =- k dT/dz (unit: J m-2 s-I =W m-2)
is large. The proportionality constant, k, is called thermal conductivity (unit:
W m-1 K-1) and describes how well a given type of soil is able to conduct heat.
As heat flows into the soil, heat flux G decreases with increasing depth (and
so does the gradient dT/dz), because some heat is stored in the soil which the heat
flux passes through. The heat storage ~Q in a layer of thickness dz will be equal to
the product of the flux decrease dG across that layer and the time interval at
during which that flux 'loss' dG has been stored. In formula: ~Q/dz = (dG/dz) at.
Substitution of the formulae for heat storage and for soil heat flux in the differ-
ential formula gives:

21
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

which, in turn, gives the equation of Fourier, describing the speed with which
temperature changes at the surface do move down into the soil (example: Figure 3.3):
2
dT k d T
~= pC dz 2
This implies that soil temperature changes are fastest at the depth where the
soil temperature profile is least linear. The combined quantity k/(pC) = K is called
the thermal diffusivity (unit: m 2 s- 1). The absence of heat quantities in the unit is
formally correct.
So we have three soil-specific constants for various aspects of soil heat flux.
For instance, if we are interested in the flux per unit area and unit time as one of
the terms of a surface energy balance, the thermal conductivity k is the relevant
characteristic. In agriculture the resulting temperature changes are important for
the below-ground parts of plants, and also for example to investigate the size of
diurnal temperature changes at some depth we need the thermal diffusivity
k/(pC). The heat capacity accounts for storage effects.
For a soil block, its thermal conductivity k depends upon:
(a) The conductivity of the particles of soil material and their size;
(b) The compaction of the mixture, i.e. the porosity or ratio of space between
particles to total volume and the degree of contact between particles;
(c) The soil moisture.
Because of the last two factors, thermal conductivity is not constant.
Addition of water to a dry soil leads to replacement of air by water in the cell pores
and an improvement of thermal contact between adjacent soil particles, and both
processes result in an increase in conductivity. However, the effect eventually
decreases (Table 3.1).
Table 3.1
Water content Pore volume percentage
Thermal conductivity of sand
(Percentage
for varying pore volume
of soil volume) 30% 40% 50% 60% 70%
percentage and water cont~nt.
0 0.377 0.251 0.168 0.126 0.084
10 1.676 1.173 0.838 0.587 0.461
20 2.179 1.634 1.173 0.880 0.629
30 2.430 1.843 1.383 1.006 0.712
40 2.011 1.550 1.131 0.838
so 1.676 1.257 0.964
60 1.425 1.048
70 1.173

Source: Scharringa (1976). Thermal conductivity unit: W m-1 K-1.

The importance of soil water for temperature changes is also evident from
the variation in heat capacity. · This will increase with its density and with the
specific heat of its constituents. Water has the highest specific heat of all ordinary
constituent materials. Table 3.2 shows that the increase is almost linear, but still
soil-dependent.
There is near-equality between soil types at 50 per cent moisture level. So in
the absence of actual measurements, a (pC)-value of 2 MJ m-3 K-1 is often assumed
for a common range of ordinary soils - but see Table 3.3 for exceptions, e.g. clay
or peat.

Table 3.2 Soil moisture as a percentage of maximum possible capacity

Heat capacity of three soil
Soil 0% 20% - 50% 80% 100%
types for various levels of soil
moisture (MJ ~-3 K-1). Sand 1.264 1.612 2.135 2.655 3.002
Expressed as a volume fraction Clay 1.005 1.495 2.227 2.956 3.446
of water to a volume fraction Peat 0 :620 1.256 2.198 3.144 3.777
of air-space in oven-dry soil.
Source: Sellers, 1965.

22
CHAPTER 3 - THE SOIL AND ITS HEAT BALANCE

In an initially very dry soil, the diffusivity can increase by an order of magni-
tude when a small amount of water is added, because relatively large amounts of
heat are transferred by the evaporation and condensation of water in the pores.
Conductivity increases rapidly and heat capacity gradually, and accordingly
K = k/(pC) increases. With further increases of water content, however, conductiv-
Table 3.3 ity approaches asymptotically to some final value, while the heat capacity
Thermal properties of materials continues to increase. Thermal diffusivity, which governs the transmission of
(%=of pores). Sources: Lowry temperature waves in the soil, thus has a (soil-specific) maximum for some inter-
(1967), Sellers (1965), mediate value of the water content. Typical values of physical and thermal soil
Monteith (1975), Oke (1987) . properties are given in Table 3.3.

Application for Heat Heat Temperature

analysis of: flow storage change
Density Specific Thermal Thermal Thermal
Material heat conductivity capacity diffusivity
p C k pC k I (pC)
Mg rrr3 kl kr1 K-1 W rrr1 K- 1 MJ rrr3 K- 1 lo-6 m2s-1
Ice 0.9 2.14 2.30 1.88 1.22
Still water 1.0 4.19 0.63 4.19 0.15
Still air 0.001 1.01 0.021 0.0011 20.8
Granite 2.6 0.84 4.61 2.18 1.92
Quartz 2.66 0.80 8.79 2.14 4.2
Clay minerals 2.7 1.0 2.9 2.7 1.1
Wood (typical) 0.6 2.1 0.15 0.8 0.2
Dry clay soil (40%) 1.6 0.89 0.25 1.42 0.18
Wet clay soil (40%) 2.0 1.55 1.58 3.10 0.5
Dry sand (50%) 1.4 0.84 0.16 1.26 0.13
Wet sand (50%) 1.6 1.26 1.67 2.01 0.8
Wet marsh soil 0.9 3.35 0.84 2.93 0.3
Wet peat soil (80%) 1.1 3.65 0.50 4.02 0.12
Dry peat soil (80%) 0.3 1.92 0.06 0.58 0.10
Forest litter 0.3 1.84 0.06 0.55 0.11
Old snow 0.5 2.1 0.3 0.9 0.3
New snow 0.1 2.1 0.08 0.21 0.4

It will be noted that the diffusivity of air is relatively high, even when at rest
(due largely to its very low volumetric heat capacity). When stirred, the air diffu-
sivity becomes very much larger, because molecular conduction is replaced by
convective turbulent exchange. The diffusivity of unstable air can be about a
million times larger than that of still air. The effect of this is that maxima in the
diurnal course of temperature above the ground are simultaneous with the surface
maximum, while the maxima in soil temperature have lag times of many hours,
and even months at greater depth.

3.3 SOIL FREEZING, In freezing winter weather, the depth of frozen soil and the speed of soil freezing
AND THE ROLE OF can be directly related to the mechanisms explained above and the role of the heat
SNOW COVER flux constants as a function of soil composition. Assuming that T<0°C at the
surface, the following aspects are important:
References: Ventskevich
(a) Large soil heat capacity slows down change. In early autumn, subsoil is still
(1961), Gray and Male (1981),
warm. Wet soils have relatively large pC-values and so freeze more slowly
Geiger (section 21).
than dry soils.
(b) Freezing of wet soils is slowed down still more by the latent heat of water
freezing. First, this is a temporary energy source; and second, a freezing
layer will be near 0°C in its entirety, so that the temperature gradient dT/dz
across the layer nearly disappears, making the layer a temporary isolator
which blocks heat flow. The same effects play a role during melting in
spring.

23
WMO 1:ECTURE NOTES - AGRICULTURAL METEOROLOGY

(c) Thermal conductivity and diffusivity decrease with increasing organic

content of the soil. Therefore, clay freezes to a greater depth than peat with
the same water content, and soil covered by grassy sod freezes slower than
bare ground. Another complication is that deeper soil is often inhomoge-
neous and layered (section 3.5).
The role of a seasonal snow cover in insulating the ground and any vegeta-
tion against excessively low temperatures is well known. Indeed, it renders
autumn sowing and subsequent overwintering of crops a practical technique in
some regions, where the soil temperature would have otherwise been too low. Of
course, a minimum depth is needed - the protection given by a light covering of
snow, which is easily blown about, is negligible. Observations at Davos in 193 7/38
for snow cover depths of 0.3 to 0. 7 m showed that, even when (snow) surface
temperature fell as low as -33°C, the temperature of the ground surface seldom fell
below 0°C. Effectiveness of such insulation is also illustrated in
Table 3.4, where the snow density, however, is not given.
Table 3.4
Month Thickness of Depth of measurement in soil (m)
Relationship between soil
snow cover (m) 0.0 0.2 0.4 0.8 1.6
temperatures and thickness of
snow cover. Temperatures are November 0.08 0.1 1.9 2.9 4.5 6.0
monthly averages ( C) 0
- (bare -soil) -L6 0.6 2.0 3.8 6.0
(Ventskevich, 1961).
December 0.21 -1.8 0.4 1.3 2.9 4.4
- (bare soil) -6.5 -3.8 -1.9 0.9 3.9
January 0.37 -LS 0.0 -0.-8 2.1 3.4
- (bare soil) -8.0 -6.3 -4.2 -1.2 2.2
February 0.55 -1.6 -0.6 0.2 1.6 3.0
- (bare soil) -8.8 -6.9 -5.2 -2.3 1.3

Figure:3.1 Additional information is included in Figure 3.1, which indicates how soil
Dependence of minimum soil minima at a depth of three centimetres are related to air minima and depth of
temperature at 3 cm depth snow cover.
(tillering node depth) on The thermal conductivity of snow cover depends largely on its density; loose
minimum air temperature and snow in particular insulates well. A frequently used relation was given by Abels
thickness of snow cover · (1892):
(Ventskevich, 1961). k = c p2 , with kin W m- 1 K-1 and ping cm-3 = Mg m-3.
00,__.....,..-..,....--,.----..--"""1 Various determinations of the 'constant' c give values of 2.0 to 2.9, with prefer-
-2 '1---t---+---lf---+---::::~"--I~ .ence for higher values. Assuming c = 2. 9, we get:
-4 ·e <I> p 0.1 0.4 0.8 0.9 (ice)
-6 0..
E k 0.03 0.46 1.88 2.35
-8 1--~--t>""~f-7"'9--::;;,,""1""':;;.-I ~
,_
-10 l-----l,,'----+-,,,"---if,,<'--,C...,,,'b,,<"---,1<( Freezing changes the structure of any soil which is not quite dry. When water
~ - lil,,C----1--,;L---l,,C.~l,£-..,L..,iC-.,,;L.--1----1 freezes in soil pores, the stress forces on remaining liquid water are larger parallel
::::,
~ -14i--,,'-+--L--+-,,~i,,,<-,,'-+---I----I to the soil grain surfaces than perpendicular to them. As a result, small ice lenses
QJ
0..-16 form in the ground. When the freezing front stops advancing, such lenses can
E ·
~ -18 grow and displace ground in a damaging way. On the other harid, when thaw
6 -20----..+---lf---+---I----I comes without rain, then after the disappearance of the ice lenses the ground
:n
. -22 ~'-+-+--+---lr---t--+----t structure is better than before freezi~g.
-24t+-+-tt~-+---lr---t--+----t Meteorological monitoring of ground freezing is a typical subject for temper-
-26 t-+-++---+---lr---t--+----t ature sum application (see section 8.2 for a general discussion on summation
-281+-,1-+--+---lr---t--+----t models). As soon as average day temperatures start to fall below 0°C, their nega-
-301--+--+--+---lf---+--+---t tive values are added daily to the index sum J (e.g. three days with average Tat
-32----------
-5°, +2° and -7° make J = 10). Then approximately the freezing depth will be Zt =
0 10 20 30 40 50 60 Ct ✓J with as soil-specific constant:
Thickness of snow cover (cm)
• Ct =0.06 m K-1/2 day-112 for bare dry sand,
• Ct =0.04 m K-112 day-1/2 for dry grassland,
• Ct =0.02 m K-1/2 day-112 for wet peat .

24
CHAPTER 3 - THE SOIL AND ITS HEAT BALANCE

As approximate empirical correction for insulation by snow cover of thick-

ness d (in m; density?) it was found, that the 'constant' cf should be multiplied
with =(0.1/(0.l+d)).

3.4 DIURNAL AND Periodic fluctuations of soil temperature are caused by the daily and yearly fluc-
ANNUAL VARIATIONS OF tuations in solar radiation received at the surface. The maximum temperature of
SOIL TEMPERATURE the surface is reached when the flow of heat into the soil is exactly equal to the
AND MOISTURE flow outward, and so depends not only on the incoming radiation but also on
heat transfer in the soil and in the air above the surface. Thus it is not surprising
that the incidence of the maximum temperature on the surface occurs some time
after local noon. At night the surface usually continues to cool until the fall of
temperature is checked by the input of solar radiation after dawn, when the curve
of temperature takes an abrupt upward turn. Observations even over a period
limited to a calendar year clearly reveal that:
(a) There is a diurnal variation of temperature to a depth of about 0.5 m, below
which changes become too small to measure with conventional equipment;
(b) This diurnal variation is superimposed on a seasonal variation. Providing
there is sufficient depth of soil, the seasonal variation becomes negligibly
small at depths between roughly 5 and 20 m depending upon the soil type
and condition - an average figure of 10 m would be reasonable to assume;
( c) Plots of both the daily and the seasonal march of soil temperature shows
that:
(i) The amplitude of the fluctuation decreases with increasing depth; typi-
cally, it is halved for each 0.1 m of descent into the ground.
(ii) With increasing depth, the times at which the maximum and the
minimum are registered lag increasingly behind time events at the
surface. This is most clearly shown in the annual curve, and least clearly
for the diurnal minimum.
(iii) Depending on soil type and structure, soil temperatures remain
constant at depths of about 10 metres or more. Ground water with a
water table below this depth has a constant temperature, which is
approximately the annual mean temperature at the relevant station.
The relationship between the three variables of temperature, time and depth
give rise to three methods of diagrammatic representation - two are given in
Figure 3.2.

(a) Temperature ( C) 0
(b)
Figure 3.2 10 15 20 25 30 35 25
Relationships between time,
G
depth and temperature in the ~ 20
soil during the course of a day. 10 ~
::,
E ~ 15
~ ~
0.
..... 20
.r:.
0.
Hour of day
E 10
0
Q)
~
30 5 - . - . - . 1.~0. - .- .- .- .
00 06 12 18 00
Hour of day

In the third diagram type, the vertical (downward) axis is for depth, the hori-
zontal for time, while temperature is plotted as isopleths (see Figure 3.3, overleaf).
The reduction in regular variations is clearly seen: diurnal fluctuations being
markedly less below about 0.2 m; the annual ones markedly less below about 3 m.
Examination of soil temperature data reveals that if there is a well-marked regular
diurnal fluctuation, the time of the daily maximum is relatively constant for
shallow depths (e.g. at z = 0.1 m about three hours after local noon). However, the
time, for any chosen level, at which the daily minimum occurs varies with the
time of year.
Observations at fixed hours may have a different physical significance
according to the time of the year they are taken. At 0.1 m depth an observation at
1400 hours (local time) will generally be at, or near, the diurnal maximum. But the

25
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

minimum is reached some time after sunrise, when the effect of surface warming
penetrates downwards and arrests the fall of temperature. Therefore observation
made at 0800 hours (local time) will be close to the diurnal minimum in early
spring and late autumn - but in summer in temperate latitudes, when sunrise
may be at 0400 hours (local time) or earlier, an 0800 hours observation will take
place well up on the rising portion of the temperature curve.

In addition to the diurnal and annual courses, temperatures in the upper layers of
the soil also show irregular variations due to weather. The exchange of air between
atmosphere and soil cannot influence soil temperature much, because of the low
density (and hence heat capacity) of air compared with that of soil, but the day-
to-day variations of radiation are certainly felt in the upper soil.
Also, a sudden fall of soil temperature can be caused by heavy rain or
showers. When rain or showers occur, the decrease of the soil temperature in the
layers below the water front is more due to the decrease of net radiation than to
water. In heavy soils, the percolation speed is only about 0.2-0.3 mm h-1, so only
the uppermost soil layers can be cooled directly by rain water. But in sandy soils,
where water is able to percolate faster, a cooling effect of rain is frequently meas-
urable, especially after dry spells. The cooling effect of rain lasts often as long as
the soil moisture and thus the heat capacity of the soil is changed.
The diffusivity K for a good moist horticultural soil will be about 0.45 mm2 s-I,
that for a peat soil about 0.15 mm2 s-1, and for dry organic matter 0.1 mm2 s-I. In
so-called 'good' soil, temperature fluctuations reach deeper levels, and more
rapidly, than in peaty soils or in organic mulch. Also, given the same net energy
at the surface, the surface and near-surface temperature of the dry organic layer
(i.e. the layer with poor thermometric conductivity) may reach higher maximum
and lower minimum than the 'good' soil.
Soil profiles often exhibit markedly different layers. On a daily basis, calcula-
tions based on homogeneous soil characteristics may not be useful when, for
example, the topsoil has a markedly higher organic content than the subsoil.
Similar problems also arise in the analysis of seasonal fluctuations, e.g. when a
shallow layer of soil lies on rock, or when a water table occurs at a few metres depth.
Geiger (1965, section 19) gives observations by Yakuwa (1945) for a typical
Figure 3.3
sunny day, which is shown in Figure 3.3 (data are given in Table 3.5). Here we have
Temperatures in three soils on
a situation with a two-layer soil. In the topmost 0.05 m layers, the value of K
15 August 1929 (Yakuwa,
increases from 'sand' to 'loam' to 'clay', and as expected the surface temperature
1945). Spaces between
maxima decrease in this order. But in the layer below, the conductivity order is
5° isotherms are alternately
reversed. Therefore the high 'sand' surface temperatures are conducted down to
clear and shaded for
greater depths, while the 'loam' and 'clay' remain comparatively cold.
comparison, and a 22.5°
When the soil is covered with vegetation, the upper side of the canopy forms
isotherm (dashed) is added to
a new surface, which absorbs a considerable portion of incoming radiation. The
illustrate the deep temperature
remaining part of radiation goes .through the plant cover and is absorbed by the
. situation.
lower leaves, the rest by the soil surface. The radiation reaching the soil surface
below a dense plant cover can be diminished to a few per cent of the total. The
amount of reduction depends on the so-called 'leaf area index', i.e. the ratio of the
20 whole leaf area (one side) to the soil surface (see section 2.3).
The transport of heat absorbed at the soil surface below plant cover occurs in
40 ~ - - - - - - - - - - the same way as with bare soil. But in daytime, some heat also goes down from
E o -- the cariopy to the soil surface by turbulent transport of the air within the plant-
~
£a.
------ - air layer. Hence, while the temperature of the covered soil surface is lower than the
- - 'J
~ 20 temperature of a bare surface during daytime, it is not decreased as much as the
"O
C

e 40 Loam incoming radiation is decreased . .

0 --- -------
The temperature at the surface (at z = 0) is frequently not known. For the

~ ~ ~~ ~~ ~
annual temperature wave, De Vries (1958) ·uses the ordinary screen air tempera-
ture, and suggests:
Temperature range at surface
{ --- -_ S~L- -- - Range of air temperature at 2 m
= 1. 1 to 2.0 (average 1.3) for bare soil
4
0Oh 3h 6h 9h 12h 15h 1gh 21 h 24h = 1.25 for short grass.
l+M MaMbMcl No such relationship is suggested for the daily cycle.
lW 2W 3W 4W SW

26
CHAPTER 3 - THE SOIL AND ITS HEAT BALANCE

Within the soil evaporation, condensation and movement of water vapour

can take place because of the presence of air in soil pores. In these pores, the
relative humidity is permanently close to 100 per cent, and this implies that the
vertical temperature gradient in the soil generates a corresponding water vapour
pressure gradient. The consequence of this is that vapour will diffuse from warm
soil to cold soil. In daytime, the vapour moves downward to the relatively cold
subsoil, and at night it will move upward. At the cold night surface the advected
vapour will then condense, and this distillation process is one of the sources of
dew (see section 4.6).
Liquid water motion in the soil, apart from gravitational drainage of recent
precipitation, occurs chiefly by way of capillary motion and is therefore
dependent on soil structure. In most soils, capillary motion can transport water
up to a few meters above the water table with relative ease; above that, it
becomes difficult.

3.5 A MODEL OF SOIL A simple theory, which accounts for much of the variation and leads to results of
TEMPERATURE DIURNAL practical value, is based upon the following assumptions:
COURSE AT DIFFERENT (a) The changes of temperature imposed on the surface follow two simple sine
DEPTHS curves - the diurnal changes being due to the daily course of the Sun, and
the seasonal one due to the seasonal variation of short-wave (solar) income;
(b) The thermal diffusivity (K) of the whole soil block concerned is constant,
both with depth and with time of day/year.
This method assumes that the variation of temperature at the surface (z = 0),
and at some time t within the period, is given by T(O,t) = <T>+A0 sin (wt), where:
<T> = mean temperature over the period;
w = 2 11:/period (i.e. an angular velocity);
A0 = half-range of temperature, i.e. (maximum - minimum)/2.
The model implies that the mean temperature over the period is the same at
all depths. What varies is the daily (or annual) fluctuation about this mean figure.
Suppose A(z 1) = amplitude (i.e. half range) at depth z1;
A(zz) = amplitude (i.e. half range) at depth z2>zi, then:

Ac,,>= Ac,,> -exp r-(z -z ✓~;1

2 1)

where P = period of oscillation (in seconds) for 24 hours and for 12 months in the
two cases being considered. If we write w = 211:/P (w being an angular velocity), an
alternative expression is:

Ac,,>= Ac,,>· exp r-(z -z,)

2 ~l
At the surface z 1 = z2 = z0 and Ac2 r-~A0 , thus the range 2A(z) at any depth below
the surface is given by:

A,=A 0 exp[-,~l
The time (t) at which the maximum is reached at depth z is given by:

t= _!_(2: + z
(l) 2
/ro)
~~
The speed of penetration of the temperature maximum into the soil is w ✓2K.
The time interval between the maxima at a depth z and that at greater depth is
given by:

In this analysis the difference between soils is expressed completely by K, that

between periods by w. The parameter ✓ (2K/w) is usually called damping depth.

27
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

Table 3.5
Depth Type of soil
Transmission of temperature
Sand Loam Clay Bog
fluctuations into four different
soil types in sunny weather. K (mm 2 s- 1) 0-0.05 m 2.2 3.2 3.6 2.8
0.05-0.3 m 4.4 2.8 2.6 1.1
Maximum surface 53.5 46.7 35.9 39.0
0
temperature ( C)
Minimum surface 13.5 13.1 14.4 15.8
temperature C1C)
Diurnal temperature .fluctuation surface 40.0 33.6 21.5 23.2
0.05 m 19.4 18.5 13.7 13.9
0.1 m 12.3 10.7 7.7 5.4
0.2m 4.8 3.0 2.2 0.73
0.3m 1.6 0.7 0.6 0.3
Depth (m) at which diurnal
fluctuation is reduced to 0.1 ° 0.57 0.47 0.47 0.40
Rate of motion (h / 0.1 m) of 3.30 3.95 3.94 5.31
temperature wave

We discussed the observations by Yakuwa (1945) for a two-layer soil in section 3.4,
and the corresponding data ar.e 'given in Table 3.5. The low value of K for the bog
soil and the lower surface amplitude are both attributed to moisture in the
organic layers.
It is interesting to check the validity of the theory given here on these data,
assuming a uniform thermal diffusivity. if we do this for the two K-values for sand
on the diurnal ranges ~T (K), starting from 2 A0 = 40° at the surface, noting depths
in cm, we get the following re·sults:
K (mm2 s- 1) ✓ 7t /{PK) As/Ao A10/Ao Azo/Ao ~Ts ~TIO ~Tzo
2.2 0.1285 0.526 0.276 0.021 21.0 11.0 0.8
4.4 0.091 0.643 0.403 0.065 25.4 16.1 2.6
Compare this with the values tabulated above.
If the proce'SS was that of temperature movement into a homogeneous
block, the ratio between successive amplitudes at equal increments of depth is the
same, i.e.:
A A
_2
= -Z3- .
etc., when (z 4 - z3 ) =(z 3 - z2 ) = (z 2 - z1 ).
AZ 1 AZ 2
The assumption of a single annual sine wave renders this method inapplica-
ble in polar and equatorial regions and - on the daily basis - in the absence of
an adequ.ate daily range (as occurs on overcast days in winter). In fact, the daily
fluctuation near the surface (say, the top .10-15 cm) is rather inadequately repre-
sented by a simple sine wave, and a two-term harmonic has been found more
appropriate. The seasonal variation in the time of the daily minimum is one
reason for the· difficulty, and the frequent absence of a marked diurnal fluctuation
,in the 'winter half-year' a second reason. The second reason also rules out the use
of the method where there is extensive or prolonged snow cover.

Review questions o~ Chapter 3

A. Questions on basic understanding:
3.1 Mountain climbers lost in the snow may improve their chances of
survival by burying themselves under the snow. Suggest reasons for this
advice.
"The best protection for winter crops is a loose, freshly fallen snow
cover." (Ventskevich, 1961, p . 50). Discuss this claim.
3.2 Geiger (section 19) quotes the following observations (August 1924) of
temperature at stated times and depths:

28
CHAPTER 3 - THE SOIL AND ITS HEAT BAIANCE

Time of day about 5 a.m. about 3 p.m.

Condition of ground Firm Loose Firm Loose
Surface 1 l.6°C 9.6°C 3l.0°C 36.4°C
5 cm depth 13.8°C 12.4°c 27.6°C 29.0C
10 cm depth 1s.4°c 16.4°C 24.2°C 23.8°C
Explain these results in terms of the concepts mentioned in section 3.2.
3.3 Under certain circumstances, snow is referred to as a 'black body'.
Explain and justify this description.
3.4 Discuss the effects of cold rain on soil temperature.
3.5 Geiger (see section 21) reports that the buried walls of houses in a
medieval village have been located by the pattern of hoar frost in aerial
photographs. Explain this phenomenon.
3.6 To reduce the risk of frost damage in grassed orchards, it is often advised
that the grass be removed and the soil surface left bare and preferably
consolidated (i.e. not loose but compact). Justify this advice.
N.B. In many cases this advice might be overruled in view of other
advantages arising from a grass surface, e.g. in the reduction of erosion
by water.
3. 7 How much heat energy will flow by conduction in three hours through
a column of wet marshy soil with a cross-sectional area of 10 cm2 and a
depth of 5 cm, if the top of the column is 5.5 K warmer than the
bottom?
(This question assumes that a 'steady-state' situation has been reached
and is maintained. Let k = 0.84 J m-1 K-1 s-1.)
3.8 Show that the speed with which the annual temperature wave pene-
trates into a homogeneous soil block is about one-nineteenth of that of
the diurnal temperature wave. Show that the depth at which the ampli-
tude is reduced to one per cent of that of the surface is given by 4.61 I!!,
where I!!= ✓(2K/w) or ✓(KP/w). (I!! is often called the 'damping depth'.)

B. Questions on extended application:

3.9 Geiger quotes the following data (1939/1940):
Type of soil Depth of frost Speed of penetration
(cm) (cm/day)
Humus 32 0.6
Loamy sand 40 1.1
Loam 52 1.1
Sand 52 1.7
Are these results in general conformity with the data in Table 3.3?
3.10 Taking homogeneous soil blocks with values of (a) k/(pC) = 2.2 x 10-7 m 2
s-1, and (b) k/(pC) = 4.4 x I0-7 m2 s-1, find the depths at which a diurnal
surface temperature range of 40 K is reduced to 0.1 K. Compare these
results with that for the two-layer soil block referred to in Table 3.5 for
'sand'.
3.11 Confirm the relationship:
AZ 3
=- - etc., when (z 4 - z3 ) = (z 3 - z2 ) = (z 2 - z 1 ).
Az2

To what extent are the observations in Table 3.5 consistent with the
above relationship?

C. Questions involving practical work or local data:

3.12 Obtain (or initiate) continuous records of soil temperature at depths
between about 5 cm and SO cm or 100 cm. Examine the relationship
between the magnitude and timing of the daily maximum, daily

29
WMO 'LECTURE NOTES - AGRICULTURAL METEOROLOGY

minimum and daily mean and the temperatures observed at specific

times of observations, e.g. at the standard climatological observing hours.
Classify the results according to the time of year and synoptic situations,
e.g. clear or cloudy weather, dry or wet periods, and formulate empirical
relationships between the standard climatological observation and other
features of the course of diurnal temperature.
3.13 In areas of seasonal freezing it has been found that the depth of freezing
is related to the 'degree days' (in K) of air temperature below 0°C. Start
from a suitable date at the beginning of the freezing season (e.g. the date
when mean daily temperature has been~ 0°C for three successive days):
(i) Examine the way in which soil temperature at several depths
changes in response to the cumulative 'degree days';
(ii) For as many years as possible, compute the seasonal total of
'degree days' (T~ 0°C), and relate this to the maximum depth of
seasonal freezing.

30
CHAPTER 4
WATER AND THE HYDROLOGICAL CYCLE IN AGRICULTURE

4.1 WATER AND The basic processes involved in the hydrological cycle are reviewed in the simpli-
VEGETATION fied presentation of Figure 4.1. As will be seen later, the properties of the Earth's
surface significantly modify this simple picture of water circulation.
Almost every process occurring in plants is affected by water. However, four
main functions can be distinguished in which water plays a major role:

, ,,
,'ll/,,' l,:t.,, ,' t I t ,

Figure 4.1 (a) Water is the major constituent of physiologically active plant tissue;
The hydrological cycle (b) Water is a reagent in photosynthesis;
correlated with the air-mass (c) Water is the solvent in which salts, sugars and other solutes move from cell
cycle (Trewartha, 1954). to cell, and from one part of the plant to another;
(d) Water is an essential element for the maintenance of plant turgidity
(hydraulic inflation) necessary for cell enlargement and growth.
As with most biological conditions, the availability of water has an optimum -
there can be too much, as well as too little water for good progress of the above
processes. An example of this is when, after a long dry period, sudden rain may
lead to bursting of cells. The roots have been working hard to provide water from
the soil and do not immediately stop doing so, leading to excessive cell turgor
because evaporation is no longer proceeding at the original rate.
In reviewing the physiological significance of internal water relations on crop
yields, Slatyer (1969) proposed that only two valid generalizations could be made:
(a) The growth and development of most crops can only proceed fully unim-
paired, and crop yield is only maximal when high water status is maintained
throughout the life of the crop;
(b) The harmful effects of water deficits are usually most pronounced in tissues
and organs which are in the stages of most rapid growth or development.
The second proposition implies that, as far as economic crop yield is concerned,
sensitivity to water stress varies in the course of time. In this context it is very
interesting to note that with some crops it is a standard practice on farms to
impose stress at certain growth stages, with a view to influencing quality devel-
opment and improving the economic (though not necessarily the maximum)
yield - an interesting conflict between quality and quantity. Other existing
dangers of over-irrigation (see section 6.5) make Slatyer's first generalization quite
doubtful. Apart from the physiological difficulties in assessing effects, it seems that
water/crop yield relationships can only be simple in the dry regions of the world.

31
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

Figure 4.2 shows the effect of total annual rainfall on the yield of wheat in
France and Israel. Geslin and Bouchet (1966) suggest an inverse relationship
between seasonal rainfall (above a rare minimum) and yield in the moist Paris
basin region. Maximum yields occurred during years of below-normal rainfall. By
contrast, Lomas and Shashoua (1973) suggest a linear relationship under arid
conditions. Thus the response of agricultural crops to soil moisture varies in the
different climatic regions. In arid areas, where water demand exceeds water supply
and is the yield-determining factor, the simplest water-supply indices have proved
successful to estimate yield. In areas with more adequate water-supplies, a more
sophisticated approach is required (see section 4.6 and section 8.4).

Despite the great range in form, all types of vegetation share certain charac-
teristics that are important in relation to water - yield control. Vegetative cover
Figure 4.2 is made up of three parts: the canopy of living and dead stems and leaves that
Relationship between wheat stands clear of the soil, the accumulation of dead and decaying plant remains that
yields and total seasonal lie on or in the soil surface, and the living and dead roots and subsurface stems
rainfall in an arid region that permeate the soil. The canopy above the ground and the surface litter accu-
(Israel) and a humid region mulation operate as barriers interposed between the atmosphere and the soil.
(France). They intercept precipitation and thus temper the delivery of water to the soil, and
o Paris basin, France, 1900-1936 they retard wind movement near the soil surface. The heat-insulating properties
(Geslin and Souchet, 1966). of canopy and litter reduce the interchange of energy between soil and atmos-
• Mishmar Ha-Negev, Israel, 1948- phere - this slows the evaporation of water from the soil in daytime, and may
1968 (Lomas and Shashoua, 1973). increase it at night. This effect of canopy and litter, however, may be offset by loss
of soil water through transpiration of living parts of the canopy.
2 500
At the soil surface, vegetation also obstructs the overland flow of water and
the pick-up and transport of soil material. This is accomplished by stems and litter
2 000 accumulations that form miniature anchored dams and diversion walls, and by
1, roots and decaying organic matter which obstruct the soil or bind it.
.c
g Beneath the soil surface, to depths reached by roots, vegetation acts both
~ 1500 directly and indirectly. Roots, by their growth and subsequent death and decay,
·s:,
"'
Cl>
permeate the soil with material that both binds it and aids the flow of water
.c
3: through it. Roots also absorb water and carry it to the above-ground parts of plants,
1 000
thus contributing to the return of water to the atmosphere. The parts of vegetation
that are below the soil surface act indirectly by creating the favourable environ-
5.0 0
ment for the activity of small animals, insects, fungi, moulds and bacteria. To all
these organisms vegetation means shelter and favourable living conditions; to most
of them it means food as well. Some of them, like worms, burrow in the soil and
thus aid in keeping it permeable to water. Others break down organic matter and
0 200 400 600
Seilsonal· rainfall (mm)
thus perform the same kind of function, while at the same time permeating the soil
'with the binding materials of their own bodies.
This brief discussion provides some indications of the hydrological effects
exerted by vegetation itself. It suggests that vegetation affects water yield by
slowing or diverting surface flow, by inducing water to enter the soil, and by influ-
encing water losses through evaporation and erosion processes. In order to
describe these hydrological effects in greater detail, attention must be focused
upon a thin layer above and below the Earth's surface, namely the layer that lies
between the tops of plant crowns and the lowermost tips of plant roots. It is
within this layer that vegetation interacts, in the fullest sense of the word, with
hydrological and meteorological processes.
Some of the processes undergone by water between its precipitous delivery
to Earth and its release into streams and ground water basins (moving from the
sphere of meteorology into that of hydrology) are the following:
Interception: Precipitation is divided into a part delivered to the
soil surface and a part caught on vegetation and
evaporated into the atmosphere.
Evaporation: Returns water to the atmosphere from water
surfaces and the soil.
Transpiration: Returns water to the atmosphere from the soil
through plants.

32
CHAPTER 4 - WATER AND THE HYDROLOGICAL CYCLE IN AGRICULTURE

Sublimation: Returns ice- or snowpack-water directly to the

atmosphere.
Melt: Delivers snowpack-water to the soil surface.
Infiltration: Divides water reaching the surface into surface
flow and soil water.
Surface flow: Conducts water not absorbed by the soil, and
water diverted to the surface after absorption, to
streams and the sea ('runoff').
Accumulation: Determines the quantity and areal distribution of a
snow pack.
Storage and drainage: Retains water and carries it through and below the soil.
Erosion: Provides sediment load in gullies and in channels,
by detachment and transport of material on the
soil surface.
Deposition: Separates the sediment load from flowing water.
Here we are dealing mainly with the first four processes of this list. For land
surfaces, the mixed term 'evapotranspiration' describes the joint action of evapo-
ration and transpiration in moving water from soil to atmosphere.

4.2 MOISTURE Types of soil differ in the size, shape and density of packing of their particles and
CHARACTERISTICS OF aggregates, and therefore in the amount and distribution of spaces ('pore spaces')
SOILS into which water and air (and also plant roots and animals) can move. Moreover
there are differences in the physical and chemical properties of the surfaces of
various soil materials, especially in their ability to absorb and release moisture.
Such properties vitally affect the movement and retention of water in the soil.
Therefore available soil water cannot be assessed sufficiently only in terms of the
amount of water per unit volume of soil block, or per unit weight of 'undisturbed'
soil or per unit weight of oven-dried soil. Different soil types will release different
amounts of water to crops, even if their total soil moisture content S (expressed as
a percentage of dry weight) is exactly the same.
The state of soil moisture can be conveniently classified into four categories:
(a) Gravitational water, which occupies the larger pores of the soil and drains away
under the influence of gravity. An excess of this water can damage plants if
drainage is too slow, for instance by choking the roots;
(b) Capillary water, which is held by surface forces as film around the particles, in
angles between them and in capillary pores. In the form of liquid, capillary water
moves slowly from thicker to thinner films and along moisture-tension or diffu-
sion-pressure gradients. It can also move in the form of vapour. Capillary water is
the only important source of water for most cultivated plants;
(c) Hygroscopic water, which is held as a very thin film on the particles by surface
forces. This water is held so firmly that it can move only in the form of vapour.
The moisture remaining in air-dry soil is usually regarded as hygroscopic water
and is not available to plants;
(d) Water vapour, which occurs in the soil atmosphere and moves along vapour-pres-
sure gradients. It probably is not used directly by plants.
The movement of liquid water through soils occurs predominantly under the joint
action of gravity (unidirectional) and surface-tension forces (all directions), asso-
ciated with the degree of curvature of water films around and between particles.
Accordingly, these forces depend upon the size and the shape, i.e., the geometry
of the pore spaces. Other forces, hydrostatic and osmotic, may also operate (see
Lowry, 1967; 1989).
If water is added to a block of otherwise 'dry' soil, some of it will drain away
rapidly through any relatively large cracks and channels. The remainder will tend
to displace some of the air in the spaces between particles, beginning with the
larger pore spaces. Broadly speaking, a well-defined 'wetting front' will move
downwards into the soil, leaving an increasingly thick layer retaining all the mois-
ture which it can hold against gravity; that soil block is then said to be at field
capacity. The drying of soil shows similar behaviour, with a relatively sharp 'drying
front' moving down into the soil as evaporation proceeds from the surface.

33
WMO LECTURE NOTES-AGRICULTURAL METEOROLOGY

Therefore, after a period with wet and dry spells, the soil block will tend to have
alternate wet and dry layers. The state of field capacity, generally desirable for agri-
culture, must not be confused with the undesirable situation of 'saturated' soil,
where all the pore space is occupied by water (see section 9.3.1).
When the moisture content falls below field capacity, the subsequent, rather
limited, movement of water in the soil is largely in the vapour phase, and this
occurs mainly by distillation, related to temperature gradients in the soil
(section 3.4). Plant roots within the block will, however, extract liquid water from
the water films around soil particles with which they are in contact. Hence it is
clear that more water can be lost from a soil block carrying vegetation than from
another block, identical in initial water status but without vegetation. As water is
extracted and the moisture content drops further below field capacity, a point is
reached at which the forces holding the moisture films to soil particles cannot be
overcome by root suction - plants are starved of water, lose turgidity, as the soil
moisture content has reached the wilting point.
What is important for the plant is not the absolute amount of water content,
the volume percentage in a unit volume of soil - but rather how much is avail-
able to the plant, and this is strongly linked to surface tension forces in the soil.
This 'available water' is specified by the soil moisture potential \If, which is the
suction energy (as measured with a tensiometer, see section 9 .3.2) that roots must
exert to extract water from the soil. The appropriate unit for \If is J m-3, which
usually is translated into a unit of pressure called pF, equal to the logarithm (base
10) of the (negative) head of water expressed in cm.
For practical purposes the wilting point is reached for all species when the
forces attaching water films to the soil particles reach a value of about 15 atmos-
pheres or 1.52 MPa = 155 m water column ➔ pF = 4.2. This corresponds to
capillary action in spaces of less than 1 µm width, while most root-hairs are at least
a few µm in diameter. So the usual location of the wilting point at pF = 4.2 makes
physical and biological sense.
Another major moisture level, field capacity, corresponds in actual field soil
situations to pf-values between about 2.0 and 2.5 - not quite precise, since grav-
itational drainage of water does not cease abruptly. Therefore, conventionally,
field capacity is defined to lie at pF = 2.0 (i.e. suction of 104 Pa). The 'readily' avail-
able water for use by plants is then the volume percentage difference of pF = 2.0
and pF = 4.2, and this varies strongly with soil structure - see Figure 4.3. Extreme
cases in this graph are the sand dune, where only 9 per cent of its volume can hold
available water, and peat whose volume can contain 44 per cent of available water.
When water supply is adequate1 as under irrigation, most of the water will be
extracted from the upper 0.3 m of the soil. Only in a limited water situation does
'more water extraction occur from the deeper soil layers (0.6-0.9 m). Under normal
conditions, wheat will develop a root system that can exceed 1.5 m depth, while
Figure4.3 7 Sand dune
Relations between soil moisture Light sandy clay
(in percentage of total volume) Humoussand
and its soil moisture potential Heavy day
\If for various types of soil.
Peat

.... Available
.... . . soil•
' ·:.,.. moisture

Field capacity
....
···....______\ '
0
' \ Saturation
10 20 30 40 50 60 70 80 90
Volume % water ·

34
CHAPTER 4 - WATER AND THE HYDROLOGICAL CYCLE IN AGRICULTURE

rice crops have a much shallower root system not exceeding 0.5 m. However, most
roots of the wheat crop are to be found in the first 0.6 m, and those of rice in the
first 0.3 m, so that a majority of the roots are in the near-surface soil layers.
Robertson, et al. (1980) reported 39 per cent of the total rooting length in the top
0.15 m of soil, and 55 per cent in the top 0.3 m.

4.3 DETERMINATION Hydrologists do not always measure evaporation in their mass-balance studies of
OF WATER LOSS FROM the water economy of a catchment, but consider it as the remainder term in a
LAND SURFACES water balance, of which the other terms (precipitation, run-off, subsurface-flow)
have been assessed. Evidently, the difference between several rather uncertain
References: WMO TN No.83;
hydrological processes cannot, to put it mildly, be determined to high accuracy.
Lowry (1967; 1989); GM II,
For actual determination of evaporation, the first hydrological alternative has
Ch. XII; Brutsaert (1982).
been to use the meteorological data which are available everywhere, namely
monthly average temperatures - sometimes adding humidity information. For
details on such methods by Thornthwaite (1948), by Blaney and Criddle and by
Turc, reference is made to hydrology handbooks.
Here we review basic meteorological alternatives for estimating water loss
from the surface by evaporation from bare soil or open water, and by transpiration
through the plants. Combinations of these alternatives are discussed in
section 4.4.

4.3.1 Fundamentals of the The phase change of liquid water into water vapour requires energy, i.e. the latent
evaporation process heat of vaporization which is about 2 500 kJ of heat per kg of water (at T =20°C).
Given this energy, evaporation can take place, given, of course, that evaporable
water is available and that the resulting vapour is removed. Evaporation is a three-
stage process, and all three are essential to the process.
Further details on the first stage, the supply of water from below the soil
surface, are given in Chapter 3. Water moves upwards in the soil, against gravity,
by two main processes: capillary rise and distillation. The effectiveness of both
processes depends upon the soil structure, and distillation also depends on the soil
temperature gradient - when the surface is colder than the deep soil, then soil
water vapour moves upward, to condense again at the surface. Let us assume that
water is available at the surface, either as a result of these two processes or because
of recent precipitation, and look at the next two stages, vaporization and atmos-
pheric transport.
Unusually critical comments are needed in this context, because popular
concepts on the evaporation process contain three major misconceptions, namely
that:
(A) High temperatures imply high evaporation;
(B) Evaporation is proportional to the vapour pressure deficit, or to the relative
humidity, as measured in the Stevenson screen;
(C) Evaporation is proportional to the (daytime average) wind speed.
These statements are at best half-truths, indicating a correlation but not a
causality. In a number of non-trivial situations, they are just plainly wrong.

For the second stage of evaporation, the phase change of liquid into vapour,
we need heat, which is not the same as a high temperature. Temperature is a
measure of the kinetic energy of average single molecules - heat is the total
energy of all molecules of an object. A cubic metre of air contains 4 000 times less
heat than an equally warm cubic metre of soil (section 3.1).
Net radiation is the principal source of energy for evaporation. This means
that evaporation from the land is usually negligible at night. Also in dense shade,
e.g. deep in a dark forest where solar radiation barely penetrates, all surfaces are
gaining and losing long-wave energy at the same rate, i.e. there is radiative equi-
librium. The only sources of energy for evaporation are then upward soil heat flux
and the heat of the air itself. In the latter case, the air will be cooled noticeably, a
similar process as in the so-called oasis effect (see section 4.4). Over a lake, the
necessary heat can be obtained from the water body below, which means that
evaporation there can be as strong at night as in daytime.

35
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

If there is no surface moisture to evaporate, incoming radiant energy must

be diverted to other energy sinks - in particular by heating the ground until its
temperature has become sufficiently high, that surface radiation emission
can balance the short-wave input. Heating of the air over such dry surfaces
occurs, of course, but is not sufficient as a sink because of the low thermal
capacity of air.
Clearly, this is one of the cases where statement (A) is wrong: surface temper-
ature is high, but no evaporation. Another reason why (A) is unreliable is, that at
high temperatures the vegetation may close its stomata, blocking evaporation.

Evaporation is stronger when the surface is wet and the air is dry than in the
opposite case. However, there are misconceptions on how this is quantified. When
it is stated, that "the driving force (of evaporation) is the vapour pressure difference
between the surface and the overlying air", such a statement is a misleading half-
truth. The actual measure for the humidity difference which drives evaporation is
the vapour pressure gradient, the difference divided by a distance. This may seem
trivial, but if under given wet surface conditions a certain (low) vapour pressure e(a)
is measured in a 2 m high Stevenson screen (as is customary in France or Norway),
the evaporation is likely to be only two-thirds of what it would be if the same e(a)
is measured in a 1.2 m high sci:een (as is customary in England) - other conditions,
like radiation and wind, being equal. In this context, the ecological habit to omit
any specification of humidity observation height is downright misleading.
It is true that evaporation is proportional to the difference [e(s)-e(a)], where
e(a) occurs at a fixed screen height and e(s) is the vapour pressure at the surface.
However, this formula is quite often interpreted as a vapour pressure deficit, i.e. as
the difference at a single level between actual and saturation vapour pressure. This
error may arise either because of misreading (taking 's' to mean 'saturation'
instead of 'surface'), or because of the assumption that surface vapour pressure is
always at saturation point - it isn't. If a moderately dry breeze, for example, blows
over very dry sand, condensation may occur at the sand surface instead of evapo-
ration. The only 'driving force' of evaporation is the strength and direction of the
vertical gradients of water vapour pressure. Blindly following misconception (B)
may be punished by biased or just wrong answers.

The third stage of the evaporation process is removal of the water vapour by
the atmosphere. Some mechanism to aid vapour transport is needed and is
provided by the movement .of the air - but by turbulence rather than by the
average wind speed. This means for instance, that a given average wind speed will
cause more transport over a rough surface than over smooth flat ground. It also
·means that in strongly radiative noontime conditions the transport will be high
because of the small convection eddies, even when in such weather the h.ourly-
average wind is very weak. Behind a windbreak of height H, evaporation will be
relatively strong in the turbulent wake, typically at downwind distances of
between 2 Hand 12 H (Marshall, 1967; section 6.4).
Various evaporation formulae - in particular those of the 'bulk aerody-
namic' type, and also most of the practically used combination types - have an
aerodynamic term, which increases linearly with average wind speed. In regions
with sustained high climatological wind speed (e.g. north-west Ireland) this can
lead to unrealistically high values of cakulated evaporation. Qualitatively it
appears reasonable that, once the wind speed is sufficient to cause a vigorous
turbulent exchange process within and above the crop stand, further wind
increases can have no effect in raising evaporation beyond the biologically
sustainable water transport in the crop. For short crops evapotranspiration
is unlikely to increase when average wind increases beyond 5 to 7 m s-1. However,
it must be recalled that wind speed within a crop is not the same as that above the
crop, and therefore upper limits to the 'maximum effective wind speed' vary
according to the nature of the evaporating surface.
It is customary to measure evaporation in terms of the depth of water lost
from a unit horizontal surface (analogous to the measurement of precipitation).

36'
CHAPTER 4 - WATER AND THE HYDROLOGICAL CYCLE IN AGRICULTURE

Thus if E mm of water is evaporated, the energy required per m2 surface is LE,

where at ordinary temperatures L ==2.5 x 106 Joule kg-I (==590 cal g-1). In strongly
evaporative conditions, the radiative energy available for evaporation is about
500 W m-2, which means that in such conditions 1 mm of water will take between
one or two hours to evaporate.

4.3.2 Existing methods to In agricultural meteorology, particular attention is directed towards determining
determine evaporation the loss of water, due to evaporative processes, from crops of all types over
surfaces of varying degrees of wetness. Given the above review of the three stages
of evaporation, the methods used in agricultural meteorology can be considered
under four headings:
(a) Surface water loss measurement. Observations are made of water loss from a
device, such as a pan of water or a surface kept wet by a supply of water from
an attached 'reservoir', that can be monitored (section 9.2.7);
(b) Measurement of energy used for evaporation ('energy budget method', section
4.3.3). Physically-based relationships are used in which evaporation is
regarded as one of the items in the surface energy budget;
(c) Atmospheric vapour transport measurement ('aerodynamic method', section
4.3.4). Physically-based relationships are used, in which evaporation is
regarded as a process of diffusion;
(d) Combined application of methods (b) and (c) ('combination method', section
4.4). Joint application of (b) and (c) eliminates the need for hard-to-get
surface information.
All methods also have simple operational versions. To develop these results
careful experiments to determine evaporation need to be undertaken at well-
equipped research stations, and care needs to be taken to empirically relate these
experiments to standard climatological measurements. The operational formulae
obtained are often only useful in similar climates.

We will begin here with a look at method (a), the measuring of actual water
losses from a test-body. The apparent simplicity of measuring evaporation in this
way is illusory. The main difficulty is to obtain a reading which is representative
of the area or situation which the instrument is designed to monitor.
Direct measurement of the water loss by extensively monitoring a part of the
actually investigated terrain soil is called lysimetry (see section 9.2.7). This method
suffers only moderately from difficulties arising from differences in exposure
between test body and represented area. However, the work and costs required to
accurately monitor the actual water content of a few cubic meters of terrain are so
extensive that operational application is not possible - even for the simplest
version, measuring the change in weight of a sample of soil (there, too, the ques-
tion of getting a representative sample still remains). In some forms of lysimeter,
arrangements are made to keep the surface fully wetted, i.e. evaporation is not
restricted by water shortage. In others, the soil block involved is left exposed to
the atmosphere without interference, similar to the situation in the surrounding
terrain. The first measures a 'potential', the second an 'actual' loss.
The alternative approach to surface water loss determination is to measure
water loss from some artificial vessel, which is exposed in the field. Evaporation
pans and atmometers fall under this heading (section 9.2.7). Given some measured
'open water' evaporation from a pan, we still require some method to find the
corresponding water loss from a soil surface - usually some pan coefficient or
correction factor. Several reviews have been published of conversion factors
between evaporation from various sources and the readings of pans, in particular
for the United States Class A pan (see references in section 4.4.1). The very large
variations of pan coefficients (e.g. 0.22 to 1.56 from the Lake Hefner results)
underline the fact that a few square meters of open water are not really represen-
tative for most field situations. Pans, and certainly atmometers which are smaller
still, have a microclimate of their own, are heavily influenced by the presence and
physical properties of the water container, as well as being 100 per cent humid.
They are mainly useful for comparative purposes.

37
WMO LECTURE NOTE.S - AGRICULTURAL METEOROLOGY

4.3.3 Energy balance In the case of a crop, the most general energy balance formulation is:
estimation of evaporation Q* + A = H + LE + G + g + p

Four components are operative in almost all situations (arrows indicate the
direction of a flux for which the sign is taken positive), namely:
Q* J, Net radiation;
H i Flux of 'sensible' heat (generally turbulent) from ground to air;
LE i Evapotranspiration (unless there is no water to evaporate);
G i Flux of heat into the ground. ·
Other components which may possibly be present include:
g Heat storage in the crop is present, if we deal with vapour transfer in a
vertical column of atmosphere, but not if we are concerned strictly with
the vertical flux at the surface;
P Energy requirements for photosynthesis exist whenever vegetation is
present, although on a long-period-average basis . this process absorbs
rather less than 1 per cent of incoming energy, so it is frequently
regarded as negligible.
A .J, The second term on the left-hand side is the advection term, the hori-
zontal divergence of sensible and latent heat. This term gives rise to
much confusion at elementary levels of analysis. At this point we must
assume that it is n@gligible, and it will be discussed at more length below.
On the surface, the general balance equation reduces to:
Q* = H +LE+ G
(Q*-G) represents the energy available for the fluxes of heat and water vapour.
If the ground is completely dry, then E = 0 and H = Q*-G. If the ground is satu-
rated, or the surface is open water, the maximum possible value of E will be
E = (Q*-G)/L.
If soil temperature is then measured at intervals of 24 hours, the temperature
profiles may be approximately the same. That means that, although heat will have
moved into and from the soil block, the net effect after a 24-hour cycle of temper-
ature change can come close to zero (G➔O). If G is 'made zero' by choice of a
24-hour analysis period, or neglected in situations where it is indeed relatively
small (e.g. low soil heat conductivity), then the upper limit of evaporation is
(Q*/L).
However, the negligibility of G cannot always be assumed. It is easy to find
situations where the temperature profile in the soil block will differ markedly
between the beginning and end of a 24-hour period - or of a week. For example,
such changes may arise from differences of snow cover within the period, or from
·the percolation of water from snow melt, or from marked changes from very dry
to very wet conditions due to heavy rain, or the reverse change from wet to dry -
or just from advancing spring or fall.
Over a year, the soil temperature profile (measured at the same time of day)
will be approximately the same, and again G➔O. Budyko, quoted in Sellers (1965),
used this to achieve maximal simplification. He first found that if Q* is the annual
net radiation for a wet surface with an assumed albedo of 0.18, then:
Q* = 10 LT (LT= sum of daily mean temperatures greater than 10°C).
Next he argued that, if it is assumed that the convective heat loss H is zero,
then:
£0 = Q* /L :::::0.2 LT mm/yr.
Observations show that, when water is not limited, a high percentage of the
available energy is devoted to evaporation and a relatively small proportion to
heating the air. When water is less available, then H takes over the main load of
dissipating incoming radiation. G is a major heat balance component over water
bodies, because the downward heat flow is, in this case, increased by convective
action within the water surface layer.
Lowry (1967) gives.an interesting review of the division of labour in dissipa-
tion of heat by various surfaces. Table 4.1 is extracted from his listing.

38 .
CHAPTER 4 - WATER AND THE HYDROLOGICAL CYCLE IN AGRICULTURE

Table 4.1
Kind of site Site energy budget
Dominant factors in
dissipating energy from Lush growing crop E does 80 to 90 per cent of the dissipation.
different surfaces Meadow E largest dissipation mode during growth.
(Lowry, 1967). When the grass is cut and mats to form an
insulator and a vapour barrier (litter effect),
then H dominates.
Rocks, city pavement, desert E is almost nil; H and G are comparable,
depending on the soil heat conductivity.
Open water E and G completely dominate H.
Still water G is reduced as compared with open water,
since convective mixing is reduced. So the
water surface becomes warmer, increasing
both E and H.
Snow (mid-latitude), T < 0°C Q* is rather small because of large albedo.
G is very small (excellent insulation), Eis very
small (low vapour pressures), so H is a large
fraction of Q*.

It was stressed in this section that energy fluxes at the .surface were being
examined. A brief consideration of the energy and water budget of an air layer
near to the ground serves to illustrate some of the complications introduced as
soon as measurements are taken above the surface.
Given a uniform extended area with a homogeneous plant cover, or a
uniform bare soil surface, or an extended open-water surface, there will be no
change in temperature and humidity of the atmosphere in the horizontal direc-
tion; this implies no horizontal divergence of sensible or latent heat, and no
advection of energy. Various energy balance formulae used for the estimation of
soil moisture and irrigation needs, and certain applications of the 'combination
formulae' (mentioned later), postulate no advective effects. This is often stated in
the form that "the area is homogeneous for large distances around the point of
interest."
When the advective terms can be neglected, the vertical fluxes of sensible
and latent heat change slowly with height, and these vertical fluxes can be esti-
mated from measurements made a short distance from the ground. Observations
are required for deriving LE and H at the surface; in practice this is very difficult
to achieve.
H and LE are customarily derived from measurements of temperature and
water vapour at two levels (at least) near the ground, the depth of the layer of air
involved being kept as small as possible in order to reduce buoyancy and advec-
tive effects to a minimum. However, making the observations only at low levels
raises two problems. For one, even when vertical temperature and humidity gradi-
ents are large, their determination over an interval of at most a few metres does
require high-quality instrumentation and professional evaluation, not an easily
achievable objective for station use. And second, the validity of various generally-
used formulations of boundary-layer behaviour is open to doubt at very low levels
above the vegetation. Garratt (1978) shows that minimum profile elevation
should be at least three times the height of the roughness obstacles. Over low
vegetation this requirement is no problem, but it handicaps its application over
tall crops, let alone above forests.
A second relationship between H and LE is needed. In the conditions
postulated:

13 is the Bowen ratio, and y the psychometric constant (here y = Cp/L , since we
use specific humidity q instead of the vapour pressure e which is used in
psychrometry).

39
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

Divergence becomes important at the boundary between different surfaces

(the so-called 'edge effect'): water to land, arid desert to irrigated field, to quote
two extreme examples. Sufficiently far from the boundary into the crop, condi-
tions can become homogeneous; in general terms a 'sufficient' distance is of the
order of hundreds of metres. In an analogous problem of the siting of a pan evap-
orimeter in a desert area, Thornthwaite and Mather (1955) suggested a buffer zone
of at least 400 m in diameter. Other work on the build-up of a boundary layer
matched to the new surface (see section 5.2) suggests a figure of the order of one
vertical unit for 30 to 100 units of horizontal fetch.
However, even with measurements at the surface and/or with horizontal
homogeneity, Slatyer and Mcllroy (1961) point out the possibility of unsuspected
advective influences:
"During the daytime the net radiant flux (or more strictly Q*-G) is
often thought to represent the total 'available' energy, which must be
divided up between sensible and latent heat fluxes into the atmosphere.
It has therefore been suggested that average Q* must set an upper limit
to average LE, even for a thoroughly wet surface. However, where
strong advection is present, the signs of H and LE may well differ. For
instance, in the irrigated 'oasis' case heat may actually be supplied to
the ground from the air for much of the time, while strong evaporation
occurs at the moist surface, consuming as latent heat appreciably more
than the net radiant energy available at the time.
In any reasonably homogeneous moist area, large enough for local
or low-level advection to be definitely ruled out, average latent heat
consumption will still be found sometimes to exceed average net
radiant energy received, often over months at a time. This arises from
what is in effect large-scale or upper-level advection due to the move-
ment of weather systems, in the above case from the passage across the
evaporating surface of relatively warm, dry air masses originating over
large, relatively hot, dry surfaces elsewhere. It seems probable, in fact,
that certain regions, even though quite well provided with incoming
radiant energy, will act as long-term sinks of sensible heat, due to the
general pattern of circulation over them - provided, of course, that
sufficient moisture is available for the amount of evaporation involved.
This effect has been observed strongly at Aspendale (Australia), and to
a lesser extent at Davis and other places in the United States, and in
Israel and England."
For further information on advection and internal boundary layers (section
5.3.3.), see Brutsaert (1982), Garratt (1990), and Kroon and De Bruin (1993).

4.3.4 Aerodynamic Measurement of the removal of evaporated water by atmospheric motion has one
estimation of evaporation advantage over the energy balance approach in that it does not require data at
ground level on the state of the surface. The basic principle of the aerodynamic
approach is that vapour transfer will occur along a vertical moisture gradient and
with an efficiency dependent mainly on the intensity of vertical turbulent
motion. The most fundamental method for using this process is measuring the
actual turbulent air movements which give rise to eddy diffusion, while at the
same time measuring the associated fluctuations in humidity which determine
the air motion effectiveness for water vapour transfer.
Experimental application of this fundamental method is called eddy flux or
eddy correlation observation. It requires measurement of vertical components of the
wind and of humidity fluctuations at high frequencies of 1 Hz or faster. This can
be achieved with hot-wire, sonic or fast-propeller anemometry, but (as with
lysimetry, another 'fundamental' method) the work and the costs are such that
operational application is not really feasible. For further information on eddy flux
observation, see Fritschen and Gay (1979), Friehe (1986) and Kaimal and Finnigan
(1994).
However, the method can be simplified so that it is suitable for operational
purposes by concentrating on the working principle that the general level of

40
CHAPTER 4 - WATER AND THE HYDROLOGICAL CYCLE IN AGRICULTURE

turbulence is a good measure of the efficiency of atmospheric transport. This

means that measuring vertical wind components can be omitted, and that the
evaluation of a full 'high-frequency' time series is being replaced by simpler
determination of standard deviations of wind or temperature. For boundary-layer
wind analysis this has proved practical (Delaunay 1984, Christoffer and Jurksch
1985, Wieringa 1986, Beljaars 1987b), and for fluxes of sensible heat as well
(De Bruin, et al., 1993). Field-applicable methods to determine evaporation from
mesoscale-averaged turbulent flux are currently in development (De Bruin, et al.,
1996). Further information on this can be obtained in section 5.3.
The aerodynamic profile approach is based on the fact, that the turbulent
exchange is linked to the vertical wind speed gradient, i.e. to dU /dz. This was first
developed by Thornthwaite and Holzman (1939) - and although their equation
is the least reliable of all methods relating evaporation to the wind profile (it is
only 'valid' in neutral stratification), it is still exclusively quoted in most text-
books, both because it was the first and because it is easiest to derive. The formula
has been superseded by more reliable approaches.
Another, theoretically equivalent but more versatile, wind-humidity profile
formula for evaporation by Deacon and Swinbank (see WMO-TN No. 83) is:
2 q2 -ql
E = -p C D(z) U (z)
U2-U1
where p is air density and Co(z) is the drag coefficient, a surface roughness measure
related to the roughness length 2 0 (section 5.3.3) by Co(z) = (0.4/ln (z/z0 )) 2, where
Co(z) and U(z) should refer to the same height. The dependence of Co(z) on
thermal stability is known, so Deacon-Swinbank is much more applicable than
Thornthwaite-Holzman.
However, aerodynamic profile methods like this are now less often applied
in the field because people gradually found out the size of the uncertainty in
short-mast profiles of wind, temperature and humidity in non-homogeneous
terrain. Profiles cannot represent turbulent fluxes sufficiently accurately for evap-
oration estimation, unless upwind terrain is homogeneous over distances of about
1 km (Wieringa, 1993). If you need reliable flux information, it is better to
measure gustiness or variance than to use profiles.
The bulk aerodynamic method can formally be seen as a simplification of
aerodynamic profile methods. For example, when we take the lowest profile level at the
surface (z1 = 0) in the Deacon-Swinbank formula, we retain a formula of the type:
E = constant x (U-function) x (q air - q surface)

The first consistent description of the evaporation process given by Dalton

(1802) had the same form: evaporation is proportional to the vertical water
vapour gradient and to the wind speed - involving, of course, a proportionality
constant.
Formulae of this type abound, with wind and vapour pressure expressed in
all possible kinds of units. All are founded on the general idea that the transfer of
water vapour from a surface is the result of the action of a mixing process (specif-
ically, turbulence generated by the wind) operating between two levels where
vapour pressures (or vapour densities) are different. All of them are based on
careful experiments in some particular setting of terrain and climate, using the
observation heights which happened to be available or practical locally.
Bulk formulae are never generally applicable, because the proportionality
constant(s) has (have) to account for the relation between average wind speed at
a locally chosen height and the overall degree of turbulent transport. Since such
a relation is terrain- and stability-dependent, a bulk-aerodynamic formula devel-
oped for e.g. estimating lake evaporation from observations at 2 m height is only
directly applicable to lake evaporation estimates in a similar climate from similar
data - although it may do that excellently. However, naming any particular bulk-
formula here, even by way of example, would be a very confounding general
endorsement. Examples given elsewhere in this compendium must only be used
in the contexts for which they are specified.

41
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

Bulk aerodynamic formulae are, strictly speaking, not fully aerodynamic.

While the eddy-correlation and the wind-profile approaches to evaporation esti-
mation only use atmospheric measurements, a bulk formula requires information
on vapour pressure at surface level, which is difficult to get. Bulk formulae have
therefore been most successful in estimating open-water evaporation, assuming
full saturation at the water surface, because water surface temperatures are easy to
measure.

4.4 'COMBINATION' Both the energy balance and the simplified bulk-aerodynamic . (Dalton)
METHODS OF PENMAN approaches require knowledge of surface properties such as temperature and
AND OTHERS vapour pressure or specific humidity, which are difficult to measure at the surface
with any accuracy, and which certainly do not form part of normal climatologi-
cal observing schedules. Moreover, the energy balance method and the real
(non-bulk) aerodynamic method require profile measurements, which again do
not feature in regular station observation schemes.
Penman (1948; 1956) developed a method to eliminate the 'unknown and
unknowable' properties at surface level by taking the inventive step of combin-
ing the energy balance with the aerodynamic approach. His original method only
required measurements which are actually made at routine weather stations:
temperature, humidity, wind speed and sunshine duration; it resulted in what he
called 'open water evaporation', which he then related to the evaporation of wet
grass.
This sounded too good to be true, and in fact the translation of his result to
the evaporation of actual open water, or of vegetation, needed additional research.
Modified 'Penman' formulae, some by Penman himself, some by others, have
been developed in such a number that it is nowadays impossible to know what is
meant by 'Penman-evaporation', unless the formulae being used are specified.
Next to this, various simplifications were developed, in particular the Priestley-
Taylor evaporation formula.
Additionally, Monteith (1965) and Rijtema (1965) developed an approach
which required fewer empirical output translations, using transfer resistances. This
approach allows for the systematic treatment of the more complex situations
encountered in vegetation. The so-called Penman-Monteith equation was
accepted by FAO in 1990 as the most reliable way to estimate evaporation from
meteorological data; this does not imply that it is always the best feasible way to
do so in practice.

4.4.1 Development of the Penman developed his original formula to describe water loss from the sunken
original Penman equation evaporation pan at Rothamsted, and he fitted some of his empirical constants to
· his pan data. One consequence of this starting point was that he could safely
ignore the soil heat flux G, which for such a pan is small. Therefore, if we specify
below the basic heat input as (Q*-G), this differs from the original presentation
by Penman (1948) who only considered Q*, the net radiation at the surface.
Penman used the following equations to determine E:
• The energy balance Q*-G = H + LE
• A bulk-aerodynamic heat flux formula H = CP ftt (U) (T0 - Ta)
• A bulk-aerodynamic evaporation formula LE= fE (U) (e0 - ea)
• Assumed vapour saturation at surface level eo = es (To)
where T0 and e0 are at surface level and Ta and ea are at screen height (in
England: 1.2 m). He made the assumption, generally accepted in micrometeo-
rology, that the turbulent transport functions for the scalars T and e are similar:
ftt (U) = fE (U).
An essential point in the formula derivation is now the elimination of the
variables at surface level by use of the temperature derivative of saturated vapour
pressure:
= des
eo - ea -- -= s --- uA (b oth notations
. are used)
T0 -Ta dT

42
CHAPTER 4 - WATER AND THE HYDROLOGICAL CYCLE IN AGRICULTURE

The value of the derivative fl. for given Ta is a known physical function (see
e.g. question 4.10). The result of Penman's analysis is:
LE = ~ (Q * - G) + --1.!::_ E
~+y ~+y a
Here y is the psychrometric constant, and Ea accounts for the evaporative
power of turbulence: if (Q*-G) = 0, the bulk-aerodynamic evaporation remains.
Generally, Ea is called 'wind function' and written in the form:
Ea = f(U) (es(Ta) - ea)
The two terms of Penman's equation are often referred to as the 'energy balance'
and 'aerodynamic' terms respectively.
The parameter values must be obtained from station data, in particular
radiation:
Q* = K.J.. (1-a) + L.J.. - Li
In section 2.1.2 and section 2.2.3, we discussed how to obtain this from whatever
is available. If global radiation is not measured, we can estimate it from sunshine
duration or from cloud cover. Penman used an Angstrom-type regression of radi-
ation on the sunshine ratio s/S (as discussed in section 2.1.2); · for Southern
England this had the form:
K.J.. = [K] (1-a) (0.18 + 0.55 (s/S))

Penman used an albedo a = 0.05 for open water, appropriate for his evaporation-
pan surface. Subsequent investigators have used a vegetation albedo a =0.25
instead (Doorenbos and Pruitt, 1975).
Incoming long-wave radiation can be estimated according to Brunt - as
Penman did - or (better still) to Swinbank or Idso and Jackson (section 2.2.1).
Outgoing long-wave radiation (cff4) was estimated by Penman directly from Ta; in
section 2.2.2. we discussed that correction is desirable when strong vertical
temperature gradients occur.
Temperature and humidity are measured in the screen. There is no generally
accepted way to obtain the wind function Ea; Penman had wind measurements at 2 m
height (acceptable at the open Rothamsted site) and used f(U) = 0.26 (1 + 0.54 Uz).
Unfortunately this historic fact has been copied elsewhere because "Penman did
it". In rough terrain, wind observation at higher level is needed (see section 5.3).
Brutsaert (1982) gives some wind function alternatives, including also stability
effects.
The greater the dependence on estimated values of the components and
upon empirical relationships, the longer the averaging period necessary for reli-
able results with Penman's formula. Given reliable station data, it has been found
satisfactory for daily values (provided the heat storage term G is not ignored), but
if the data quality is not assured, it should not be used for periods shorter than
about five days.
In his pioneer paper, Penman (1948) compared his derived and measured
evaporation from his pan with measured water loss from vegetated lysimeters. The
transition from water to vegetation surface required a correction coefficient of 0.6
for winter, and 0.8 for summer use, to convert his 'open water evaporation' (E0 )
to green crop evapotranspiration. He justified his coefficients (Penman and
Schofield, 1951) by noting firstly, that transpiration takes place through open
stomata; this is a response to light, and hence one which is affected by day length.
Secondly it was noted that crops reflected more solar radiation than did open
water and, thirdly, that there is a resistance to the diffusion of vapour through
stomata. A further relevant consideration in comparison between evaporation
from an open water surface and that from vegetation may be the role of the
greater surface roughness of the latter in promoting evaporation.
These comparisons gave rise to the concept of 'potential evapotranspiration'
(PET), originally defined by Penman as 'the amount of water transpired in unit
time by a short green crop completely shading the ground, of uniform height and
never short of water'. Brutsaert (1982) rightly remarks that this (widely employed)

43
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

concept is quite ambiguous, because it contains biological components such as the

type of crop and the stage in its growth cycle. There can be quite large differences
between PET and potential evaporation, if the latter is defined as the evaporation of
a large uniform surface which is sufficiently wet to saturate fully the surface air.
Moreover, if potential evapo(transpi)ration is calculated from observations under
non-potential conditions, the result can be quite a large bias in the estimated
partition of energy between H and LE.
So the practical solution therefore is: first, we define exactly the type of
surface for which evaporation is calculated from meteorological data, and call this
grass reference eyapotranspiration ET0 (Allen et al., 1994). Then, we investigate how
actual evapo(transpi)ration of various crops in various stages of phenological
development relates to ET0 , and call this ratio the crop coefficient; similar coeffi-
cients can be determined for unplanted surfaces. The practitioner has a number of
choices for the method to determine ET0 • FAO now supports the big-leaf formula
of Penman-Monteith (see section 4.4.2), but since not all measurements needed
for this method are readily available worldwide, there are alternatives. The FAO
gives as alternatives: (1) methods based on temperature only (Blaney-Criddle
seems unreliable, Hargreaves' approach using maxima and minima may be better);
(2) a useful radiation-only method based on Makkink (see section 4.4.2); (3) the
good old evaporation pan, where that is the only available information source.
The FAO review reports (Doorenbos and Pruitt 1975, Doorenbos and Kassam 1979)
contain usable tables of pan-crop coefficients.
The reference hypothetical 'grass' crop of the ET0 -definition is: height
0.12 m, roughness length (section 5.3) 0.015 m, displacement height 0.08 m,
albedo 0.23, and surface resistance (section -4.4.2) r5 = 70 s m- 1 . Like the crop of
the PET-definition, it is assumed to be 'actively growing, completely shading the
ground and never short of water' . If an actual crop is tall and moist, its actual crop
evapotranspiration (AET) can exceed ET0 ; however, it is usually AET ~ ET0 , either
by reason of the plant's control over the stomata! openings, or the failure of the
physiological mechanism of moisture transmission in the plant, or exhaustion of
the soil water available to the plant. It is in the latter situation that supplementary
supply of water by irrigation becomes desirable. A great many theoretical discus-
sions and numerous practical difficulties revolve around methods of assessing the
extent to which soil moisture falls to a level below which extraction by the plant's
root system becomes inhibited (i.e. in the assessment of the Soil Moisture Deficit
(SMD)). This is further discussed in section 4.6.

4.4.2 Evaporation formulae The evaporation of vegetation is a multistage process: from root to leaf, from
of Priestley-Taylor and stomata: to canopy and finally into the free atmosphere. Use of concepts of resist-
Penman-Monteith . ance of plants and canopies to transfer of properties, e.g. vapour flow, has
considerably increased our understanding of the process of transpiration.
In transport of electricity through a wire, Ohm's law states that the amount
of current is equal to the potential difference between the ends (voltage) divided
by the resistance of the wire. As analogue for flow of any other quantity we can
assume:
flux rate= (concentration gradient)/(resistance to flux)
The advantage of this approach is the ease of combination. Resistances in a
chain can be summed, resistance networks can illustrate multi-alternative situa-
tions, etc. Variations in the mechanism of moisture loss by leaf surfaces have been
allowed for by applying ideas of resistance to water flow within the leaf, then at
the leaf-air interface and through the crop stand.
In particular, the 'wind. function' Ea in the Penman evaporation determina-
tion can be made more accessible to analysis by rewriting it as a resistance ra, which
decreases when the wind speed increases: ra = (pCp)/(yf(U)). With this notation,
including some physical constants for convenience, the Penman equation reads:
A(Q * -G) pCP es(~)- ea
LE=----+-- X
A+y A+y ra

44
CHAPTER 4 - WATER AND THE HYDROLOGICAL CYCLE IN AGRICULTURE

where now the 'wind function' is made visible as 'air resistance' ra. It has been
shown that ra can be derived as follows from the length-dimension parameter for
surface roughness, z0 (see section 5.3), and a similar terrain parameter for heat
transfer, Z0 v = 0.2 z0 :

r3 = - I [ (z - d) (z - d)]
2
- fn -- Rn - -
K Uz Zo Zov

Here K =0.40 is the Karman constant, and d is a correction height interval; over
crops of height H, approximately d =0.7 H (see section 5.3). Air resistance ra is
small, when the wind is strong and the terrain is rough, so that the turbulence is
vigorous.
It is found that for many types of short vegetation, if they are not actually
wet but have sufficient water available, the value of the ratio between the two
terms of Penman's equation is similar. Some compensating effects between stom-
ata! resistance and turbulence-generating roughness may account for this.
Accordingly, the Penman equation was simplified by Priestley and Taylor (1972) to:
LE = a d (Q * - G)
d+y
The constant a here has in many cases values between 1.25 and 1.30, though
smaller values (e.g. for coniferous forest) or larger values (e.g. for winter situations)
can occur (De Bruin and Keijman, 1979). In winter the aerodynamic component
has greater relative importance, resulting also in sizeable wintertime evaporation
from forest, while in the same season the short-crop evaporation is small.
The Priestley-Taylor simplification has given rise to some more drastic 'short-
cuts'. An example is the Makkink relation, proven to work acceptably in the
Netherlands in summer (Makkink, 1957; Hooghart, 1987) by making good use of
global radiation measurements which are available in the national station network:
LE = 0.65 KJ.. (M(/1 + y)) mm/day
This formula is applied locally with an appropriate set of crop factors
(Hiemstra, et al., 1994). Makkink performs quite well in arid regions (De Bruin and
Lablans, 1998).
Another variant is the possibility of combining energy balance and profile
methods to obtain evaporation estimates in case that radiation is measured, but
no humidity data are available. If we can then determine H from temperature
profiles and good wind observations, determination of LE is also possible (Stricker
and Brutsaert, 1978). This approach has an interesting future in remote sensing
applications, if these can provide good estimates of Q* or KJ.. and of H.
A fundamental improvement to the Penman approach has been the addition
of a bulk stomata! resistance rs, to model transfer between leaf stomata and
atmosphere. This is often called the 'big-leaf approach to account for the active
influence of vegetation on its transpiration. Monteith (1965) and Rijtema (1965)
showed that this extension of the Penman equation results in:
pC
~(Q *- G) + __
P (es (Ta)- ea)
fa
LE==

If in this Penman-Monteith equation we substitute rs = 0, i.e. we assume that the

leaves are covered with water, the original Penman equation is recovered. If we
take rs= 00 ,meaning that stomata are closed and leaves are dry, we get the trivial
extreme, LE= 0.
The 'big-leaf' canopy resistance is an integral property of a type of vegetation
and must, therefore, be determined experimentally. Some typical values are
rs =30 s m- 1 for low crops and rs =150 s m- 1 for pine forests. FAO assumes
rs =70 s m- 1 in their definition of grass reference evapotranspiration ET0 , and it is
cogently argued that evapotranspiration estimates from weather data using this
approach are superior to lysimeter data as benchmark values (Allen, et al., 1994).

45
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

The Penman-Monteith equation is a good example of the power of resistance

formulations as it provides a way to account for various canopy effects. It shares
with the original Penman ·equation the basic advantage that it explicitly includes
two of the three stages of evaporation. However, temperature-only evaporation
estimates and the evaporation pan will remain in use for the many regions
without basic weather data.
The above discussion of evaporation methods is a good illustration of the diffi-
culties of applied meteorologists. So often the employment of methods based on
sound physical principles is only possible with accurate equipment and data, which
are not available in actual field work. This makes it necessary in practical field prob-
lems to use more than one method, and to employ as many cross-checks as are
within the capacity of the investigator. An interesting example are attempts to esti-
mate evaporation water losses of the Aswan dam lake on the Nile, where data are
rather scarce. Several of the above methods are applied, and only if all results agree
approximately, we can say that they have probably some validity (Sadek, et al., 1997).

4.5 SPECIAL FORMS OF' From a narrow point of view, dew is not considered to be a precipitation as its
PRECIPITATION transport distances are small and are measured in meters only, as compared with
the kilometers which a raindrop may fall. Yet in some cases dew can make a signif-
4.5.1 · Dew icant contribution to . the moisture balance by depositing water where it is
necessary, or in other cases undesirable.
The positive effects of dew in arid zones are various. As long as it is evapo-
rating from leaf surfaces, the stomata are open and assimilation can take place,
and some plants can absorb water from their leaves. An example of vital impor-
tance of dew in drought is given in section 6.2.3. An important negative aspect of
dew is that prolonged leaf wetness can lead to fungal diseases. A positive aspect is
the retardation or prevention of frost.
There are three forms of dew. The first is dewfall, condensation from the
atmosphere. It requires conditions which are opposite to those of high evapora-
tion: a very negative radiation balance (night with clear sky), light wind and moist
air. If in conditions of strong nocturnal cooling there is no wind, downward
vapour transport is too small for dewfall; if there is a lot of wind, a thick layer of
fog may form instead. Broad leaves have the strongest radiational cooling, and can
already show nocturnal condensation at relative humidities of <90 per cent, while
at the surface soil heat flux still delays dewfall.
The second form of dew is distillation from the soil, dewrise. After a sunny
day the soil surface is warm and dry and evaporates; when at night the vegetation
cools off, the vapour condenses on the vegetation. In some conditions, condensa-
tion can begin several hours before sunset and be maintained after sunset.
'However, in the majority of cases dew begins after sunset and stops at about
sunrise (Monteith, 1957). The already formed dew may persist in the morning, but
it usually evaporates within a few hours after condensation has ceased.
The third variation of dew is water tran.s port from another place within the
vegetation, including the fact that water can be exuded from plant tissue ('gutta-
tion') due to large root-generated pressure.
Quantities of dew are generally small. Distillation occurs most frequently,
but generally deposits much less than 1 mm. Dewfall is more variable in occur-
rence and delivers up to 1 mm Qacobs and Nieveen, 1995). Such quantities are not
always adequately recorded at weather stations. Dewfall can be very large in
regions with humid nocturnal advection, for example in coastal Nigeria the
dewfall due to nocturnal advection is 500 mm annually.

4.5.2 Snow Snow and ice play a two-fold role in the hydrological cycle; first, by providing a
source of moisture and, second, by delaying the release of moisture to the
ground, to rivers, etc., until some time after precipitation. This delay has
some consequences for water supply, budgeting and control (see Gray and Male,
1981).
Even when restricting ourselves to hydrological aspects, we can list some
other effects on agriculture of the occurrence of snow, ice and frozen ground:

46
CHAPTER 4 - WATER AND THE HYDROLOGICAL CYCLE IN AGRICULTURE

(a) Snow insulates the ground and shelters buried crops from excessively low
temperatures (section 3.3), but is hazardous to vegetation protruding above
the snow;
(b) A mild degree of freezing can have beneficial effects on soil structure
(section 3.3);
(c) Snow, ice and frozen ground are a physical hazard to livestock, and partially
or completely hide herbage from grazing animals when snow lies thicker
than 0.1 m;
(d) Regarding mobility, after a wet autumn frozen ground gives a firm surface for
field and track vehicles, and snow allows sledge transport (e.g. of fertilizers
and farm produce). In spring the saturation of soil by melt water may give
rise to a period when the use of mechanized farm equipment is either inad-
visable or impossible;
(e) Snow and ice place stress on all types of structures following snow and ice
loading.
The balance of advantage and disadvantage depends partly on whether these
climatic elements are a regular annual feature, or whether they occur intermittently.
For the hydrological input of snow, its water yield is of crucial importance.
Conventionally, it is assumed that 10 cm depth of fresh snow = 1 cm depth of
water. Air temperature during snowfall affects the density of fresh snow cover.
Ventskevich (1961) gives the following data to illustrate this point:
Air temperature (°C) -10° -10° to -5° -5° to 0° 0° to 2° 2°
Snow density (g crrr3)
Mean value 0.07 0.09 0.11 0.18 0.20
Minimum value 0.01 0.02 0.04 0.07 0.16
The density of the cover (about 0.1 g cm-3 to begin with), increases with settling,
with the mass of any new snow deposit, with duration of snow cover, and with
alternate partial thaw and freeze - the density may double within a day. At the
beginning of the spring thaw season, snow density is 0.3 to 0.5 g cm-3, or higher.
On mountains, the density of permanent snow cover increases to 0.8 g cm-3 as it
changes to firn ice. In cold climates, duration of snow coverage can be many
months (see Question 4.12, p. 52).
The snow disappears mainly by melting under the influence of solar radiation,
by sufficient rain (very light falls of rain may merely add to the snow cover), and
most effectively by advected moist warm air, since with advection of dry air the
direct evaporation ('sublimation') requires latent heat and cools the snow surface.
The duration of the period obviously depends on the depth and extent of the
cover: thawing typically lasts two weeks in Russia. Some delay is to be expected between
the end of thaw and the date when the soil is in a condition fit for cultivation.
Subsequent disposal and movement of melt water depends on whether the
snow cover developed on an unfrozen surface or whether the ground was signifi-
cantly frozen before the snow cover built up. In the first case, melt water can
percolate into the ground; in the second, percolation will be delayed until the soil
has thawed out.

4.6 SOIL MOISTURE The relationship between actual transpiration and potential transpiration contin-
BUDGETS- ues to be a matter of argument. In spite of this, however, perfectly satisfactory
IRRIGATION NEED water-budgeting procedures - devised to meet specific situations - have been
developed and have proved to be of the highest practical value in cropping
programmes, irrigation scheduling and water-resource management.
In practical applications soil water is usually expressed in terms of 'depth of
water per stated depth of soil'. The various working hypotheses relating to the rate
of water extraction are briefly as follows:
(a) Water is equally available between field capacity and wilting point; or,
(b) A linear decrease of evapotranspiration with increasing water tension
(proposed by Thornthwaite and Mather, 1955); or,
(c) A compromise in which actual evapotranspiration proceeds at the potential
rate for some time and then decreases rapidly in an exponential (or some

47
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

roughly equivalent) manner. The point .at which the changeover occurs still
·requires decision.
A simple example of water budgeting is given in Table 4.2. This illustrates
practice in the humid temperate regime of the British Isles. It is assumed that:
(i) There is only evaporation through the pl.ants (i.e. transpiration) which
proceeds throughout at the potential rate, i.e. the root range is effectively
unlimited;
Table 4.2 (ii) The .soil is at field capacity on 1 April (a reasonable assumption for the region);
Simple water budget for dry (iii) Any excess rainfall over transpiration is used to replenish the soil to field
summer in the United capacity; excess beyond this amount is deemed to have drained away and/or
Kingdom run off;
(Rainfall Rand potential (iv) Soil Moisture Deficit (SMD) is equal to the amount of water required to bring
transpiration PT expressed in the soil back to field capacity (note: this emphasizes the importance of the
mm of water). concept when planning irrigation).
March April May June July August September
R(mm) 53 20 91 29 14 25
PT (mm) 51 84 79 97 83 58
· R-PT 2 -64 12 -68 -69 -33
SMD (cum.) 0 0 64 51 119 188 221
:· surplus. 2 0 0 0 0 0

The ,e xample leads to a cumulative deficit of 221 mm, which is is unusually

high. It would require a homogeneous soil block of about a metre in depth,
initially at field capacity, and a crop stand with .a vigorous root system -c apable of
extracting water from such a soil block. A possible practical instance could be that
of a well-established ·fruit orchard. In this case our model assumes water and is
regarded .as non-limiting.
A more Tealistic computation of the same situation, ·shown in Table 4.3, takes
into account limitations imposed upon tran~piration by the plant/soil system. It
is assumed that:
(i) The soil is at field capacity on 1 April;
{ii) In response to anev:aporative potential, the first x mm is drawn at the poten-
tial rate. When this sup.ply is exhausted, the next y mm can only be released
if a potential ny is -exerted. When {x + y) mm have been drawn, a further
z mm requir,es a residual ·p otential of mz to rel.ease it (where m>n);
(iii) In this particular case it is assumed that;
Stage I storage - the first 50 mm is drawn at the :P otential ,rate;
Table 4.3 Stage II storage- the next SO mm is drawn at half the potential rate (n = 2);
Specimen water budget for dry Stage III storage - the final 25 mm is drawn at quarter potential rate (m = 4);
summer in the United · (iv) When R>PT, the surplus replenishes storage Stage I before affecting Stage II,
Kingdom water in Stage I being available for transpiration at the potential rate;
(Rainfall Rand potential (v) Any surplus from Stage I is made .avail~b1e to Stage II and this water .c an be
transpiration PT expressed in called upon at half-rate; surplus from the second stage passes to Stage III and
mm of water). can be released at one quarter potential rate.

March April May June July August September

R(mm) 53 20 91 29 14 25
PT (mm) .51 84 79 97 83 58
R-PT 2 -64 12 -68 -69 -33
.SMD (cum.) -0 0 64 51 119 188 221
Storage (0-50) 50 50 0 13 0 0 0
Storage (51-100) so 50 43 43 16 0 0
Storage (100-125) 25 25 25 . 25 25 16 8
Actual transpiration:
From rain 53 20 78 29 14 25
From Stage I 50 13 0 0
From Stage II 7 27 16 0
From Stage III 0 9 8
Total actual transpiration 53 77 78 69 39 33

48
CHAPTER 4 - WATER AND THE HYDROLOGICAL CYCLE IN AGRICULTURE

There are minor differences between the quantities in Tables 4.1 and 4.2
following rounding off. Derivation of actual transpiration for July and August will
illustrate the procedure.
The July rainfall of 29 mm is evaporated at the potential rate, leaving a
potential need equivalent to 68 mm. To meet this, Stage I storage can supply 13
mm at full rate (so reducing Stage I storage to zero), and leaving a potential need
equivalent to 68 - 13 = 55 mm; but this is only sufficient to extract 55/2 = 27 mm
at a half potential rate from Stage II storage. This leaves 43 - 27 = 16 mm in Stage
II storage. Stage III storage is untouched.
The 69 mm unsatisfied potential demand for August has nothing available
from Stage I storage and 16 mm from Stage II storage, which requires a transpira-
tion equivalent to 16 x 2 = 32 mm potential, thus leaving 69 - 32 = 3 7 mm to be
obtained from Stage III storage. The 37 mm potential can only release 1;4 x 37 = 9 mm
from Stage III, where 25 - 9 = 16 mm remains.
It will be noted that the actual transpiration totals amount to 348 mm
compared with a total of 452 mm (see second line of table for PT).

A large number of models have been devised to take into account the
following factors:
(a) The seasonal changes in the colour, degree of ground cover and stage of
growth of the crop;
(b) The associated changes in the extent of the root mass - a possible assump-
tion is that of steady downward growth of the roots at a rate sufficient to
extract water in successively deeper layers to meet and satisfy potential need.
A meteorological soil water budgeting procedure, called versatile budget, uses
the daily estimates of potential evapotranspiration and observed data on precipi-
tation as input (Baier and Robertson, 1966). This technique simulates variation in ·
daily soil moisture content by making use of physical and biological concepts of
water movement into the soil and water loss from different soil depths, and
provides estimates of Actual Evapotranspiration (AET) and soil moisture contents
in six zones of the soil profile. Validity of the soil moisture estimates has been veri-
fied directly by comparison with soil moisture measurements in the field (Baier,
1969, 1973), and indirectly through correlations between these soil moisture esti-
mates and wheat yields (Baier and Robertson, 1968).

Review questions on Chapter 4

A. Questions on basic understanding:
4.1 Distinguish between the moisture content of a soil block and the water
available to plants growing in that block. What properties of the soil
contribute to these differences and so to the soil's ability to store water
and to release it to plants?
4.2 Suppose a cubic block of compact soil material is successively broken into
four equal cubes, then each into four smaller cubes and so on. Investigate
how the total surface area of the cubes alters. To what extent are changes
of this type relevant to matters raised in 4.1? (For further discussion of this
topic, see Baver, 1948.)
4.3 What is meant by 'field capacity'? Explain why this concept is important
in studies of the water-balance and, particularly, in the management of
irrigation.
4.4 Suggest reasons why increasing rainfall increases wheat yields in Israel,
but why, in France, rainfall beyond a certain limit is detrimental (see
Figure 4.2).
B. Questions on extended application:
4.5 Water balance data are given as follows, where:
PET = potential evapotranspiration }
AET = actual evapotranspiration
R = rainfall all in mm/month
T = mean monthly temperature

49
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

Table 4.4 (1) Colombia, S. Carolina (Water-holding capacity in root zone= 300 mm).
Month I F M A M I I A s 0 -N D Year
T(OC) 8.2 9.1 12.9 17.5 22.2 26.0 27.1 26.6 24.0 18.1 12.4 8.4
PET n 15 37 65 115 158 172 157 114 64 29 13 952
-R 76 95 91 80 78 99 141 139 98 61 73 77 1108
AET 13 15 37 65 113 147 162 151 107 63 29 13 915-

Table 4.5 (2) Bismark, N. Dakota (Water-holding capacity in root zone = 200 mm).
Month I F M A M I I A s 0 N D Year
' T(OC) -13.4 -12.1 -4.3 5.6- 12.5 17.6 2-J.O 19.6 14.5 7.2 -1.9 -9.6 -
PET"'* 0 ('.) 0 31 78 115 140 121 76 31 0 0 592
R 11 11 23 39 59 85 57 46 31 24 14 14 414
AET 0 0 0 31 69 99 87 64 39 25 0 0 414

** (a) When monthly mean (i)

Examine the relationships between monthly values of PET and AET.
temperature s 0°C , evaporation is Under what circumstances will AET = PET? Explain why in other circum-
nil. stances AET<PET;
(ii) Graph the values of PET, R, AET. Identify on the graph, and by distinctive
shading, the periods and the amounts of water involved when there is:
(a) water deficiency viz., PET>AET
(b) water utilization viz.., AE-R (and water is drawn from soil
reserves)
(c) water surplus viz., R-PET in early season
(d) recharging of soil viz., (R-PET) after end of period
reserves when PET>R.
4.6 Consider a column of soil from the surface down to a depth at which the verti-
cal exchange of water is negligible, and without a transpiring plant cover.
Suppose G = change in storage; R = rainfall; E = evaporation; fi = horizontal
movement of water into ('runoff into') the. block; f0 = horizontal movement of
water from ('runoff from') the block; M = net runoff.
(i) Then confirm:
G =R - E + f i - f 0 = R - E. - (f0 - fi) = R - E - M
(ii) Suppose G➔0. Show that R = E + M. Under what circumstances will this
occur?
(iii) Under what circumstances may we assume Af = 0? (Sellers, 1965, p. 82).
(iv) Show that the runoff ratio M/R = 1 - E/R.
(v) Complete Table 4.6:·
Table 4.6 Annual water balance of continents (mm water year 1)
Continent Evaporation (E) Rain(all (R) Net runoff ('1.f) 11(/R
Europe 360 600 240
·Asia 390 610 220
North America 400 670 270
United States 560 760· 200-
South America 860 1 350 490-
Africa 510 670 160
Australia 410 470 60
Antarctica 0 30 30

Explain, in general terms, the reasons for the contrasts in the ratio M/R
· as between, for example: North America and United States; Africa or
Australia and Europe; Antarctica and the other continents.

50
CHAPTER 4 - WATER AND THE HYDROLOGICAL CYCLE IN AGRICULTURE

4. 7 Evaporation from an open water surface is sometimes estimated from:

_1 30U
E (mm day ) = --(es - ea)
T
U average wind speed (m s-1) at z :::::6 m;
T average air temperature (in Kelvin) at z :::::6 m;
es average daily vapour pressure (in mb) at the water surfaces (viz., satu-
ration vapour pressure at the temperature Ts of the water surface);
e3 = average vapour pressure (mb) of the air at z :::::6 m.
With the Tetens-formula es (in mb) can be found at temperature T (here
in °C): 1010g es= [aT/(b + T)] + 0.786, with a= 7.5 and b = 237.7 above
water.
(Above ice, a= 9.5 and b = 265.5 in Tetens). Let the mean value of various
elements during the 10 'night' hours and the 14 'daylight' hours be:
Night period Day period
T (OC) 23 30
Ts (°C) 25 25
U (m s- 1) 4.1 7.2
mean relative humidity(%) at z = 6m 80 so
Form a weighted average of these elements for the 24-hour period and
compute the evaporation. Compare these results with the values for the
two separate periods: Night evaporation rate= 3.7 mm day-1;
Day evaporation rate= 7.8 mm day-1.
4.8 Table 4. 7 gives the rate of depletion of an initial soil moisture reserve of
300 mm in response to an increasing cumulative sum of unsatisfied
potential need.
Table 4.7 Water retained in soil
PET 0 2 4 6 8
0 300 298 296 294 292
10 290 288 286 284 282
20 280 278 277 275 273
30 271 269 268 266 264
40 262 260 259 257 255
80 229 228 226 225 223
110 207 206 204 203 202
120 200 199 198 196 195

Confirm that up to an accumulated deficit of 24 mm the loss (i.e. the

withdrawal of water) is at the potential rate and that at a potential need
of 126 mm water, only 104 mm can be derived from soil reserves.
Contrast this 'model' of water withdrawal with that adopted in
section 4.6.
4.9 Starting with the general form of the energy balance (section 4.3.3) state the
assumptions (or conditions) permitting the successive approximations:
Q* = H +LE+ G
Q* = H + LE [= LE (1 + ~) when ~=Bowen ratio]
Q* = LE
The quantity Q*/L (=E) represents the maximum evaporation possible at
a given site. Is this strictly true in all circumstances? Discuss (see also
Question 8.4).
4.10 The ratio of the psychrometric constant y (:::::0.66 hPa K- 1) and~= desfdT
can be approximated for non-freezing temperatures by:
y/~ ::::1.44 - 0.145 (T - 273.2)0.63

51
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

Use the Angstrom formula of Question 2.10, this y/A relation and the
Makkink evaporation formula to estimate the monthly evaporation at
De Bilt, the Netherlands (52°N) from s/S and T data listed below
(1951-1980). Compare your results with additionally given figures of
evaporation at De Bilt, which have been calculated with the full Penman
Table 4.8 formula, and the actual Kt-averages measured at nearby Wageningen.

I F M A M I I A s 0 N D
s/S (%) 19 24 30 39 42 42 37 40 37 31 20 18
r (OC) 1.9 2.7 4.7 7.9 12.1 15.2 16.5 16.5 14.0 10.3 5.7 2.8
E (mm/month) 3.9 15.0 40.7 73.5 105.7 120.8 112.6 92.8 58.5 26.8 9.1 2.7
Kt (W m-2) 26 52 93 153 193 207 184 164 119 69 32 20

4.11 The following data (from Thornthwaite and Mather, 1955) are part of a
complete water balance sheet for a soil with moisture reserves of
Table 4.9· 300mm.

Month J; F M A M r I A s 0. N D Year
PET 3 2 19' 43 93 131 156 1318- 97 52 20 2 756
R 87 93 102 88 92 21 112 113 82:. 85 70 93 1108
R-PET 84 91 83 45 -1 -40 X X X X X X 3S2
SMD -1 -41 X X X X X X -
. Storage 300 300 300· 300 299 X X X X X 2.80 300 -
Storage 0 0 0 0 -1 -38 X X X 33 X X -
Actual T 3 2 19 43 93 129 X X 92 X X X 734

(SMD = accumulated potential water loss). Fill in the missing entries (x).
For relevant information see question 4.5. Note: As referred to in section
4.6, rainfall is presumed to evaporate at the potential rate. The balance
required to make up the deficit (R- PET) is drawn from soil reserves, but
a 'potential' demand will not, in general, be fully met, and the soil only
releases a quantity associated with the cumulative deficit.
4.12 Data on the average maximum snow depth (D in cm) and the average
duration (Nin days) of snow cover for nine stations in Russia and four
stations in the British Isles are given below:
(a) Perm Kirov Penza Novosibirsk Moscow Saratov Orel Poltana Rostov
N 181 168 152 TtiS 149 130 126 85 70
· o, 65 n 48 47 48 40 36 11 n
(b) Balmoral Aberdeen Eskdalemuir W Linton (Edinburgh)
N 59· 3'6 32 35
D 27 13 16 16
Plot these results. Estabfish the 'best' linear relationship for Russia and Britain
separately, judged by eye, and calculated by 'least squares'.
C. Questions involving practical work or local data:
4.13 If you live in a cold climate, obtain year-by-year data for your area on:
(i) Total snowfall; (ii) Duration of snow cover; (iii) Maximum depth of
snow; (iv) Mean temperature. Plot these values two at a time, and see
whether any useful relationships emerge.

52
CHAPTER 5
SMALL-SCALE CLIMATE, REPRESENTATIVITY, AND THEIR
DEPENDENCE ON TOPOGRAPHY

5.1 MICRO-, TOPO-AND It is customary to divide atmospheric phenomena in various scales. However, the
MESO-CLIMATOLOGY terms 'macro'-, 'meso'- and 'micro'-scale, are arbitrarily used depending on the
viewpoint of the user. The term 'mesoscale' has been applied to a study of the
entire high plateau of Tibet (halfway between global scale and weather forecast
scale), but also to the study of a clump of plants (halfway between the scale of a
single leaf and the scale of an entire field). For meteorological investigations, an
entire range of 'mesoscales' was listed by Orlanski (1975) and is used in a number
of fields, e.g. cumulus research.
However, knowledge of the atmospheric planetary boundary layer provides
us with at least two natural scales, which by non-arbitrary arguments can define
meteorological working spheres. The first is the height of the daytime convective
planetary boundary layer (PBL) of about 1 km, with a factor two uncertainty
depending on local climate. The PBL has in daytime some interaction with the
surface, be it earth, vegetation or water.
The second scale is the height of the so-called surface layer, around 50 m
(again with a factor two uncertainty). Below this level the atmospheric behaviour
is permanently dependent on the state of the underlying surface, and additionally
some macroscale forces, such as the Coriolis force, do not matter in the surface
layer. Therefore surface layer models (e.g. of evaporation) are two-dimensional -
height and the main horizontal flow direction - while PBL models (e.g. of
seabreezes) must be three-dimensional.
Of course, as soon as we get into the air layer below the top of the vegeta-
tion, three dimensions are again required for modelling. In a forest or at crop
stand edges, wind direction can vary by as much as 90° (Smith, et al., 1972; Popov,
1975). Therefore a separate canopy layer is defined, approximately below the so-
called displacement height, about 2/3 of the vegetation height (Brutsaert, 1975).
The behaviour of air in this canopy layer is often only intermittently coupled to
the atmospheric surface layer, and its weather can be called microweather in the
full sense of the word.
Above the vegetation, horizontal scales of layers with a given height are indi-
cated by the horizontal distance, which must be traversed before layer properties
have adapted to a sudden change of the underlying surface. The ratio between
adaptation distance and adaptation height is typically 100: I for full adaptation
(Wieringa, 1993) or 30:1 for moderate adaptation (see section 5.3 on windbreaks).
The time scale of the layer can be taken as the time needed by wind flow of
:::::5 m s- 1 to traverse the adaptation distance.
From these features we can derive a series of climate scales which are approx-
imately matched, with some practical shifts and overlap. The two entries for time
scale are the sampling period and the largest extent of processes at the named scale.
Table 5.1 The category 'mesoscale' is defined here according to the most common
Summary of atmospheric scales usage, e.g. for investigations of sea breezes, valley flow phenomena or thunder-
in time and space. storm development; it could also be called a 'regional' scale. In weather forecasting

Scale name Time scales Horizontal scale Vertical scale

Micro scale 0.1 second to 1 minute <1 mm to 100 m <1 mm to 3 m
Toposcale 3 seconds to 30 minutes 10 m to 3 km 1 m to 100 m
Mesoscale 1 minute to 3 hours 300 m to 30 km 10 m to 1 km
Synoptic scale 1 hour to 1 day 3 km to 1000 km 100 m to 10 km
Macroscale 1/2 day to 1 week 30 km to 104 km 1 km to 20 km
Global scale 3 days and longer 300 km to globe 1 km to 100 km

53
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

terms, 'mesoscale' corresponds to 'nowcasting' for a few hours ahead, which is

done by outdoors people (sailors, farmers) by simply looking at the clouds, and by
weather services through radar use combined with simple extrapolation.
Forecasting for one day on a synoptic scale is defined as that which can be
done by classic techniques of analysing weather maps; and macroscale forecasting
for a week is only done by a few international centres in the world, using the
largest computers available.
Of course, application categories .of climate are not rigidly defined by these
scales. For example, a typical problem is that of identifying the frost-liable portions
in quite a small area such as an orchard, occupying an area of 100 x 100 m 2 . Though
the effects of frost occur on a small scale, investigation procedures may be those
typical of topoclimatology. If, however, leaf or bud temperatures are incorporated,
the project would take on many features of a micro-climate project. If the orchard
lies on the slope of a valley, mesoclimate phenomena like possible katabatic winds
must be considered.

Most agricultural activity takes place within an atmospheric layer of a few metres
from the ground surface - · up to about 5 m in orchards, in the case of forests
maybe extending to 20 or 30 m. So the atmospheric phenomena, which act most
directly upon crops, animals, agricultural buildings and equipment and opera-
tions, belong mainly to the toposcale. In some cases the atmospheric mesoscale
situation at higher levels, such as stability, can be important for agriculture -
notably in connection with the transport and diffusion of spores, pollens, insects
and pollutants.
A somewhat arbitrary but useful definition of topoclimatology (the word
means 'local' climatology) is that it is the study of the modifications of regional
climates that are due to features that can be identified on maps of a .scale between
1:103 to 1:105. Features mapped on these scales give rise to climatic patterns which
can be called the topoclimate. When topoclimate is considered in relation to agri-
culture it has been termed 'agrotopoclimatology' (e.g. WMO TN No. 133). The
word topography relates formally to features of level as well as non-level terrain, but
its frequent association with undulating and mountainous areas may sometimes
render the term mesoclimatology more appropriate for studies of hilly country
climate.
Agrotopoclimatological surveys and investigations may deal with:
(a) Determination of the suitability of sites for the growing of particular crops,
e.g. vines, fruit types, .trees, or selection of cultivars which promise good
return at some site;
(b) Identification of sites in which damage or injuries to soil, plants and animals
may occur due to topoclimatological reasons;
(c) Investigating possibilities to take measures to improve the topoclimate
(Chapter 6). .
A characteristic feature of small-scale climatology is the emphasis placed on
changes of heat and mass balances, resulting from the interaction between, on the
one hand regional climatic elements, and the geometrical and actual physical
properties of the ground surface on the other hand. Topoclimatology recognizes
differences in low-level conditions as between adjacent areas of fallow soil and
pasture, and between sand and open water. The type of vegetative cover, from
fallow to densely forested areas, must be added to this list. As stressed by
Thornthwaite (1954) in a pioneer paper, it is essential to ascertain and map 'the
aspects of the soil (and its cover) that affect the heat and moisture balance', and
he continues: 'We must include maps of albedo, of aerodynamic roughness, of soil
moisture (a variable!), of heat capacity and conductivity, etc.' All such features
occur in both level and hilly terrain. Additionally, horizontal place-to-place
climate differences arise on the toposcale and mesoscale due to rivers, reservoirs,
lakes and, of course, human settlements of various sizes.
To assess the economic importance of any proposed action aim~d, for
example, at reducing frost risk or gaining benefit from wind shelter, it is essential
to undertake a statistical analysis of frequencies of occurrence of potentially relevant

54
CHAPTER 5-SMALL-SO\LE CLIMATE, REPRESENfATIVITY, AND 11-IEIR DEPENDENCE ON TOPOGRAPHY

conditions, using data over many seasons. Many schemes have failed because
inadequate attention was given to this aspect and only averages were taken into
consideration.
It is frequently stated that microclimatic and topoclimatic phenomena are
essentially fine weather effects. Although vertical temperature gradients are most
marked in clear, quiet weather, this is certainly not true for patterns of wind distri-
bution. All climatic elements in the full range of weather situations need to be
examined when topoclimatological surveys are being undertaken, e.g. when
assessing the suitability of an area for the production of a particular crop. Surveys
during a limited period of the year, e.g. the growing season, might well point to
satisfactory yields, but ground conditions in other seasons (specifically frozen
ground and saturated soil) might well prohibit pre-sowing cultivation and success-
ful harvesting.
Restricting observations to particular weather situations can only be justified
if the phenomena under study can only occur in such situations. The most
common example is that of radiation frost damage which is only likely on quiet,
clear nights. Experience suggests, that the pattern of frost spots at some particular
time of year in a given area can be constructed reliably by using data from obser-
vations of a relatively few cool clear nights (of the order of ten). The study of
windbreaks forms a contrasting example; periods with moderate or strong winds
must be investigated for a range of directions to establish the pattern of wind
reduction in the neighbourhood of such barriers.
With regard to climatological equipment and experimental procedures, the
necessity for non-standard instrumentation and observation in small-scale clima-
tology depends on the investigated questions. Only microclimatological
investigations always require instrumentation and processing facilities with a
much finer resolution in time and space than is used in the standard climatologi-
cal network. Work at toposcale requires knowledge of atmospheric turbulent
fluxes and sometimes these must be directly measured, but in other cases their
estimation from macrostation-type observations is sufficient.
Historically, investigations in the context of 'micrometeorology' in the 1950s
and the 1960s (Sutton, 1953) were only done over highly homogeneous level
terrain, because at the time, this was the state of the art in the study of physical
processes in the lowest layers of the Earth's atmosphere. The purpose was to
obtain models of exchange in the vertical of momentum, heat, water vapour and
particulate matter, and so every effort was made to ensure uniform and horizon-
tal site situations. Once having been checked in such theoretically impeccable
conditions, however, the models have proved quite useful in small-scale meteor-
ology within a wide range of environmental conditions - if wisely used. It is
wrong to say that micrometeorological basic relations (e.g. logarithmic wind
profiles) only apply above flat grassland, and that elsewhere one gets better results
from empirical relations. But in real-life conditions it is also wrong to apply basic
relations in the same way as if in a homogeneous situation.
One way to estimate the variability of results is to replicate observing sites
(some equipped with only the simplest of 'probe' instruments). If only one meas-
uring point is possible, not a grid of stations, then the degree of uncertainty of
analysis results must be arrived at from a scrutiny of the literature and/or from
previous experience. In whatever manner it is obtained, the magnitude of this
uncertainty must be explicitly stated in any numerical expression of results.

5.2 OBSERVATION Discussion of the topic of observation in the middle of a meteorology course is
REPRESENTATIVITY, quite unusual. Commonly, the educational treatment of this subject is restricted
EXPOSURE AND to an appendix with some explanations of working principles of various instru-
SAMPLING ments. Luckily, a few good books (e.g. Oke, 1987) also add comments on choosing
and handling the instruments.
The instrument, and particularly its accuracy, is only one of the links in the
chain of meteorological observation. Other important links in that chain are:
(a) The exposure of the instrument, i.e. its siting and mounting, and surround-
ing barriers;

55
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

(b) The sampling procedure, i.e. duration and frequency of observations;

(c) The observation and recording procedure (including e.g. calibration).
For standard meteorological stations, the manner of handling these links in the
chain was optimized as long ago as the 1950s. Weather forecasters and computer
climatologists have since been provided with standardized products, designed to
monitor the macroscale atmosphere. Therefore GM-II does not discuss exposure or
sampling at any length - it does not need to do so. If you follow WMO observa-
tion rules for forecasting, you are working correctly.
However, agricultural meteorologists, who work at smaller scales, cannot
assume that this observation product will also serve their own information
requirements adequately. So they have to figure out which exposure and what
form of sampling will match their own problem well enough. Only after answer-
ing these questions can they make a sensible selection of instruments. If it then
appears that different hardware and different observation procedures are required
for agro-networks, their synoptically-minded colleagues may ask why their station
measurements are not good enough? The appropriate answer to this is that they
are - for weather forecasting.
Some instruments (if not already adequately presented in GM-II) will be
discussed in Chapter 9. But this is the right moment to address representativity.

A primary requirement for a standard station is that all the observations obtained
should be 'representative', but this is all too often misunderstood. When someone
says that representativity means that 'at a given time of observation, at two or
more observing stations judged to have the same exposure, observations of some
meteorological element should be approximately the same', he is not
talking about representativity, but about accuracy of observation. In his case, the
issue of representativity is hidden behind his phrase 'judged to have the same
exposure'.
The representativity of an observation implies that its exposure, instrumen-
tation (if any) and sampling procedure are well matched to provide a reliable
average value of the observed parameter in a pre-determined space which we want
to represent.
For standard network stations, the space to be represented is usually the
atmospheric boundary layer in the region within a radius of 20 to 100 km from
the station. This determines that the appropriate sampling procedure is averaging
over a period of 10 minutes to one hour. This, in turn, means that the choice of
instrumentation is not a critical issue as even slow-responding meteorological
instruments are usually capable of adapting to parameter changes within an hour.
The optimal exposure for the various parameters will be discussed below.
For agrometeorological observation, the issue is less predetermined. Some
stations may be established to provide climatological values in, for example, a
.particular valley, in which case sampling a~d instrumentation requirements of
standard synoptic stations may apply. Other stations, however, are installed to
represent smaller spaces, such as individual orchards. If we want to detect an hour
of frost in an orchard, we cannot observe it in a Stevenson screen - however
excellently such a screen may perform at climatological stations - because in
light winds. a thermograph in a screen will require at least an hour to adapt to a
temperature change (assuming no other differences caused by other factors such
as radiation error), and also the screen height of 1.5 m is not optimal for detect-
ing what is happening at ground level. Screen temperature is then
unrepresentative for the orchard.
This example shows that representativity is not a universal quality like accu-
racy, nor a fixed attribute of some (well-executed) observation. It depends both on
the properties of the parnmeter (radiation, precipitation, wind, etc.) and on the
intended application. Just calling an observation 'representative' is a nonsensical state-
ment, unless the situation for which it is representative is specified. Representative
observation requires good and well-placed instruments, but is not the same as
measurement accuracy. Other important aspects are observation location and
sampling frequency. It is possible to optimize the observations so much towards

56
CHAPTER 5-SMALL-SCALE CLIMATE, REPRFSENTATIVTIY, AND 1HEIR DFPFNDENCE ON lDPOGRAPHY

one application, that other applications are impossible. An example would be a

precipitation sampling frequency of once a month - useful for climate atlases,
but useless for an operationally active hydrologist or meteorologist.
So representativeness requirements are user-dependent, and the spatial
scales of their needs appear to vary widely. This leads to a practical problem of how
to set up a station network so that most of the major applications are served well
enough.
A general solution to the problem is to shift our attention from the applica-
tions to the sources. There are three major source parameters which have an origin
high up in the atmosphere: incoming solar radiation (both short-wave and long-
wave) as the main source of heat, large-scale wind (say at geostrophic level) as the
principal source of motion, and precipitation as the source of water.
Clearly, for 'representative' radiation measurement, it is necessary to have as
free a horizon as possible. Surrounding obstacles should not project more than a
few degrees above the horizon, certainly not in the direction of the Sun (e.g. South
in the Northern Hemisphere). Sunshine meters or radiation meters should there-
fore be placed at the highest point, which can still easily be reached for card
exchange (Campbell-Stokes, Robitzsch) or dome cleaning (of a radiation meter).
For wind in the surface layer, the source is the large-scale airflow, well above
retarding influences of obstacles and roughness at the surface. If one wants to get
wind data which are most useful for many applications, both local and distant,
then anemometers should be placed rather high- but not on a building, because
such obstacles modify airflow much more than most people are aware of. The
useful compromise of WMO is to prescribe a 10 m mast in an open field - it is
one of the few requirements which is mandatory, and not just a recommendation.
Climbable 10 m masts can still be constructed relatively easily, in most soil situa-
tions even without guy wires. The fact that FAO only uses wind observed at a
height of 2 m (Allen, et al., 1994) is a major source of operational error.
For precipitation, the problem is that we want to know how much actually
falls to the ground, and measurement of precipitation is highly wind-dependent
(see GM II). Again a compromise is proposed by WMO, namely measurement at
as low a level as is practically feasible, provision of some form of wind shelter, but
the absence of obstacles of height H within a radius of 4 H (in order to prevent
obstacle drip into the raingauge).
Practical representativity problems for all parameters are discussed in section 5.4.

5.3 WIND BEHAVIOUR Now is the place to consider air motion as well as the thermodynamical behav-
IN COMMON iour of the lower atmosphere. In evaporation treatments often the observation
INHOMOGENEOUS height of wind is not even mentioned, let alone argued. However, wind is a major
TERRAIN factor in transport of heat, water vapour and CO 2, and some insight in its behav-
iour in actual terrain (not just over hypothetical infinite grass plains) should not
be considered a luxury.
The discussion on wind is based here on the behaviour of airflow around
surface obstacles, particularly around shelterbelts. These are important in agrom-
eteorology and a fundamental discussion of their action leads naturally to a more
general treatment of wind behaviour as influenced by surface features. Practical
application of shelterbelts for active modification of microclimate is discussed in
section 6.4.

5.3.1 Wind around barriers Wind is not necessarily harmful to crops and even necessary for certain purposes.
of varying porosity However, excessive wind speeds can be damaging and can also cause soil erosion.
The damaging effect of strong winds is therefore often reduced over a limited area
References: WMO TN No. 59;
by the use of shelterbelts and windbreaks.
Rosenberg, et al. (1983);
A shelterbelt is a single or multiple row of forest trees or a block of trees, with
Yoshino (1975); Gandemer and
or without shrubs below. The term windbreak usually refers to a hedge of lower
Guyot (1981); Oke (1987);
height, or to some artificial means of shelter, like a wall or a fence made of non-
Geiger (Section 52/53).
vegetative material. The term barrier refers to all types of obstruction to the wind,
but particularly to those of simple geometrical shape. The terms barrier, windbreak
and (shelter)belt will be used here interchangeably.

57
WMO LECTURE NOTES -AGRICULTURAL METEOROLOGY

When comparing effects of different sizes and types of barrier, a valid

generalization is to express all linear dimensions (i.e. crosswind length of a
barrier, downwind distances from a barrier, vertical distances above ground level)
in multiples of the height (H) of the barrier. In addition, it is necessary to specify
for a barrier the percentage porosity or permeability, defined as the ratio between
the open area in the barrier and the total barrier area, where both these areas are
regarded perpendicularly to the barrier axis.
Wind speeds within the zone of influence of the barrier are usually expressed
as a percentage of the 'free' (i.e. unobstructed) wind speed. However, not all publi-
cations clearly specify whether their 'free' reference speed is at a standardized
reference height, or at the same height as the considered point in the barrier influ-
ence zone.
In the fundamental discussion given below it will be assumed that the barrier
has infinite length along its axis, so that a two-dimensional flow description is
sufficient.
Figure 5.1 shows the general way in which airflow normal to the axis of a solid
barrier (with a porosity O per cent) passes over that barrier. The flow is pictured by
means of streamlines, which are drawn parallel to the local flow direction.
Figure 5.1
Airflow as influenced by a solid
barrier placed normal to the
flow; dimensions expressed as
multiples of the barrier height.
Upper picture: streamlines;
lower picture: generalized flow
zones. (After Oke, 1987).
Undisturbed flow
T ., ,. --- . . --- .. --- .. --- . --- ..
~3 H . .,... Displacemeo..t.-- -- --- -- --- :-:--.._ _ ---- .
J_ J/·
· .'
zone/ .,,,. -
1 Cavity
-... ,
\ Wake
·

//f/ 7777Ji777777777777777777777
l--3 H------~10-15 H - - - - -

Upwind of the barrier, its presence is first noticed by the flow at a distance
of 3 to 10 H, since the air nearer the barrier is blocked in its progress and therefore
experiences pressure increase. Immediately in·front of the barrier, stagnating air at
levels below H forms a kind of cushion, often with a 'bolster eddy'. Over this the
'flow rises into the so-called displacement zone. In this zone a relatively large
amount of air, rising from lower levels, must pass through the layer just above the
top height of the barrier. So right above the barrier, and for some distance down-
wind in the displacement zone, wind speeds· are relatively high, creating a 'jet' of
accelerated air. Having passed the barrier, the accelerated air is able to move down-
wards again, but because of inertia continues moving mainly ahead.
Below this jet, a low-speed and low-pressure region, a cavity zone, is formed
behind the barrier; its lower pressure draws.the jet downwards. In the cavity zone
large lee eddies are formed, reaching to a maximum height of 1.5-2 H at a distance
of 3 to 5 H downwind from the barrier edge. At ground level at this distance the
air is very turbulent, as evidenced by damage to crops and by surface wind erosion.
The downwind extent of the cavity zone is about 6 H to 12 H, depending on
atmospheric stability (Counihan, et al., 1974; Seginer, 1975a). Modelling of evap-
oration in the cavity zone formed by a solid barrier is difficult, because it must
combine increased eddying close to the belt, local decrease in mean wind, and
recirculation of unrefreshed air.
Gradually the displaced jetflow and the turbulent cavity flow merge into the
barrier wake, where the lee eddy turbulence is dissipated into smaller eddies. This
wake with significantly reduced wind speeds and increased turbulence extends to
20 to 30 H downwind of the barrier Qacobs, 1985), sometimes detectable to 50 H.

58
CHAPTER 5-SMALL-SCALE CLIMATE, REPRFSENfATIVITY, AND 1HEIR DEPENDENCE ON TOPOGRAPHY

Figure 5.2 Height

Actual wind speeds at various 4H
heights before and behind a
2H
barrier of height H, if the
oncoming flow is undisturbed H
and has a logarithmic wind
profile (Nageli, 1943).

) porous
barrier
dense
Wind
speed --barrier
u
downwind distance

-lOH barrier lOH 20H 30H

location

Field observations suggest that the flow above about 3 H is not seriously disturbed,
except maybe immediately above, or within 1 or 2 H downwind of, the barrier (see
Figure 5.2).
If now we exchange the solid barrier for a permeable one (say, 50 per cent
porosity), there are two major changes in the airflow across it. First, the stagnation
upwind is less because some airflow can now pass through the barrier. This results
in smaller wind speed reductions in the cavity zone, often much smaller lee eddies
in that zone. Second, the wake now reaches further downwind, and velocity
reductions are easily noticeable to a distance of 30 H, depending on porosity. This
is due to the fact that the pressure in the cavity is not very low any more, so that
the displaced jetflow is no longer drawn downwards, resulting in slower surface
speed restoration.
The resulting degree and extent of reductions of wind speed behind a two-
dimensional shelterbelt are given in Figure 5.2 for a dense and for a porous barrier.
The above-mentioned features are clearly noticeable: speed decrease upwind of the
belt, acceleration above it, maximum retardation directly behind the belt depend-
ent on porosity, wake extending to 20-30 H downwind, little effect at heights
above 3 H.
Figure 5.2 shows that there is a trade-off in the choice of porosity for optimal
wind reduction. The reduction given by a porous barrier is less, but it lasts longer
(see also Figure 5.3). For significant wind protection of large areas, medium-dense
barriers with a permeability of about 35 to 55 per cent seem to give the most satis-
factory results.
While good estimates of permeability are possible for an artificial barrier of,
say, bamboo fences or paling, or stone walls, no accurate method exists for
calculating the permeability of natural shelterbelts of trees with or without
undergrowth and thick bushes, etc. (Bean, et al., 1975). Moreover, the belt
thickness is greater for an oblique wind than for wind approaching the belt
perpendicularly.
In judging the porosity or permeability of a belt for airflow (dynamic poros-
ity), both the overall screen density (judged optically) and the size of the
individual holes (mesh size) need to be considered. For a given optical density, the
smaller the mesh size the greater the shelter effect. When the meshes are less than
a few centimetres across, the properties of the material are important. Compared
with the smooth surface of wire or board, natural belts composed of several rows
of twigs, leaves and branches have a higher effective density than would be judged
optically. Also, if the thickness of the screen is more than a few times the hole
diameter, the flow resistance of the hole increases (Gandemer and Guyot, 1981).
So far, we were still discussing barriers with a homogeneous porosity. But a
barrier may be more open near the ground (row of broad-crowned trees without
undergrowth), or instead more open near the top (scrawny trees and much
undergrowth). Large porosity near the ground does induce 'jet' effects below the

59
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

belt (section 5.3.2), but normally generates a larger downwind extent of the wind
reduction pattern Gensen.,. 1954; Lawrence, 1955; Gandemer and Guyot, 1981;
Rosenberg, et al., 1983).

5.3.2 Wind reduction by Shelterbelts may have very different purpos-es, as · will be reviewed in
shelterbelt arrays section 6.4. They may just serve to somewhat diminish the general windiness near
the ground, they may have been made to cope with occasional violent winds from
a very specific direction, or they may be meant to ward off encroaching sand, salt
or snow.
Actual shelterbelts have a finite length; when a steady wind blows perpen-
dicularly to a belt, air flows around the end of the barrier and encroaches from the
side into the protected area, thereby eroding the sheltered zone. For a belt with a
limited crosswind length (i.e. perpendicular-to the wind), this erosion reduces the
maximum extent of the protected area which its height and structure could theo-
retically achieve - even near the middle of the belt. The crosswind length of a belt
needs to be at least 12 times its height if it is to protect a useful area despite lateral
erosion and fluctuations of wind direction.
Given adequate crosswind length, the shape of the protected area is roughly
parabolic. Figure 5.3 illustrates some of the features of that area for the case of an
impermeable and a 54 per cent porous shelterbelt. It should be noted that the
impermeable belt offers strong protection (a decrease in average wind with a factor
3 near the belt), but su'tfers from a relatively large erosion of the protected area
through lateral encroachment because the flow around the ends of this solid
barrier is quite strong. On the other hand, the porous belt provides less wind speed
diminishment, but provides some reduction over an area which is about three
times as large as the area protected by the impermeable belt.

Figure 5.3 0% porosity 54% porosity

Isolines of protection factor 010m 40m 60m 80m 100 m 120 m 010m 40 m 60 m 80 m 100 m 120m

(original wind/sheltered wind) I I

I
I I
I I
I I
I I l I I I I I I I I I I
0 5H l0H 15H 20H 25H 0 5H l0H 15H 20H 25H
at height 0.3 H = 1.5 m behind I

an impermeable wall (left) and

c:>: ~---------
I

a barrier with 54 per cent

l
.. .
.
H
I
porosity (right). (Measured;
Gandemer and Guyot, 1981). I I
I
---
I I

r
E I

n
I
I I

~
I
0 I I O•
N NI
I
I
I #
r-- -, '
.2

t3 .· ;i.2
I
I

'.. .. Su = 4 600 m2 C>, : --· · · ·-" Su = 12 996 m2 :

~ 0.8 S2 = 3 43i m2 , 52 = 3 540 m2 :
H;,,5 m : S3 = 868 m2
0 •8. : H=5 m : 53 = =0 :
I - - - - - - - - - ""' - - - - - - - - - - - - - .,. ,_ - - - - I - - - - - - - - - · - · ,.. - - -- - - - - - - - - - - - - - - - - -

The high-speed flow around. the ends of the impermeable belt in Figure 5.3
is not v~ry strong because only a single belt is present, so excess flow can disperse
sideways. But in a series of belts, with gaps in between for the passage of machin-
ery and livestock, the wind .will tend to blow through the gaps in strong 'jets'.
Trees or bushes at either side of the gap are very exposed to wind damage - forest
'windblow' can easily start from such gaps,. which operate as starting points for
subsequent progressive blow. The hazards from 'jets' cannot be eliminated, but
can be diminished by screening the sides of the gap protectively with hardy
bushes or wire mesh; or, if conditions permit, staggering the shelter belts to place
the openings in a zig-zag course.
A similar 'jet' phenomenon is noticeable for belts of the 'alley' type, i.e. rows
of tall trees without branches or shrubs as an understorey. Usually such belts

60
CHAPTER 5-SMALL-SCALE CLIMATE, REPRFSFNfATIVITY, ANDTIIEIR DEPENDENCE ON lOPOGRAPHY

consist of a double or single row of trees along roads and highways. Because of the
gaps between the tree trunks, airflow passes through the lower part of the belt,
with the result being that the surface wind speed may even be accelerated below
the tree canopy in this type of belt. However, although the maximum wind reduc-
tion at distances of about 5-6 H is, of course, somewhat less than behind more
homogeneous types of belts, the limit of the zone protected by alley belts extends
much further downwind.
If the wind approaches a shelterbelt at an angle, the overall extent of the
downwind protected area is approximately proportional to the cosine of the angle
to the normal (Seginer, 1975b). So the protected area is not much decreased, until
the wind angle to the normal exceeds 30°. The general parabolic outline is
preserved even with larger angles, but little shelter remains if the angle exceeds 45°.
The influence of thickness (or distance through) of single shelterbelts has
been studied mainly in relation to wide forest-type belts, with cross-axis dimen-
sions of 5 H or more. The general finding is that the sheltered zone beyond the
downwind forest edge does not extend as far as with an equally dense, but narrow,
belt. At the windward side of such a belt the main air stream is deflected upwards,
but it progressively becomes horizontal flow over the very rough surface presented
by the treetop canopy, and finally leaves the lee edge moving horizontally. The
displaced air stream then reaches the ground at a shorter relative downwind
distance than would occur for a single barrier. Effectively, a very thick belt acquires
the properties of an impermeable barrier, with a correspondingly short protection
distance downwind.

With respect to a series of parallel belts, the shelter effect between belts tends
to be greater than to the lee of a single barrier. Flow studies (e.g. Beljaars, 1982)
indicate that effects on the wind climate over a larger area are even significant
with large interspaces, and that the presence of belts with intervals of about 100
H can still exert around twice as much shear stress on large-scale airflow as a tree-
less but vegetated plain. However, when the belts are less than about 20 H apart
in the wind direction, downwind wakes merge with upwind shelter zones of the
next belt. This provides all-area wind protection, and it also has consequences for
wind profile modelling which are discussed in section 5.3.3.
The choice of shelterbelt spacing has to be made on the basis of effectiveness.
There is no general optimal separation distance for rows of parallel belts, because
the best solution depends on the wind climate and other factors. When damaging
winds are predominantly from one direction or from two directly opposing direc-
tions, the belts can be much further apart than when the same total frequency of
strong winds is distributed over many directions. The optimal spacing of shelter-
belts also depends upon the desired amount of the average wind reduction over
the whole area which is to be protected. Many other factors have to be taken into
account, such as the shading of the ground from direct sunshine, soil moisture
and air temperature variations, and the types of plants and crops to be sheltered.
The distance between belts also has to be large when the economic use of farm
machinery calls for large-sized fields.
Usually the distance between belts can be greater, when the aim is to bring
about a general reduction of the mean wind speed over a wide area than when a
threshold value of wind (e.g. a soil erosion threshold value of about 8 m s- 1) is to be
excluded at all times and at all points. The most general system of shelterbelts on
large plains consists of parallel belts, placed relatively close together perpendicu-
lar to the main local wind direction; these belts are complemented by some more
widely spaced belts parallel to the main wind direction. In windy areas of
Germany, recommendations exist for a network of parallel and perpendicular belts
ten metres in height and 250-300 and 500-600 m apart, so that the enclosed area
totals about 0.15 kffi2. In southern parts of the Russian plain, especially in
Ukraine, systems of forest belts are recommended that enclose standard plots of
400 x 1 200 m2. In parts of Western Europe, for example in the Netherlands, shel-
terbelts are planted at 30-50 H separation distances in order to achieve general
reduction of wind speed.

61
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

Very few investigations are known on special aspects of shelterbelt use in

hilly country (Rutter, 1968). One obvious requirement, though, is that their loca-
tion should be chosen with consideration of their possible role in forming, or
preventing, pools of nocturnal cold katabatic air (see section 6.6.1).
In special circumstances, the direction of damaging winds is confined to a
narrow directional range. Such cases are associated with valley winds constrained
to blow in one (or in two diametrically opposed) directions; or on a larger scale,
special regional winds. In the horticultural area of the Rhone Valley in the south
of France, a very dense network of parallel belts (at about 4 to 6 H apart) is neces-
sary for protection against the frequent, very strong and very cold dry north wind
known as the Mistral.
At shelterbelt separation distances of less than 3 H, the airflow becomes
essentially different from the situations discussed above. This can be illustrated
from wind-tunnel studies by Perry, et al. (1969) of airflow over a surface with trans-
verse rectangular trenches of the same depth, but at different distances apart. Two
types of flow could be recognized, as illustrated in Figure 5.4.
Figure 5.4
Two-dimensional surface flow
patterns over two
arrangements of surface
obstacles; for further details,
see text (Perry, Schofield and
Joubert, 1969).
d-type

--- Transitional. zone

---► ---~

k-type
In broad qualitative terms, it was found that for D/H :::::1 (D = distance
between objects) separate circulations of roller eddies were generated within the
cavities, only loosely coupled with the flow above. The upper flow related mainly
to an effective base plane at some fraction of the obstacle height just below the
tops of the trenches. Perry, et al. (1969) called this flow 'd-type'. In nature, this
approaches the situation of a dense forest or crop, with the flow 'skimming' over
the vegetation tops.
_As D/H increases, the external wind begins to penetrate the gap (Geiger,
'Figure 38-4, gives an example from forestry). When D/H :::::3 or more, the external
flow circulates in the cavity, and the edges generate a turbulent wake which is fed
into the main stream. Consequently a trans_itional zone is formed, above which
the steady external flow is established. Perry, et al. (1969) called this flow 'k-type'.
It has since been shown that, in nature, such transitional zones also occur over
d-type skimming flow; they are called 'roughness sublayer' (Raupach and Thom,
1981). The zone height depends on interspace distance D, but reaches at least to
0.5H above the tops (Wieringa, 1993).
Anyhow, it is obvious that common shelterbelt-type airflow cannot establish
itself, unless there is at least space for a cavity zone, say if D :::::lOH. In smaller inter-
spaces, like those of the Provence, wind will not quite behave as pictured in
Figure 5.2.
The assemblage of 'roughnes_s elements' (i.e. tree plantations of all sizes, from
bushes to forests, tree belts, hedges) covering extended tracts of countryside tend
to reduce the velocity of wind by exerting frictional drag. Jensen (1954) reports
such effects based on a comparison of winds across two transects in Jutland, one
relatively free from natural obstructions, and the other with considerable surface
roughness.
These matters take on greater importance when large changes of land use are
under examination. While the removal of a single hedge or belt or forest block

62
CHAPIBR 5-SMALL-SCALE CLIMATE, RFPRFSENfATIVITY, AND TiffiIR DEPFNDENCE ON10POGRAPHY

cannot be shown to affect the mesoclimate of an area, the destruction of these

features on a large scale can result in an important increase of wind, and maybe a
deterioration in the ecological quality of the area - accelerating existing erosion
processes, and certainly seriously reducing the variety and number of different
habitats.

5.3.3 Wind representativity Common terrain, even if it is level and if no shelterbelts are present, is generally
at toposcale full of obstacles such as bushes, single trees, plots with sizeable vegetation, human
constructions and maybe hillocks or ditches. All these make the wind change in
the same way that shelterbelts do. A change of surface roughness will show up in
the windfield: if fallow flat terrain is followed by crop vegetation, or by a meadow,
wind turbulence and its rate of increase with height will change gradually to
match the new surface. The layer close to the new surface, where wind structure
adjustment has already taken place, is called an Internal Boundary Layer (IBL; see
Garratt, 1990); the IBL grows in height with the downwind progress of the wind
across the new surface. Above the IBL, surface conditions from the previous
stretch of surface still determine the windfield structure. Obviously, the windfield
very close to real terrain surfaces is very difficult to describe.
However, shelterbelt flow (Figure 5.2) shows reasonably regular behaviour at
heights above 3 H. At that level, or higher, we can define an average flow speed U,
which does not change dramatically over horizontal distances of 30 to 50 H. It is
found that over terrain covered with obstacles of height H, the average wind speed
at levels above 3 H increases with height z in the following manner:
Uz/U 1 = In (Zz/z0 )/ln (z 1/ z0 )
where Ux is the wind speed at height Zx· In this logarithmic wind profile formula,
the so-called roughness length z0 depends on the degree of irregularity of the
surface, and an unreliable approximation gives z0 = 0.15 H (see Tables 5.2, 5.3).
The logarithmic profile is valid for neutrally-stable conditions, i.e. a dry-adiabatic
lapse rate, and then it can be used at heights above 20 z0 , i.e. for z >3 H, up to
about 100 m (Tennekes, 1973).
Above 3 H, turbulence caused by the surface roughness does also conform to
simple laws. If cru is the standard deviation of wind speed samples of 3 seconds
duration (call them gusts) around a wind speed U averaged over 10 minutes, we
find that:
CJu = Uz/ln (z/z0 )

if cru and U are measured at the same height z. Moreover a good estimate for umx,
the largest gust with duration t at height z within those 10 minutes is:
Umx =U + E,._ cru = U (1 + E,._/ln (z/z0 ))

where E,._ = 1.42 + 0.3 In ((103/Ut) - 4) is the statistical ratio of the largest probable
extreme and the standard deviation (see App.I); e.g. if U = 10 m s-1, E,._ = 2.44 for
gusts of 3 s duration. With averaging periods of an hour instead of 10 minutes, the
probability of strong gusts increases: (umx/U) 60 min = 1.10 (umx/U)i 0 min· For
finding the most probable duration of the largest gust measured by a given
combination of anemometer and wind speed recorder, information is given by
Wieringa (1976, 1977, 1980 and 1996).
Not all terrains - in fact, very few - are homogeneously covered with obsta-
cles of a single height. If we have a field with crops of 0.3 m height, and say
z0 = 0.05 m, and if we have measured U at some height z ~ 1 m, we can now esti-
mate average wind and turbulence at other heights which are not less than 1 m.
But suppose that there is a large hedge or a tree at the upwind side of the cropped
field, how does z0 change?
The question indicates that z0 shows a scale-dependence. If we have a field
of mown grass, then z0 = 0.01 m only for the average wind profile within the
height of the IBL fully adapted to the grass - and this IBL-height is about 1 per
cent of the fetch, i.e. the distance to the upwind edge of the (grass) roughness. If
we want to obtain U at higher levels, or to estimate turbulence, we have to take

63
WMO LECTURE NOTFS - AGRICULTURAL METEOROLOGY

into account aspects of the (rougher?) terrain further upwind. Before calculating
the change in z0 , we must extend the problem.
Win:d at the surface has its source in the larger-scale airflow, driven by pres-
sure differences, maybe modified by large-scale topography, usually with a
reasonably steady value over a period of an hour and over a distance of several
kilometers. If the terrain is covered with obstacles with maximum height Hmx, the
lowest height at which we can expect occurrence of the larger-scale source wind is
3 Hmx· Assuming that for trees, or buildings, Hmx is not larger than 20 to 30 m, we
can define a blending height zb at 60 to 100 m, where the influences of separate
surface features on the larger-scale wind field are no longer recognizable individ-
ually, but where the average wind is still well correlated with average wind
observed near the surface. For heights over 100 m the correlation with surface
wind decreases rapidly for various reasons (Wieringa, 1996).
The wind at blending height can therefore be regarded as a source wind,
which is reasonably homogeneous at toposcale - and which we can still relate
quite reliably to the wind at heights from 2 to 10 m by means of the logarithmic
wind profile. The practical problem is the choice of an appropriate roughness
length for the relation between, say, 10 m and 60 m height. The appropriate
z0 -value is one, which includes the integrated effect of all roughness variations in
the upwind terrain near enough to influence the profile. A good estimate for this
upwind fetch is 30 zb, i.e. 2 to 3 km, and we must find the overall effective rough-
ness for an upwind area of that size.
To determine effective roughness from profile measurements we would need
masts of at least 50 m height, which are very uncommon. However, an effective
z0 can also be determined from turbulence information, since that retains infor-
mation on roughness of previously-crossed terrain over appreciable distance. The
formulas given above for both cru and umx can be inverted to determine a z0 -value
from measured gustiness, and this proves indeed to be the correct effective z0
Table 5.2 (Beljaars, 1987a).
Revised Davenport effective Results of observations on high masts, collected by Davenport (1960), have
roughness classification - been supplemented by Wieringa (1992) with gustiness-derived effective z0 -values
(Wieringa, 1992; 1996). obtained -over all possible types of terrain. These are summarized in Table 5.2.

Class Roughness Landscape description

Number Name length (m)

1 Sea 0.0002 Open water, tidal flat, snow, with free fetch> 3 km.
2 Smooth 0.005 Featureless land with negligible cover or ice.
3 Open 0.03 Flat terrain with grass or very low vegetation, and widely
separated low obstacles; airport runway.
4 Roughly open 0.10 Cultivated area, low crops, occasional obstacles separated
by more than 20 ob.stacle heights H.
5 Rough 0:Z5 Open landscape, crops of varying height, scattered-shelter-
belts etc., separation distance ::::15 H.
6 Very rough 0.5 Heavily used landscape with open spaces ""10 H; bushes,
low orchards, young dense forest.
7 Closed 1.0 Full obstacle coverage with open spaces ::::H, e.g. mature
forests, low-rise built-up areas.
8 Chaotic ~2 Irregular distribution of very large elements: city centre, big
forest with large clearings .
.

_With this effective z0 we can transform wind speed data between two nearby
locations. When at an observation station we have measured a wind speed Us at a
height Zs, and the station upwind roughness is estimated to be z0 s, we can obtain
an estimate of the wind speed at height zb. Subsequently, we use the obtained
U (zb) to estimate the wind Ur at some nearby reference location with height Zr,

64
CHAPTER 5-SMALlrSCAI..E CLlMATF., REPRF.SENfATIVITY, AND TI-1EIR DFPENDENCE ON TOPOGRAPHY

where zor is the local upwind roughness. Taking 60 m as a working value for zb,
the upward-downward transformation formula is:

In (60/z 05 ) In (z/zor)
Ur = Us --------
In (z/z 05 ) In (60/z 0 r)

Though this transformation is based on the logarithmic wind profile, which

assumes neutral stability, it has been shown in the field that it also applies to
situations with unstable or stable stratification (Wieringa, 1980). The reason is
that any stability-caused deviations in the upward transformation to zb are
compensated by similar but opposite deviations in the downward
transformation. However, such compensation requires that the height intervals
of the transformation are comparable. Therefore the transformation nomogram,
Figure 5.5 is restricted to heights Zs and Zr which ares 30 m.
In the nomogram below (Figure 5.5) the above transformation formula has
been used to ease wind transformation between two locations with known height
and upwind roughness. The centre point is the observation situation officially
required by WMO, namely 10 m over open terrain, for which z0 = 0.03 m (rough-
ness-class 3). This WMO-reference wind is generally called 'potential' wind UP. It
is the wind speed which would have been observed locally, if the upwind terrain
had been open and measurement had been made at 10 m height. This terrain-
normalized UP-value is a good parameter for large-scale wind climate studies,
because it has been corrected for purely local effects (e.g. the presence of the
observer's house).
Figure 5.5 30
Transformation diagram for
observed average wind speeds 20
U, if the height and the
upwind roughness are known 15
at the observation site.
Diagonal lines with (bracketed) 10
z0 -values represent Davenport
roughness classes. Broken
roughness lines at the left-
5
hand side indicate that there
z <20 z0 , so that the validity of
the transformation is doubtful 3
there. d is displacement height,
see text (Wieringa, 1986). 2
0.7 1.0 1.1

For example, if we assume that our station is located in 'roughly open' loca-
tion, and that it observes at a 4 m height a wind speed Us = 6 m s- 1; the
appropriate UP-value is 8 m s- 1 (6/0.75), and at a nearby 'very rough' location we
will observe at 10 m height 6 x 0.82/0.75 ::::6.6 m s-1 .
In the upper left-hand corner of the nomogram, the roughness isoline of
'suburb, forest' (roughness-class 7) is displaced. This relates to what was said in
section 5.3.2 on obstacle situations with interspaces comparable to the obstacle
height (Figure 5.4), namely that in skimming flow the base plane of the flow lies
near the obstacle tops. The base plane level, called 'displacement height' d, typi-
cally lies at 2/3 H (Brutsaert, 1975). Then:
U2 /U 1 = In [(z 2-d)/z0 ] / In [(z 1-d)/z0 ]
becomes the adjusted form of the logarithmic wind profile. Such profile descrip-
tion is appropriate for describing wind above forests, and sometimes also above
high crops.
Now we must return to the question on scale-dependent differences in z0 •
The values tabulated above are larger than those generally found in handbooks,
because they have to account for the wind structure of the entire 60 m high

65
WMO tECTURE NOTES - AGRICULTURAL METEOROI:.OGY

Table 5.3 Surface type Roughness length (m)

Roughness lengths of
homogeneous surfaces Concrete, flat desert, mudflat 0.0002-0.0005
(Wieringa, 1993). Flat snow field 0.0001-0.0007
Fallow unplowed ground 0.001-0.004
· Short grass and moss 0.008-0.03
Long grass and heather 0.02-0.06
Low mature agricultural crops 0.04-0.09
.' High mature. crops, e.g. grain 0.12-0.18
. Continuous bushland 0.35-0.45
Mature continuous pine forest 0.8-1.6
. Tropical forest 1.7-2.3

surface layer, including all surface inhomogeneities. For wind analysis up to some
height z over really homogeneous terrain, where over an upwind distance of at
least 100 z the same roughness is present, smaller z0 -values may be appropriate.
Appropriate values of these are tabulated above for some agriculturally important
surface types; this list has been updated with all reliable reported field investiga-
tions made between 1965 and 1990. It is therefore far more correct than the z0 -list
published by Sellers in 1965. Until now, most handbooks have copied Sellers'
z0 -list, which is identifiable. by an unrealistically high roughness value for fir
forest, namely a z0 -value of 2$3 m (derived erroneously by assigning to a forest of
5.5 m height a displacement height of only 0.1 m - d should be about 70 per cent
of the vegetation height).
When the situation is not really homogeneous, e.g. if the observation field is
small, then the z0 -values of the Davenport classification are more appropriate.
Also, these homogeneous 2 0 -values should not be used for transforming wind data
to other locations.
In section 5.3.2 we mentioned that, when belts are less than about 20 H
apart, downwind wakes merge with upwind shelter zones of the next belt. The
consequence of this wake interference is that, from wind profile data on a short
mast, we cannot estimate wind speed (or fluxes) reliably at levels much higher
than the mast top. The wind field in a wake is too inhomogeneous - it is not at
all what theoreticians like to call a 'constant-stress' boundary layer. However, we
have still been able to transform winds spatially within that inhomogenous layer
by way of an estimated U(zb), using an effective roughness which adds all upwind
effects on the profile (Table 5.2) - and its representativity proves to be reliable in
the field (Wieringa, 1980). That's scientific luck, and very useful.

5.4 TO POSCALE For prediction or analysis of atmosphere-vegetation interaction, representative

REPRESENTATIVITY OF ·weather data are needed. In practice, this means. temperature, humidity, precipi-
METEO RO LOG ICAL tation, wind, radiation, etc., right above and beside the crop, or near cowsheds -
OBSERVATIONS that is, at toposcale. Information is finally needed at microscale, in the crop or in
the cowshed, but microscale questions must be answered for each vegetation situ-
Reference: L.P. Smith (1975).
ation with the- active cooperation of users. The methods available to
meteorologists seeking to produce toposcale information include:
(a) Measuring parameters locally;
(b) Using measurements made elsewhere, near or far; and
(c) Supplementing measured or climatological parameter values by physical models.
For the sake of simple argument we are assuming that the terrain is generally
level, and is maybe rolling country with low.gentle hills. The problems associated
with significant topographical effects of mountains and slopes are discussed at the
end of this chapter.

We can be brief on the issue of measurement: it essentially comes down to a ques-

tion of the availability of money, time and various forms of manpower. Some
authorities think that the latter investment can be eliminated with the advance of
automatic weather stations, but that view is only partly true. First, unlike classic
weather observation techniques, automatic weather stations require highly skilled
technicians to be available at short notice (i.e. within a day, not within a month)

66
CHAPTER 5-SMALL-SCALE CLIMATE, RFPRESENTATIVITY, AND THEIR DFPFNDENCE ON TOPOGRAPHY

to guarantee the continuity which all weather observations require. Local circum-
stances determine which option is preferable, the high-technology solution, or
the employment of less technically skilled, but dedicated observers at simpler
stations.
Second, even with automatized measurement there are still routine mainte-
nance jobs to be done for several observations, notably the measurement of
humidity and radiation (see Chapter 9). Automatic registration of output from a
dirty pyranometer is worse than worthless - it is downright unfair towards the
unsuspecting data user.

The use of measurements made elsewhere, e.g. from an official weather station at
local or national level, is more economic than many users think. Operational
users such as consultants or farmers, in particular, underestimate the effort needed
to observe weather data with a guaranteed minimum level of quality. Since sunrise
and rain do not, in themselves, imply costs for the user, there is an impression
among the general public that weather observation comes cheaply. This manifests
itself often in a marked unwillingness to pay the costs either centrally (at the state
level) or directly. As a result, there are often too few weather stations and they
have too few resources to function adequately. This worldwide problem demands
from knowledgeable station data users that they check the quality of the station
data they get, maybe by a visit to the station.
When using weather data measured at some distance, whether it is five or SO
kilometers away, some corrections may need to be applied to ensure the data are
valid where they are to be used. For research purposes, a simple way to get these
is to make temporary 'probe' observations at locations of use, and then to derive
corrections in the form of linear mathematical regressions between 'probe' obser-
vations and simultaneous station data (e.g. Pedro and Gillespie, 1982). This will
work in those cases where real relations are linear, but-it must be repeated for each
project, location and season.
Operational users are generally not interested and/or able to set up such an
experimental comparison, and they either use the data 'as measured', or they
apply a standard rule-of-thumb correction (examples: topographical height correc-
tions for radiation, or 'power-law' corrections for wind speed). Particularly in this
operational case, the criterium of choosing a weather station as a data source is not
simply the shortest distance, but rather the best climatological comparability. For
example, when meteorological information is needed at a location some kilome-
ters from the sea, it may be better (depending on the purpose) not to use data from
a relatively nearby seaside station, but instead take data from a more distant
station which also is located a little inland.

The third alternative is estimation of the required parameter(s) by employing

models which use some data, some physics and some location characteristics of a
rather general nature, e.g. integral surface properties. Reliable models on this basis
can be made on a relatively large scale for three parameters (radiation, wind and
precipitation) because the primary source of these three is atmospheric (see
section 5.2).
The physical base for radiation estimation was presented earlier in Chapter 2.
It is based on the balance of three radiation components: net short-wave incom-
ing energy, Kt (1 - a), downward atmospheric long-wave radiation, Lt, and
outgoing terrestrial long-wave radiation, Li. Modelling of these by combining
astronomical information with local atmospheric and terrain parameters was
examined in Chapter 2. There it was shown that for the radiation balance it is
more feasible to estimate its long-wave parts than to derive short-wave radiation
estimation from cloud cover or sunshine duration. More availability of global radi-
ation observations is therefore desirable.
Wind speed station information can be modelled at toposcale, as was shown
above in section 5.3.3. If observations are made locally at a well-exposed site at
sufficient height, or if good wind information is obtainable from weather maps for
example, then this wind speed can be translated to the required application site.

67
WMO LECTURE NOTES-AGRICULTURAL METEOROLOGY

Figure 5.6 A knowledge of the upwind (azimuth-dependent!) effective roughness at the

Spatial distribution of station and at the application site is necessary for this purpose. A well-equipped
measured 24-hour precipitation synoptic-scale network then goes a long way.
during a showery day in the Estimation of precipitation at an application site is primarily done by interpo-
rolling country of Salland, the Jation between neighbouring stations through distance-dependent (i.e.
Netherlands (De Bruin, 1973). mathematical) weighing of their data. Gradually, some of the physical properties of
rainfall are being exploited. 'Optimum' interpolation, relating weights to statistical
properties of occurring precipitation (Gandin, 1970), seemed not to improve relia-
bility as much as was hoped (Kruizinga and Yperlaan, 1978). But new interpolation
models also account for the significant relative increase of precipitation at windsides
of hills (e.g. Benichou and Le Breton, 1987). Operationally, powerful support to
interpolation of precipitation can also be given by radar (Kessler and Wilk, 1968;
Wilson and Brandes, 1979; Collier, 1986), and radar coverage is gradually increasing.
So for radiation, wind, and to a certain extent precipitation, physically-based
translation or estimation methods to make optimal use of existing station
networks are becoming available. Their structure is similar: for larger-scale atmos-
pheric sources (global radiation, wind at blending height, rainbearing clouds) the
models account for surface influences (albedo, soil conductivity, surface rough-
ness, terrain slopes) integrated at toposcale. The net result for uncomplicated
geographical situations is:
(a) Radiation and wind speed information can be reliably extrapolated from a
station over topo- and meso-scale distances with proper handling;
0 3 10 17 30 60
(b) For precipitation, interpolation between stations is becoming more reliable
for rain produced by large weather systems, such as fronts. Because of the
spatial variability of showers (e.g. Figure 5.6), interpolation or extrapolation
of station measurements of precipitation will still require a dense
observation net, and even then results can only be operationally reliable at
local scales, if supported by frequent radar observations.
For other weather parameters, their major source is at or near the surface.
Essentially that fact makes it impossible to handle their toposcale behaviour in the
same way, and it requires modelling in combination with Geographical
Information Systems (GIS).
Air humidity in the absolute sense, i.e. the atmospheric density of water
vapour; is reasonably constant over moderate distances in situations without
condensation (fog, dew) or evaporation (advection over wet surface). Dewpoint is
also a rather constant humidity parameter over some distance in space or time.
Relative humidity (R.H.) is definitely not spatially constant, and R.H. data should
always be transformed into Pv or Td with use of the simultaneous station temper-
ature, before any attempt is made to use the information at other locations.
'Estimates of condensation are apt to be very local indeed, indicating the need for
more and better observation of leaf wetness.
Knowing the-incoming radiation (Kt + Lt) is only half the story for deter-
mining the surface energy balance. Short-wave radiation reflection, a Kt, depends
on local surface reflectivity, albedo a. Long-wave upward radiation, Li, depends
on surface temperature, which may differ much from screen temperature if the
soil has a poor thermal conductivity k (Beljaars and Holtslag, 1991). So upward
radiation depends on the local terrain parameters a and k. The complete radiation
balance depends also on slope and height above sea level, and is therefore much
more local than (Kt + Lt).
Air temperature observations, if radiation-screened and averaged over a few
hours, are reasonably constant over some distance because of advective mixing.
Maximum temperatures, occurring around noon when convection is largest, make
it possible for regional forecasts to be made with some confidence, particularly in
summertime. Still, attempts to transform temperatures between locations in the
way used for wind speeds (McNaughton and Jarvis, 1984) have not been very
successful, because temperatures still remain partly linked to small-scale properties
of their primary source, the radiation-exchanging ground. Spatial extrapolations
of temperatures have therefore been corrected empirically by using estimated net
radiation (Gillespie and Barr, 1983).

68
CHAPTER 5-SMALL-SCALE CLIMATE, REPRFSENfATIVITY, AND TIIEIR DEPENDENCE ON TOPOGRAPHY

On the other hand, minimum temperature is very variable from place to

place, also because the location- and vegetation-dependent soil heat flux at night
forms a relatively large part of energy balances. In hilly terrain, katabatic drainage
adds more complications. As knowing minimum temperature is of great practical
importance, and in complex terrain no station is representative, modelling efforts
aim to join small-scale terrain data with large-scale advective flow modelling (e.g.
Avissar and Mahrer, 1988).
Soil temperatures depend on soil type, soil state and surface cover. Only their
variations, e.g. the observed diurnal course, can therefore be taken from distant
station observations with some confidence; but that can be quite useful.
Station observations of soil moisture, of leaf wetness, of grass-minimum
temperatures and of pan evaporation cannot be extrapolated with any real confi-
dence, as it is too dependent on local surface, vegetation and radiation situations.
Soil moisture is very important for agriculture, but the usefulness of the last two
is debatable. However, until small-scale availability and reliability of satellite infor-
mation improves drastically, we must monitor soil moisture as remainder term in
the water balance equation (section 4.6), and require that the data for such moni-
toring procedures should be reliably available.

Where there are geographical complications - sea coasts, large lakes, mountains,
urban agglomerations - the resulting gradients in weather parameters need to be
taken into account, in preparing representative data. For radiation, changes in
turbidity and cloud cover must be considered. For the wind, local circulations may
require careful use of local wind direction observations. In any case, station loca-
tions should be chosen so as to monitor the major gradients. If a region can only
support, say, three good stations, they should not all three be located on the coast-
line, or all at the foot of the mountain range.
An additional approach for obtaining a dense spatial distribution of agro-
meteorological parameter data is the use of remotely-sensed information. Several
currently operational satellites offer the possibility to obtain spectral information
with a ground (pixel) resolution ranging from approximately 20 m (SPOT,
LANDSAT, ERS) to about 1 km (NOAA). In particular the last satellite appears to be
quite interesting for this purpose because of the frequent observations (twice a day
for each location) and for the 'convenient' cost of its data.
A major advantage of remotely-sensed satellite data is their synopticity,
which allows simultaneous acquisition of environmental information at different
spatial scales. This property can be used to extend climatic or weather measure-
ments, which are generally obtained locally at ground stations, over the
surrounding land surface. This is of great importance for many operational appli-
cations, and recent work has indicated that NOAA-AVHRR data are suited for
surface temperature estimation (Tarpley, 1990). For example, climatic estimates of
mean temperatures and the lengths of arid and cold seasons have been extended
over the surface of Toscana (Italy) with high accuracy by means of a joint data set
composed of AVHRR-NDVI and thermal data (Maselli et al., 1997). In another
investigation, AVHRR thermal data have been shown useful for interpolating daily
air temperatures by means of kriging (Petkov et al., 1996). In both cases, ground
information was complemented by exploiting the density of estimation points
provided by the AVHRR sensor (one every 1.1 km in both directions), which could
never be attained economically by traditional surface-based observation.
The often-voiced desire for mesoscale agrometeorological station networks
seems to be justified by the impossibility to interpolate a number of parameters
over customary synoptic network distances, 50 to 100 km. The agricultural
community will need local stations for observing temperature and humidity, leaf
wetness, two soil temperatures (say at 0.05 and 0.2 m) and precipitation - maybe
also global radiation, provided the cost of that will go down due to increased
demand. If synoptic stations can expand their agrometeorological information
utility by measuring wind with sufficient precision (gustiness included), and
providing global and diffuse radiation and additional soil heat information, then
the major needs for weather data in agriculture can be served.

69
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

5.5 TOPOCLIMAT- Climate is usually described at the scale of the basic network of forecasting
OLOGICAL EFFECTS stations, averaged over distances of the order of 50 km or more. To interpolate
ARISING FROM LAND- over these distances, primarily we will look at landscape differences at meso- and
SCAPE VARIATIONS topo-scale, say at least a few kilometers across. Topo-microscale effects are
discussed in Chapter 6.
References: CL-G,
A useful subdivision of landscape influences can be made as follows:
WMO-TN No. 133, Geiger,
(a) Effects of differences in physical properties of surface and soil (slope, rough-
Scorer (1978), Oke (1987).
ness and moisture status) on, for example, the energy flux partition;
(b) Local circulations resulting from the effects of (a), and their thermal effects;
(c) Influence of natural and artificial obstructions, mountains or hills on the
large-scale airflow and on local circulations.
The categories overlap, e.g. an extended dense forest acts both as a physical
barrier to the wind and as a generator of a special climate near the ground. A built-
up city with high-rise buildings will simultaneously present a physical barrier and,
in addition, by trapping sunlight in its streets and storing the heat in its dry stony
surface, alter the surface energy balance within the area of influence and generate
local circulations.

5.5.1 Effects of slopes on Until here we generally assumed the surface to be essentially horizontal. When
incoming solar radiation the ground is sloping, daytime climate effects are mainly caused by irradiation
differences stemming from differences in orientation with respect to the solar
References: Geiger (Section 40),
beam. Night-time effects depend more on flows of cold air, and here differences
Kondratyev (1969, and
in elevation are more important.
WMO-TN No. 152).
Dependence of direct-beam irradiation of a sloping surface on the Sun's posi-
tion (latitude, season, time of the day) and on slope orientation (angle, azimuth)
was mentioned briefly in section 2.1.1 and is carefully described by authors such
as Geiger (section 40), Oke (1987) and CL-Gin various graphs. The results can be
summarized along a few simple lines:
(a) The farther slopes with an equator-directed component are from the equator,
the steeper they need to be to maximize irridation, an effect which is more
pronounced in winter than in summer;
Figure 5.7 (b) On slopes with a poleward-directed component, the period between apparent
Daytime total direct-beam local sunrise and apparent local sunset ("apparent daylength") decreases with
irradiation on slopes at 45°N increasing steepness, and may be zero in part of the winter;
on 21 March or 21 September (c) Slopes which are not meridional (i.e. pure N or pure S) have their maximum
(equinox time) as function of input in morning (for easterly slopes) or afternoon (for westerly slopes).
slope angle and slope azimuth (d) Tropical slopes are irradiated for part of the year as if they were poleward,
(Garnier and Ohmura, 1968). and in the remainder of the year as if they were equatorward.
Figure 5. 7 presents a review of the day-integrated results of the various rules that
25.-----------,
'have been mentioned. Use of such information together with terrain elevation
maps can, for instance, show how regional irradiation variability in complex
terrain is least at noon (CL-G, Chapter 4).
As shown in Table 5.4 (Geiger, Table 40-1, from observations by Schedler),
the contrast between north- and south-facing slopes and between slopes and level
ground is relatively less in summer than in winter months.
N
Obviously, the greatest influence of slope on radiation will take place in cloud-
0 less weather. Comparing the total annual radiation of around 4 200 MJ m-2 yr1;
30° 60° 90° shown on generalized maps of average total short-wave radiation on a horizontal
Slope angle surface at the latitude of Vienna, with the total of about 2 300 units in the table
indicates the importance of diffuse sky radiation in these regions.
Table 5.4 February April August October Year
June December
Monthly totals of direct solar
radiation on N- and S-facing 20°N slope 22 175 294 250 48 0 1588
slopes. From observations in 10°N slope 49 214 324 294 86 13 1978
1930-1932 in Vienna, Austria
(48°N). Unit: MJ m-2 month- 1 . Level ground 75 245 343 330 123 34 2322
10°S slope 100 266 352 357 156 52 2584
20°S slope 121 283 355 370 185 67 2780

70
CHAPTER 5-SMALL-SCALE CLIMATE, REPRFSENfATIVITY, AND 11IEIR DEPENDENCE ON TOPOGRAPHY

It is important to understand that the total irradiation with the Sun behind
clouds may be comparable to its value in direct sunshine. If that were not the case,
many solar energy projects would be ineffective.
The effect of cloud (and of turbidity) is to cut out some direct sunshine to
increase diffuse radiation; and to distribute that diffuse component rather more
evenly over the sky canopy. Orientation and slope thus become less important
with increasing cloud cover. Geiger (section 40) quotes observations that, on a
mountain at S0°N with 30° slopes, the north-face receives the following average
percentage of the radiation which falls on the south-face:
December June
Direct radiation only 2 per cent 73 per cent
Direct and diffuse 32 per cent 94 per cent
When the sky is bright, radiation is scattered mainly forward, because small
aerosol particles have Rayleigh-type scattering. Then diffuse radiation is brightest
near the Sun and least from directions at 90° to the sunbeam. Cloudiness increases
the isotropy of radiation, because droplets are larger and have Mie-type scatter in
all directions. That is the reason why for greater degrees of cloudiness the ground
slope is less important for the total input. In mountain areas, radiation climate can
vary locally if topographically-formed clouds have semi-permanent locations.
The effect of slopes on vegetation varies significantly with latitude. In the
tropics where there is generally sufficient radiation and sometimes insufficient
humidity, vegetation may be more lush on poleward-facing slopes. In temperate
and arctic areas, and more cloudy regions it is more necessary to exploit the direct
sunshine which occurs on equator-directed slopes, which makes it important for
agriculture to pay attention to slopes. Vineyards in Germany can only thrive on
southwest-facing slopes. This is especially true for steep slopes where, in the
autumn, the maximum irradiation concurs with the warmest part of the day.

5.5.2 Soil temperatures on Probably the most important agricultural implication of the relationship between
slopes slope and direct solar input is that of increased soil temperature of the bare soil,
or sparse vegetative cover during germination and early growth of annual crops.
Because on sunny days the turbulent buoyancy helps the air to dispose of its heat
input, exposure and incline of slopes have more important effects on the daily
maximum of soil temperature than on the daily maximum of air temperature.
Because photosynthesis of plants requires capture of solar radiation by
leaves, it is desirable that the total leaf area (LAI) should reach its maximum when
solar radiation reaches a maximum. In maritime areas, especially in the higher
latitudes, there is a climatological lag of temperature (which governs leaf expan-
sion) behind solar radiation, and many crops are unable to fully utilize the high
radiant input of midsummer.
Consequently, early germination and initial growth is very advantageous,
and it may be achieved by exploiting the higher soil temperatures of slopes tilted
up to the Sun. The magnitude of slope-effects on soil temperatures is illustrated in
Table 5.5, which gives results of a pioneer investigation by Wollny (1878).
Temperature differences between slopes of different degree and aspect vary
with time of day. Work by Haude (1940), although under rather extreme
conditions, serves to illustrate certain important points. Observations were taken

Table 5.5 Direction of slope

Soil temperatures at 15 cm Month N NE E SE s SW w NW
depth in 15° slopes facing eight
main directions. Derived from May 10.7 10.9 11.2 11.4 11.3 11.3 11.0 10.9
daily means at 0730 hours, June 20.4 20.6 20.8 21.4 21.4 21.4 20.9 20.6
noon and 1730 hours (and July 18.7 18.9 19.1 19.4 19.3 19.2 18.9 18.7
August 19.2 19.4 19.8 20.4 20.5 20.5 19.9 19.2
hence approximating to
September 12.3 12.5 13.2 13.7 13.9 13.7 13.2 12.7
a 24-hour mean value).
Temperatures in °c

71
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

on a 23 m high sand dune on 12 clear days in winter in the Gobi Desert (42°N);
the angle of slope is not stated, but for a stable dune it would be much less than
the 'maximum angle of repose' of a sand dune (54°). While the air temperature
(observed in a shelter) stayed below freezing around noon, soil temperature at a
depth of 2 mm reached l0°C on the crest, and 22°C on the south slope. At
40 mm depth the average noon temperatures were respectively 2°C and 10°c.
Geiger (section 44) gives diagrams of full diurnal courses, which for agricultural
purposes are as important as mean daily values. In particular growth stages,
attainment of higher temperatures during short periods in the day may be
extremely important.

5.5.3 Effects of slopes and Influence of mountains on tropospheric airflow as a macrometeorological

hills on ·airflow problem has been intensively ·s tudied (e.g. WMO-TN No. 34). Our concern is with
mesoscale airflow, and two major points merit discussion. The first one is the way
References: Yoshino (1975),
in which some features of hilly areas modify larger-scale .airflow. The second one
· Scorer (1978).
is the generation of local circulations, particularly from the effects of solar radia-
tion on sfopes.
When looking at the effects of mountainous terrain on windflow, we must
consider a series of factors, these include: the absolute height of the hills; the
height differences between ridges and valleys; the shapes (e.g. the ratio of width
of a valley to the height of the containing slopes); the convexity or concavity of
hi11s and mountain ranges; the extension of high ground ('spurs') down to the
surrounding plains, and the orientation of all these features relative to the imping-
ing airflow. For example in studies of gale damage in mountain forestry, the slope
radiation effects discussed above are not relevant as in strong winds buoyancy
effects become secondary. Much can be learnt from results of windtunnel and
model studies, such as the windbreak barrier flows discussed in section 5.3.1.
An air current approaching a gently rising slope (of 'infinite' cross-wind
extent) will flow smoothly from the base of the slope up to the ridge. When the
ridge is steep, an upwind 'bolster' eddy will be found at the base of the slope,
particularly when hills rise suddenly from a plain.
In the case of a block of rectangular cross-section, or a vertical step (a shape
such as a vertical cliff face presents to the wind), it would appear from windtunnel
studies (Frost, 1973) that the oncoming flow is detached at about 1.0 H upwind,
and the bolster eddy extends to about 0.6 H up the face, at which point the main
flow again impinges on the face. The critical slope angle, at which this upwind
bolster eddy begins to form, appears to be about 25°.
Since most arable farming is undertaken on slopes of less than 10° and
pasture farming on slopes below 20°, we may then expect that in most cases
·oncoming wind can be expected to move smoothly up these slopes. The situation
is different for forestry (where gale damage is a major economic factor) and for
terrace cultivation.
Air flowing over a ridge may separate (i:e. no longer follow ground contours)
at sharp edges. The first likely separation point is the upwind edge of the ridge top
area- the zone immediately downwind of that edge may experience wind speeds
of double the average and a high degree of turbulence (Bowen and Lindley, 1977).
Beyond the ridge, the windflow will separate if the angle of downward slope is
great enough. This is influenced by thermal convection; Scorer (1978) states that
convection makes separation more likely over leeward slopes but inhibits it over
windward slopes. He also points out that sunshine on a lee slope may give rise to
a local upslope wind, aiding separation of the main flow and formation of a lee
eddy.
Separation generally protects vegetation on the lee slope. Geiger (section 43)
quotes results obtained by Kaiser which showed that for very shallow lee slopes
(2° to 4°) there was no tendency to detachments of flow at speeds up to 4 m s- 1,
but at 5 to 10 m s:....I there was a marked degree of protection to the lee.
Contrasting flow patterns arise when large-scale winds interact with valleys.
When the direction of the large-scale wind is similar to the line of the valley, the
wind speed exceeds that of the regional -wind because of canalization. Perhaps

72
CHAPTER 5-SMALL-SCALE CLIMATE, REPRESENTATIVITY, AND THEIR DEPENDENCE ON 1DPOGRAPHY

unexpectedly, the most marked increases in speed seem to occur in wider valleys
rather than in the narrow valleys of highly serrated terrain. This can be explained
by the greater importance of frictional drag in narrower valleys.
When the regional wind blows across the line of the valley, a number of
possibilities exist: roller eddies can form; or given an appropriate ratio of distance
between successive ridges to depth of valley floor, wind can penetrate down the
lee slope and across the floor.
Hlitte (1968) indicates the changes in wind pattern due to differences in the
approach angle of a current impinging on an (effectively) infinite ridge:
(a) When the wind direction is parallel to the ridge line, wind speed and turbu-
lence are greatest on the lower slopes;
(b) When the wind direction is at 20° to 50° to the ridge line, turbulent winds
with frequent changes of direction are seen in the midslope area;
(c) When the wind direction is perpendicular to the ridge line the wind speed
increases from the base to the top of the ridge.
Experience in the United Kingdom seems to indicate that in really severe
gales, the pattern of relatively exposed and relatively sheltered zones, evident in
moderate or strong winds, breaks down, and the severe winds envelop the whole
area, causing damage which is difficult to associate with any topography. In strong
winds, intense turbulence can be generated midway up the flanks of hills. Also,
with a series of roughly parallel ridges, the most severe damage to forests can occur
on the mid-upper slopes of a second ridge when the wind current, deflected
upwards by the leading ridge, descends and re-attaches itself to the windward base
of the second ridge.

5.5.4 Local mesoscale In mountain valleys, two phenomena are involved. First, flows are generated by
circulations differential heating and cooling of valley sides. Second, the valley walls exert
mechanical constraints to air movement generally, as discussed above. When
References: Geiger (sections
large-scale winds are strong, the circulations discussed below may be somewhat
41-45), Oke (1987).
masked.
On valley slopes facing the Sun, upslope winds will tend to develop because
of buoyancy effects. Obviously the orientation of the valley is critical. In a
north/south valley, during high-Sun hours both east- and west-facing slopes are
equally irradiated. Then, when large-scale wind is light, upslope winds will occur
on both slopes. The return flow of these air currents can come about in three ways.
First, there may be a balancing descending current in the middle of the valley.
Second, buoyant upflow can be balanced by downslope winds on non-sunlit
slopes. And finally, there is generally an along-valley flow, since valley bottoms
usually rise from the plains to the central high region of the mountainous area,
and three-dimensional return flow is possible. The phenomenon of valley circula-
tion has a typical vertical extent of 1.5 to 2 km.
At night, air cools off by radiation from all slopes, particularly those with
direct exposure to the sky - and this cold surface air behaves like water, it flows
downslope. This flow can be modified by barriers at all scales: narrowing of the
valley, forests or treelines, and in particular local ridges, uphill of which 'lakes' of
cold air form. Usually the speed of the downflow (katabatic wind) will not exceed
about 2 m s- 1 . However, in rough terrain 'air avalanches' may occur as cold air
dams are circumvented or overtopped, and air rushes down the slope until the
cold air 'lake' is exhausted. When the store is refilled, the avalanche will be
repeated, and such cold air rushes often occur at very regular intervals. Scaetta
(1935) observed violent katabatic flow of this type near Lake Kivu in Central
Africa, which shows that such phenomena are not confined to high latitudes,
since they are caused by temperature differences.
Very gentle slopes are sufficient for nocturnal cold air drainage: Geiger
(section 42) reports a frost problem caused by a 1:200 slope, i.e. a 0.3° slope angle.
Such terrain looks level to any observer. At toposcale, every terrain has so-called
'frost hollows', small depressions, natural or artificially created, with restricted or
no outlet for cold air flowing into them. This is discussed for smaller scales in
Chapter 6.

73
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

24 100
Figure 5.8
22 aJ 90
Diurnal air temperature and Cl
l'l
J 20 c 80
humidity variations at screen Cl.I
!...- 18 ~
height on clear spring days in the ~
3 16
E; 70 At peak
~ (1447 m)
mountain area of tpe Grosse ~ 14 ......- At peak .£ 60
"O
Arber (Geiger, section 45). Data ~ 12
..•·········· (1447 m) -~ 50
are observed averages in a ~
~ 10------
... - - - - - - - - - - >
40
warming season, so T<24 h) > T(o h)· 8
-~ 30
~
a:: 20
6
Oh 6h 12h 1 gh 24h Oh 6h 12h 18h 24t
Hour of day Hour of day

For agricultural planning purposes, the mesoscale effects of .valley flows are
usefully summarized in Figure 5.8, which gives data on diurnal courses in a valley
and higher upslope. At night, a lake of cold air develops at the bottom of the
valley, but cold air remains in the peak area, and if the peak is high enough its
temperature will be lower than in the valley. There is an intermediate warmer zone
called the thermal belt, typically at 50 to 300 m above the valley floor, depending
on local topography. In temperate and arctic climates, this thermal belt is for -
obvious reasons a preferential zone for agriculture and for other purposes, such as
dwelling.
Another well-known mesoscale circulation type is the sea breeze phenome-
non. At sea, or above _a reasonably-deep lake, the vertical heat flux His small, since
E and G are large. The water surface temperature is therefore steady in comparison
to the advected air. At the warm-surface side of a coastline (i.e. in daytime over
land, at sea in the night) the pressure at 1 km height is therefore larger than at the
cold-surface side. This sets up a circulation, generating at the surface daytime sea
breezes and night-time land breezes. Oke (1987) describes the diurnal course of
this phenomenon; horizontally, it is usually confined to distances inland from the
coastline of about 20 to SO km.
These coastal circulations can result in peculiar atmospheric transport. Van
Arsdel (1965) observed pine blister rust 15 km inland from a lake, possibly caused
by spores emitted near the coast at night, but with no infections in the interme-
diate area. Apparently the spores -emitted at the shore were picked up by the land
breeze, and were carried by the return circulation at 1 km height to the downward
flow at the inland circulation limit!
Other important mesoscale circulations are possible, wherever a mesoscale
surface difference occurs. Mesoscale circulations can be found in cities as well as
desert oases.

Review questions on Chapter 5

A. Questions on basic understanding:

5.1 It is proposed to undertake a mesoclimatic survey on a site planned for
development of an orchard. No climatic data are available for the actual
site. Outline the steps you would take to obtain some relevant long-
period estimates. (Procedures mentioned in 5.14 are relevant.)
5.2 It is stated that the protected zone to the lee of a broad shelterbelt (such
as might be produced by a wide forest strip or a forest) does not reach as
far downwind as that from a narrow belt of equal permeability. Discuss
this and suggest an explanation.
5.3 In a semi-arid agricultural region with a typical surface roughness
(Davenport-class 4), the use of windmills for water pumping had never
been considered, because their economical use would require climatolog-
ical average wind speeds of 4 m/s, while local climate statistics showed
averages of only 2.5 m/s. However, it was found out that these local statis-
tics were 24-hr averages of wind at 2 m height with a daytime average of
3 m/s and a night-time average of 2 m/s. The windmills required an
average input of at least 4 m/s at the WMO-standard height of 10 m.

74
CHAPTER 5-SMAI.LSCALE CLIMATE, REPRF.SENfATIVITY, AND TIIEIR DEPENDENCE ON 10POGRAPHY

(a) Reconsider the usability of wind energy from information given in

Figure S.S.
(b) The atmospheric surface layer, the lower 30 m, is unstable in
daytime and stable at night (see section 7.5.1). In an unstable situ-
ation, wind gradients are typically reduced to two-thirds of their
basic value, while stable wind gradients (increases per height inter-
val) are about twice as large as those in basic 'neutral' conditions.
Reconsider again the feasibility of wind energy application for
pumping.
5.4 In Scotland, some hilly areas are highly dissected with narrow V-shaped
valleys (glens) while in others, equal height differences are associated
with rolling terrain and wide U-shaped valleys (strathes). In broad terms,
what peculiarities and contrasts of weather and climate are to be
expected between the two types of terrain?
5.5 In Scotland and the hills of Northern England (a relatively high latitude,
maritime region), conditions at 700 m above MSL have been likened to
those of the Tundra in Iceland. In Switzerland, very productive pastures
and arable land exist at well above 1 000 m. Suggest climatological
reasons for these contrasts.
5.6 In the technical literature, much stress is laid upon the importance of
thermal belts in hilly and mountainous country, i.e. the zones above the
valley floor with upper limits well below the summit. Describe the
advantages of such situations (see Geiger, section 44 and 45).

B. Questions on extended application:

5.7 The data in Table 5.5 relate to wind speeds at four different heights and
at stated distances (in terms of barrier heights H) upwind and downwind
from a barrier. The speeds are expressed as percentages of the free wind
Table 5.5 at each height. The barrier height is 2.2 m.

Height of Windward (H) Leeward (H)

measurement 10 4 2 1 2 3 4 5 7 10 16 22 30
4m 100 100 101 104 104 103 101 100 99 98 98 99 100
2m 100 99 95 104 103 100 97 94 90 88 88 91 95
lm 100 94 82 22 26 32 39 45 55 66 79 88 94
0.5 m 100 83 61 31 23 26 37 42 46 60 83 92 95

Suppose the unobstructed (or 'free') wind speeds were:

Height (m) 4 2 1 0.5
Speed (m s-1) 10.6 9.4 8.3 7.0
(i) Construct a diagram (to scale) giving isopleths of the tabulated
percentages of free-wind speed at the corresponding heights.
Draw the approximate location of the cavity zone in the isopleth
graph;
(ii) Compute the actual speeds at all points, using the free-wind speeds
mentioned above; construct a diagram giving isopleths of actual
speeds;
(iii) For the unobstructed wind profile (windward fetch = -10 H) and for
the leeward profiles with fetches of 5 H, 10 H, 16 H and 30 H do the
following:
Calculate the roughness length z0 from the winds at z = 4 m and
z = 2 m. Then also calculate 2 0 from the winds at z = 2 m and
z = 1 m, and compute from this 'lower' z0 -value and the wind at
2 m an estimate of the wind at 4 m. Enter in the second isopleth
graph (with actual speeds) the 2 0 -values at the proper
intermediate height (==3 m and ==1.5 m), and the 4 m speed
estimates.

75
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

(iv) At which leeward distance (expressed in multiples of H), would you

trust wind measurements at 2 m height to give a fair estimate of
wind at 4 m height, or of wind at other places in the region?
5.8 The overall effectiveness of a wind barrier depends critically upon the
relative frequency with which wind blows across the belt line. Hogg
(1964) evaluated relative shelter (R) defined by:

wind speed in shelter

R = 1 - - - - - - - - x 100
free-wind speed

at two points, R1 at 3.5 H and R2 at 7 H downwind of a hedge 450 m

long running NW-SE, and defined E = effective wind fraction, the
percentage of total wind-direction observations within a sector from
WNW to SSE, and got:

Table 5.6
1960 I F M A M I I A s 0 N D

RI 31 38 17 38 33 37 46 34 25 7 35 30
R2 23 27 13 28 26 26 31 23 17 3 21 17
E 39 45 16 40 32 40 71 48 27 19 47 48

R was found to be largely independent of the wind-run when this was

between 160 km day-I and 400 km day-I.
Analyse and discuss these observations by plotting R1 and R2 against E
by computing a linear correlation coefficient and by confirming the
following regression equations:
R1 = 0.52 E + 9.41
Rz = 0.54 E + 6.31

5.9 Given that we define the vegetative season as that period of the year in
which average air temperature >6°C, list the factors and conditions
Table 5.7 which, in practice, reduce the length of the season.
Place A B C D E F G H I
Number of days (L)

Absolute maximum 316 285 249 285 Z96 282 (?) 224 256 273
Useful maximum (a) 288 259 227 273 281 222 218 243 256
('Maximum utile')
Lower quartile (QI) 279 245.5 214 257.S 258 210 202 232 238
Median (Me) 250 230 199.5 241 247.5 186 193 220 222.5
Upper quartile (OJ) 233 219 183 221 225 174.5 176.5 213.5 211.5
Useful minimum (a) 223 214 173 214 222 165 176 195 194
('Minimum utile')
Absolute minimum 217 201 165 193 208 140 155 179 190
Arithmetic mean (AM) 253.9 233.8 200.5 240.6 239.6 191.5 191.8 213.8 225.2

A. Geneva (375 m) 5.10 Primault (1972) defines climatically the vegetative period as the number
B. Neuchatel (440 m) of days in the year between:
C. La Chaux-de-Fonds (992 m) • Commencing date: the seventh consecutive day when mean daily
D. Lausanne - I (485 m) T ~ S°C;
E. Lausanne - II (485 m) • Terminating date: the fifth consecutive day when mean daily T<5°C,
F. Leysin (1 250 m) or: the third consecutive day with minimum T ~ 0°C,
G. Chateau-d'Oex (1 000 m) or: the day when minimum T ~ -2°C.
H. Marsens (71-6 m) (Seasons with a terminating date before 1 August - in the N.H. -
I. Fribourg (640 m) are ignored.)
(a) Is obtained by omitting the
For the 1931-1960 period he gives the data in Table 5.7:
highest, or lowest, 10 per cent
(i) Plot median values of L against heights above mean sea level;
of cases.

76
CHAPTERS -SMALL-SCALE CLIMATF., REPRESENfATNITY, AND THEIR DEPENDENCE ONTOP<XiRAPHY

(ii) Draw the best fitting straight line (by eye) through the mean of the
points;
(iii) Compute the best fitting straight line by the method of least
squares. (L =274 - 0.072 (height in m));
(iv) Examine the change with height of the other measures of length of
vegetative period (see also 5.9 and 5.12).
5.11 The annual mean temperatures (°C) of the stations listed in 5.10 are:
A B C D E F G H I
Arithmetic mean 10.6 9.4 6.8 9.6 10.3 6.2 6.1 7.7 8.2
Median 10.4 9.4 6.9 9 .8 10.4 6.2 6.2 7.8 8.3
Plot mean temperature against height above MSL. Investigate the rela-
tionship between mean annual temperature and the length (L) of the
vegetative season.
5.12 Conrad and Pollak (1950) define the vegetative period as the number of
days in the year when the temperature was 2::43°F (6.1 °C) derived from
the average annual course of temperature. For 168 stations in
Switzerland he derives:
L = 268 - 0.07H (H in m). State clearly the differences between defini-
tions, assumptions and methods used in this derivation compared with
those of 5.10.
5.13 If the net outward long-wave radiation towards the zenith is I0 (0), then
radiation in a direction making an angle z with the vertical, Io(z), is
given by:
I0 (z) = Io(0) cos y z
where y = 0.110 + 0.0255 e (e = vapour pressure in hPa):
(a) When y = 0.3 show that e = 7.45 hPa
When e = 23.1 mb show that y =0.7.
(b) Confirm the following values of Io(z)/I0 :

r 0.3 0.7

z
30° 0.958 0.904
60° 0.812 0.617

C. Questions involving practical work or local data:

5.14 (a) For some substantial area (say 5 000-10 000 km 2) containing a number
(say at least 10) of climatological stations, plot the daily values of a selec-
tion of:
(i) Daily maximum temperatures;
(ii) Daily minimum temperatures;
(iii) Daily mean temperatures;
(iv) Hours of bright sunshine;
(v) Wind speed and direction at some chosen reporting hours; and,
(vi) Any other element, e.g. snow cover, minimum daily humidity, for
about 10 spaced days within a month and the monthly mean
values of the chosen elements.
(b) Compare the spatial distribution for individual days with that for the
month and note the extent to which the influences of topography (e.g.
land and coastal contrasts) are revealed in the distribution pattern -
e.g., do particular geometric patterns appear consistently on all days; or
on days experiencing certain synoptic situations?
(c) For specific elements, e.g. hours of sunshine, does it appear possible to
'thin out' the network? Is it only possible to draw isopleths on a coarse
scale, e.g. daily maxima at 2°C intervals, when meaningful maps may
require the distribution of isopleths at 1°C intervals?

77
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

Note: Action on network density would need to be based upon an extended series
of observations covering all types of synoptic situations, all seasons, and all types
of weather.
5.15 Get a good terrain map of the surroundings of a nearby weather station
(better than 1:100 000, if available). Mark 8 or 12 direction sectors
around the station and classify the sector roughness within 3 km accord-
ing to Table 5.2.
• Then go and visit the station, stand next to the wind mast, observe
the actual terrain cover ·and obstacles (maybe take photographs for
future reference), and classify again the roughness for each sector.
(If the surface is sloping down away from you in a sector, diminish
its roughness class number by 1; do the opposite for rising terrain).
Note for reference faraway larger features (sea, hill).
• Try to sketch a detailed map of the surroundings (scale :::: 1: 1 000)
featuring surface cover and obstacles (woods, houses, hills - with
height and distance).
• Using the corrected sector roughness values, calculate for each
sector the station exposure correction factor, i.e. the ratio of actu-
ally measured wind averages at the actual mast top and the wind
speed which would have been measured at 10 m height when the
sector (hypothetically) would have had Davenport roughness class
3. If relevant, also use Figures S.2 and 5.3 for this calculation. (Show
correction factors to the station keeper - he may be surprised.)
• Make an estimate of the wind-related errors in bulk-aerodynamic
evaporation estimates for all sectors, without and with use of wind
exposure correction.
5.16 Select a transect across hilly terrain with a number of rainfall measuring
stations. Plot the mean annual and mean seasonal rainfall along the
transect and relate the pattern to altitudinal variation and to directional
exposure. Examine the changes associated with prolonged frontal rain,
with showering situations.
Note: A deep-going investigation would involve examining patterns in relation to
different wind directions.

78
CHAPTER 6
AGROMETEOROLOGICAL MANAGEMENT AT MICROSCALE
AND TOPOSCALE

6.1 INTRODUCTION Meteorological knowledge about the energy exchanges and transports at the
surface has useful applications in agriculture. The practices considered in this
chapter are directed towards managing the climate near the ground, partly by
influencing the moisture status and temperature of the upper soil layers, e.g. by
mulching, and partly by action above ground, e.g. by shading or use of wind-
breaks. Division into separate topics is to some extent arbitrary because of close
interrelations.
To circumscribe our purpose, some basic meteorological requirements of
vegetation are described in advance. Biological significance of soil temperatures
must first be described, before we can discuss usefully how to influence those
temperatures.
Land plants live in two media: in the air near the ground, and in the upper layer
of the soil. The soil is the only medium for most plants in their first stage of life, before
the seeds have germinated. Soil temperature can be a limiting or critical factor for seed
germination, root elongation, tuber development. It also affects decomposition of
organic matter within the soil, and thus for the amount of carbon dioxide passing from
the soil into the plant-air layer and into the atmosphere. Table 6.1 lists a number of
optimum and minimum soil temperatures for seed germination given by Schnelle
(1948; 1955), Blanc (1960) and Stanhill (1965).
Table 6.1
Crop Optimum Minimum Crop Optimum Minimum
Optimum and minimum soil
temperature (°C) for seed Clover - 1 Field beans 21 10
germination. Flax - 1.8 Com (maize) 20 8-9
Wheat 15-18 3-4 Cotton 21 -
Barley - 3-4.5 Sorghum 25 -
Figure 6.1
Sugar beet 16 4-5 Soy beans 25 8-9
The effect of soil temperatures
Peas - 7-8 Tobacco 29 13-14
(T) on the duration in days (t)
Potatoes - 7-9 Gourd (pumpkin) - 12
of emergence (I) and germi-
nation (II) (After Geslin, 1944).
The upper temperature limit for seed is usually about 30°C; for com the upper
1 .I I I I
II 5 cm depth Tt = 121 .6 temperature limit is higher than 40°C. Such figures are usually established in a
70
laboratory under steady temperatures and investigations under field conditions -
60 with changing temperatures - are rare (see section 8.6). However, in climates with
a cool season, the temperatures given in Table 6.1 define the sowing dates. If soil
~ 50 management results in an increase of soil temperature in springtime, it permits
"'C
0 \ earlier sowing.
<ii 40
Geslin (1944) carried out sowing experiments at constant temperatures in
..c
E
~ 30
\ climate chambers and at diurnally varying temperatures in the field. Figure 6.1
presents the results obtained. It is evident from this figure that a small increase in
\1I O'c~ depth Tt =40.6

'~- ----
20 soil temperature near the minimum temperature produces a large decrease in

10 \ ~........_ -.........
time for emergence. On the other hand, at higher temperatures, further increases
produce relatively little effect.
T
Soil temperature is also important for plants from which parts are harvested
0
5 10 15 20 25 below the soil surface, e.g. the tubers of potatoes develop best at temperatures of
Temperature
about l 8°C. Their growth is reduced at temperatures of above 20°C, and is stopped
entirely at about 28°C.

6.2 SOIL CULTIVATION Soil cultivation is a permanent chore. Its contribution to better soil climate by
AND TREATMENT improving soil structure is usually temporary, because both weather and agricul-
tural practice can again degrade the soil, requiring repetition after some time.

79
WMO LECTURE NOTES-AGRICULTURAL METEOROLOGY

Only subsoiling may have a more permanent effect, in particular if there is an

impermeable subsurface layer. In the past, the major functions of soil cultivation
were the creation of a seed bed and control of weeds. Although mechanical weed
eradication can be replaced by other methods' of weed control such as burning or
using weed killers, seed bed preparation remains important, because after sowing
the possibilities to do something about the state of the soil are much less.
Cultivation may increase the porosity of the soil and consequently improve
its aeration and increase its temperature; it may also improve drainage.
Cultivation of the soil also influences climatological conditions in soil and air,
evapotranspiration, as well as the air temperature near the ground and its daily
variation. Discussion here will centre on soil state changes related to changes of
surface albedo, by changes induced by a shallow 'mulch' soil cover, and as a result
of reshaping the surface as ridges and furrows.

6.2.1 Effect of surface colour Differently coloured soil surfaces have different coefficients of reflection of short-
on soil temperature wave solar radiation and hence different proportions of solar energy are absorbed
by the surface. So net radiation of the soil surface can be changed by
References: Geiger (Section 20),
altering surface colour. (Because infrared albedos are always around 1, we can do
Oke (1987).
little to control long-wave exchange). The application of coloured surface
dressings is used in agriculture to warm the uppermost layers of the soil
or to prevent high temperatures which are dangerous for young plants and
Figure 6.2 seedlings.
Reflectance of Newtonian silt When the whole spectrum is considered, marked differences in reflection of
loam as a function of diffuse radiation from natural surfaces only occur between the extreme condi-
wavelength and water content tions: e.g. black cotton soil, or soil covered with soot or carbon black on the one
(Bowers and Hanks, 1965). hand; and, on the other hand, light-coloured, dry sand, chalk and, in particular,
fresh snow. The effective colour of a soil changes with its moisture content
60 % reflectance (Bohren, 1987; see Geiger, Table 4-3). An illustration of this is given in Figure 6.2.
50
Surface dressings make it possible to increase or decrease soil temperatures.
To indicate the range of possibilities, Indian data (Ramdas, 1961) are referred to_in
40
Table 6.2. Local black cotton soil (control) was covered by very thin layers of soils
30 from other regions. Albedo values are given where known, and also soil tempera-
ture reductions ~T at the time of the diurnal maximum, when the control surface
20
temperature was 62°C.
The temperature reduction measured over grass vegetation is . greater than
1'.0 1.5 2.0 2.5 that from (white) chalk powder for two reasons. The vegetation is actually shading
Wavelength (µm)
the soil surface; also, grass transpiration requires more energy than bare (dry?) soil
radiation.
In another trial, Ramdas and Dravid (see Geiger, Figure 20-1) applied a thin
. Table 6.2 . 'layer of white powdered lime on black cotton soil. The surface temperature
Changes in soil temperatures decreased by about 15°, and at a depth of 20 cm the soil temperature decrease was
resulting from changes in soil still 5°.
colour (Ramdas and Dravid, Stanhill (1965) used a surface dressing of magnesium carbonate in order to
1936). reduce the daily maximum temperatures of the upper layers of the soil by up to

Temperature reductions

Soil Colour Albedo !ff(surface) i1T(S cm deep)

1. Gras-s-covered soil Green 0.32 26.5 ?
2. Fresh chalk White 1.0 14.2 6.8
3-. Trivandrum sea sand Nearly white 0.78 11.4 5.0
4. Soil from Sakrand in Sind Ash-coloured 0.41 5.1 2.3
5. Soil from Lyallpur Light brown ? 3.9 1.8
6. Soil from Pusa Ash-coloured ? 4.5 3.1
7. Soil from Mekrand Yellowish-brown ? 5.2 2.1
8. Soil from Bangalore Red ? 3.5 1.6
9. Soil from Sholapur Black 0.04 -2.5 -0.3
10. Poona cotton soil (control) Almost black 0.16 0 0

80
CHAPTER 6 - AGROMETEOROLOGICAL MANAGEMENf AT MICROSCALE AND TOPOSCALE

Table 6.3
Energy balance for bare and Bare soil Whitened soil
whitened soil, daytime (albedo 0.3) (albedo 0.6)
averages during July and
Kl. Incoming short-wave radiation +27.2 +27.2
August. Units: MJ m-2 day-I
K* Net short-wave radiation +19.0 +10.9
(Stanhill, 1965).
L* Net long-wave radiation +12.8 +9.8
Q* Net radiation at soil surface +6.2 +1.1
LE Evaporation (latent heat exchange) -4.1 -3.4
H Sensible heat exchange -1.9 + 2.5
G Soil heat flux -0.2 -0.2

10°C for a period of three weeks. In Table 6.3, the change in the energy balance
due to the change in albedo from 0.30 to 0.60 is presented (note the oasis effect!).
In temperate climates with a cold season, dark materials can be applied to
the soil to increase the soil temperature in springtime for earlier germinating. Van
Eimern (1968) sprayed dark bitumen mulch on soil in which corn (maize) had
been sown. The effect of this treatment is presented in Table 6.4. The increase in
yield was significant.
Table 6.4
With spray No spray Difference
Effect of dark bitumen surface
spray on growth and Sowing date of corn, 1962 30 April 30 April
development of maize Emergence of plants 13May 21 May 8 days
(Source: Van Eimern, 1968). Blossom of panicles 30 July 6 August 7 days
Height of plants 12 July 0.84m 0.60m
30 July 1.73 m 1.53 m
9 August 2.18 m 2.00m
Water content grains 10 October 42.8% 44.8%
5 November 39.9% 41.9% (harvest)
Yield of grains at 86% dry weight 6 560 kg/ha 6 220 kg/ha 340 kg/ha

Many more examples of the practical exploitation of alterations in albedo are

quoted by Geiger (section 20). Examples include vertical walls painted in different
colours, against which crops (tomatoes, peaches) were grown. Against the dark
wall, which was warmer, the growth of wood was stronger, whereas reflection of
the short-wave radiation from a light coloured wall resulted in more buds and
increased yields.

6.2.2 Mulching Mulching is the application, or creation, of some soil cover which reduces vertical
transfer of heat and water vapour. Mulches vary in surface albedo, in thermal
References: WMO-TN No. 136;
conductivity and capacity, and in porosity to vapour, and all of these are important.
Geiger (Sect .19); Rosenberg,
Mulches may consist of:
et al. (1983); Oke (1987).
(a) A loose layer of topsoil (harrowing produces a mulch, apart from other effects);
(b) Cut or gathered vegetable material, such as grass, weeds, straw, tree leaves;
(c) Redeployed surface material, such as stubble, litter, and stones;
(d) Manufactured materials, such as paper, plastics, and reinforced aluminium foil.
The growing practice of employing artificial mulches by utilizing plastics
requires special mention. The use of relatively rigid sheet material renders it possi-
ble to place it either directly on the ground (with a thin layer of air between the
covering and the ground), or deliberately supported a few centimetres from the
soil surface. In this latter case, some of the processes of low glasshouses become
operative.
Prolonged snow cover also operates in many climate situations as a mulch,
protecting the subsoil from intense cold. It also has the practical advantage over
other mulches as it applies itself in early winter and removes itself in springtime.
Usually, mulch application must be a temporary measure, because crop require-
ments vary significantly in the course of their growth cycle.
The thermal conductivity of mulches is low - in case of heat-conducting
covers, like aluminum foil, the non-convecting air layer below the cover acts as a

81
WMO LECTUR'E NOTES - AGRICULTURAL METEOROLOGY

heat flow barrier. This means that daily, and yearly, surface temperature variations
are poorly transmitted downwards, so the subsoil is sheltered from short-term
temperature excesses, from midday or summer heat, and from night-time or
wintry cold.
Evaporation from the soil is reduced by mulching because of decreasing
availability of vaporization energy below the mulch (Stigter, 1984). Moreover,
most mulches are an effective barrier to the upward flow of water vapour. If the
mulch is a loose-layered topsoil, resulting from cultivation, another reason for
evaporation decrease is the rupture of capillary connections, inhibiting the
upward motion of deep soil water.
If incoming radiation cannot be well transmitted into the soil, it heats up the
surface. Similarly, if at night the outgoing surface radiation cannot be replenished
from below, the surface cools more. As a result, the diurnal temperature variation
just above a mulch is greater than it would be on the surface of bare soil. This
difference also depends on the albedo of the mulch. One of the disadvantages of
mulches is that, as soon as young plants extend their tips above the mulch, these
suffer extremely large daily temperature variations. At that growth stage, removal
of the mulch may be necessary.
An illustration of the different action of various types of mulch is given by
Waggoner, et al. (1960), who compared three cover types over fairly dry soil with
a low thermal conductivity (0.42 W m-1 K-1) at local noon (Sun elevation 71 °).
The table gives a good idea of the size of typical energy fluxes in the field. The
screen air temperature was 25°C.
Table 6.5
Surface type Bare soil Black plastic Paper Hay (0.06 m)
Noontime thermal budgets
over mulches in Hamden Albedo 0.24 0.08 0.42 0.22
(Conn.) on 11 June 1959. 0
Surface temperature ( C) 38° 50° 40° 50°
Heat flux units: W m-2 Radiation budget
.(Waggoner, et al., 1960). K* 819 993 631 840
L* -177 -282 -199 -233
Q* 642 711 432 607
Energy balance
Sensible heat flux (H) 363 635 349 488
Latent heat flux (LE) 195 0 42 84
Soil heat flux (G) 84 77 42 35

The choice of a mulch, its conductivity and its albedo, certainly requires
much thought. Black plastic and hay heat up equally highly, the plastic because
'of low albedo, the hay because of low conductivity. Hay and paper have a signifi-
cant soil cooling effect. All mulches reduce evaporation, even hay. A new
aluminum~film cover has a high albedo and behaves like paper mulch; old
aluminum has a very low emission (\Jf=0.5). White polythene sheet leaves soil
temperatures nearly unchanged, while black and transparent polythene results in
soil temperature increase.
Mulch effects upon surface and soil temperature are illustrated by many case
histories in Geiger (section 20). One example is an experiment in Israel to ascer-
tain whether a mulch of light-coloured straw could lower the undesirably high
surface temperature, so as to permit planting of a second potato crop in August -
instead of delaying until September, as is the customary practice. In this particu-
lar case, a useful lowering of soil temperature was achieved, but the air
temperature increase at 0.1 m above the mulch was too large for the desired
purpose. At night, soil-mulch interfaces were warmer than the bare soil surface,
though the mulch surfaces were cooler.
Obviously, there may be disadvantages in using mulches. For instance, it
can be quite harmful when the activity of soil microbes is restricted, either by
lowering of temperature, or by the depletion of nitrogen. In the latter case,
bacterial activity is concentrated more in the organic mulch than in the soil
itself. Mulches can also alter the composition of microfauna.

82
CHAVfER 6 - AGROMETEOROLOGICAL MANAGEMENT AT MICROSCALE AND TOPOSCALE

A major negative effect of mulching is the increase of frost risk to crops. In

practice, the first priority in reducing frost danger in a plantation or orchard is to
remove loosely packed vegetation litter - such a layer conducts heat very badly.
Other frost-preventive action consists of increasing the contact between soil parti-
cles by compaction, in order to increase the conductivity for soil heat flux. Soil
compaction should perhaps be seen as an inversion of mulching. One investiga-
tion indicated that, in a European climate, the loosening of the upper soil layer to
a depth of two centimetres will lower the night minimum surface temperature by
2°c for moist soil and 3°C for dry soil.
Mulching has other side-effects. A positive effect, often a reason for
mulching, is reduction of wind erosion, both by surface shelter and by keeping
soil moist longer (section 6.4.2). Also the effect of cover on runoff and on perco-
lation into the soil may be important.
The problems of assembling and applying large quantities of material to the
ground are obvious. In the case of vegetable mulches, other than remnants of
previous crops, such as straw, the area required to produce the mulching material
may be considerable. In many cases, similar benefits can be obtained with cover
plants (see section 6.3.2).
In some projects, the thermal behaviour of a layer of a particular soil is
changed permanently by the addition of a soil of a different type. For example,
bog soils are improved by adding sand, either on top or mixed in, in order to
increase their thermal diffusivity and to change the evaporation (Geiger, Table 19-
3). For evaluating such soil improvement methods, physical principles are similar
to those for mulch evaluation.

6.2.3 Surface geometry We will now review the influence on microclimate of what might be called micro-
effects on temperatures topography, i.e. the effects of small terrain irregularities such as mounds, ridges,
furrows and trenches, with vertical dimensions of at most a few metres. Their
Reference: Geiger (sections 41,
main effect is to alter the amount of solar radiation (in particular direct radiation)
42), Oke (1987).
striking the surface, with effects upon soil temperature and soil moisture. The
germination and emergence of a considerable range of crops can thereby be
altered.
Daytime geometrical relationships between direct sunshine and slope and
aspect of small ridges and furrows are comparable to those discussed in
section 5.5. In particular, their effects are very dependant on their orientation
with respect to slopes and the position of the sun. In the early morning the tops
of ridges are colder than at ground level, both because of excessive sky view and
because of the (relatively) large distance to deep soil heat. Thereafter sunny slopes
heat up appreciably, depending on orientation, also because reflected radiation
may still be absorbed at the other side (radiation trapping). An interesting situa-
tion has been observed on very dry North-South oriented ridges, where plants
placed on West-facing slopes survived because in the early shade they were able to
make use of occasional morning dew.
Nocturnal temperature relationships in small hollows, trenches (with linear
and vertical dimensions of a few tens of centimetres at most, typical of furrows)
may be different from those observed in cases of dimensions of tens of metres. At
the smaller scale, five factors need attention:
(i) The gravitational flow of cold air into shallow 'frost hollows' (section 5.5.4);
(ii) The reduction in turbulent exchange because of the shape of the hollow;
(iii) Shortening of the period during which the Sun actually can reach the soil
surface ('apparent daylength');
(iv) The decrease in long-wave loss due to diminished exposure to the sky;
(v) The heat supply from the sides of the hollow.
Factors (i), (ii) and (iii) tend to cause lower temperatures, while (iv) and (v)
tend to reduce the nocturnal fall in temperature - (v) will only work at small
scales.
With respect to shape effects in (ii), it is interesting to cite an investigation
on snow accumulation in excavations of about 5 m depth (Anon, 1950). This
showed that they were 'self-cleaning' (i.e. of snow), only when the slopes of the

83
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

sides were less than 10° to 15°, depending on other factors such as wind direction.
So for steeper side angles, turbulent transport decreases significantly.
Geiger (section 42) cites some results obtained by Bolz, in which minimum
temperatures were 0°C on level ground, and respectively -3°C and -6°C in depres-
sions with side slopes of 15° and 45°, while in a steep-sided (75°) hollow the
temperature was +2°C. In another series of observations (Brocks, 1939) in symmet-
rical trenches of 20 cm depth and uniform width at base, minimum temperatures
at 10 cm height in the trenches on a particular night were respectively:
Trench side slope .15° 30° 45° 60° 75° 90° (vertical)
Temperature ( C) 0
6.6 7.0 7.3 7.5 7.5 8.1
compared with 6.3°C on adjacent level ground. Other cases were recorded (Geiger,
section 42), in which minima in a vertical-sided hole were generally higher than
in one of 30° slope, which, in turn, were higher than that on level ground. An
interesting finding was that the heat conducted from the side was in exact agree-
ment with the variation in ground temperature, and was therefore large in
· autumn when the soil still retained its summer warmth, and small in spring.

~3 CROP We have so far discussed managing soil by colouring it, covering it with mulch or
MANAGEMENT AND changing its surface shape. Temperature plays an important role as it influences
LAYOUT the germination of seeds and the growth of young plants. Next, we have to look
at a surface with appreciable crop vegetation, and this implies different
References: Geiger (sections 41,
approaches. The direct accessibility of Photosynthetically Active Radiation (PAR)
42), Oke (1987), Benincasa,
to leaves becomes important, particularly when there is little of it at high lati-
et al. (1991).
tudes. In other circumstances, excessive heat or radiation can be a danger,
depending on the type of crop and its growth stage, e.g. by causing excessive
water stress. In such circumstances, vegetative or artificial cover becomes neces-
sary, with the additional benefit that it can diminish wind erosion.
An example of the dependence of photosynthesis on solar radiation is
presented in Figure 6.3 (from Monteith, 1968) for two agricultural crops. The
diagram, based upon hourly mean rates of photosynthesis of barley and beans,
clearly illustrates the increase in photosynthetic activity with increase in short-
wave radiation.

6.3.1 Spacing of crop rows When the crop covers the ground completely, the absolute value of temperature,
humidity and winds at the surface, and the vertical gradients of these elements are
decisively altered by the crop itself. On horizontal ground the spacing of row
crops, and the seasonal changes in the ratio of crop height to distance between
the rows, drastically affects the amount of sunshine reaching the actual soil
·surface. When discussing this and related topics, it must always be borne in mind
that radiation phenomena depend upon the ratio of direct and diffuse sunshine.
In overcast weather some of the differences due to orientation vanish .

Figure 6.3 5 • •
Gross photosynthesis of CO 2
by crops, plotted against total
solar radiation (• = barley,
'.c 4
•
26-6-63; ti.= beans, 21-6-61). N
I
• !).
!).
~3 •
• •
!).
0 !). !).
u !).
0)
2

•
!).
1

200 400 600 800 1 000 Wm-2

84
CHAPTER 6 - AGROMETEOROLOGICAL MANAGEMENT AT MICROSCALE AND TOPOSCALE

Figure 6.4 North-South East-West

21h 21h
Sunshine duration in rows of
plants for varying relative row 18h 18h
distance, for two orientations
(Kaempfert, 1951). 15h 15h

12h 12h

9h 9h

6h 6h

3h 3h
I I I I I I I I I I I I I I I I I I I I I I I I
J FMAMJ J ASOND J F MAM J JASON D

The general form of the relationship between the possible hours of direct
sunshine reaching the ground and the ratio e, where:
e = (height of crop)/(distance between rows)
is indicated above in sketch diagrams (Figure 6.4) for two orientations, N-S and
E-W, and for different times of the year in mid-latitude regions. These Kaempfert
diagrams, given by Geiger (section 41), assume that the crops can be considered as
vertical walls. These two patterns are symmetrical about local noon and about the
summer solstice; for other orientations (e.g. NE-SW) the pattern is asymmetrical.
When E = 2 (i.e. height of plants = twice the width between rows), the direct
sunshine can reach the ground only during the time covered by the unshaded area.
Crop yield per unit area is affected by both the density of planting and the
production of individual plants. If densities are higher than optimal, then the
light, which can still penetrate to lower parts of the crop, suffers too much inten-
sity reduction in the photosynthetically active part of the radiation spectrum. In
addition, movement of air within the crop decreases with increasing density,
which may reduce the CO 2 concentration near leaves, and limit photosynthesis
on sunny days with weak winds. In arid conditions, North-South orientations
seem to give better yields (Anda and Stephens, 1996).
For sorghum, a number of experiments (see Thomas, et al., 1981, and Graser,
1985) indicate that, after correction for population density, different row spacing
has no dramatic effects if sufficient water is available; in arid conditions, wide
spacing is better.
For rice, Johnson (1965) shows the association of spacing with both grain
yield per plant and grain yield per hectare in Figure 6.5. It appears that modern
rice plants can be transplanted at any distance from 10 to 35 cm apart without
significant differences in yield per hectare. Figure 6.6 gives a similar picture for
clover (Donald 1963).
Grain 4 000
Figure 6.5 150 weight/plant (g)
Grain weight per plant of rice Grain yield
and grain yield per hectare of (kg/ha)
rice, at spacings from 7 cm to 3 000
100 cm Qohnson, 1965).
sin 90X)2 ·11
y= 128 ( 85
2 000

50

1 000

0 50 100 0 50 100
Spacing between plants (cm) Spacing between plants (cm)

85
WNfO'L'ECTURE NDT.ES - -A:GRICO'VfUID\'L7MET.E0"RObOGY

'Figure ·6.6
Changes with time in dry
weight per plant of trifolium
subterranea at different
densities for two years (Donald,
1%3).
1.6 plants
.:per ·d m 2
~• 36 plants
per _dm 2

fo
I
20
I I
-.60
I I I
100 140
,
;It

I
...
; I
·20
I I
·60
I .1 1
100 140
I I

Days from sowing

6.3.2 Shading, cover crops The practice of shading (i.e. interposing some type of screening against solar radi-
and weeding ation) is employed in both animal and crop husbandry. Provision of a vegetative
cover also has macroscale implications as a protection against wind and water
erosion.
The influence of shading on the microclimate is rather complicated. A
considerable fraction of total global radiation is reflected by the surface of trees
providing shading or roofs. For trees with a closed canopy, reflection factors vary
from 16 to 37 per cent (vanW¼jk, 1963)."Jl;he reflection of dry leaves is higher than
that of the same leaves when-wet. The non-refle:tted part of the global radiation is
partly absorbed by trees providing shading or by roofs and partly transmitted. If
shading is provided by a roof made of dead material, most of the absorbed energy
will be transformed into thermal energy.
It should be borne in mind, that shading provided by green leaves not only
reduces the total amount of solar radiation reaching the surface of the shaded
plants but at the same time changes its spectral composition. The principal change
in the quality of the radiation is caused by the preferential absorption of radiation
with a wavelength smaller than 0. 75 µm and within the visible range in particu-
lar, in the blue and red regions of the spectrum (Leopold, 1964). If shading is
provided by means of trees, these will use half of the absorbed radiation with
wavelengths of approximately 0.4-0.7 µm for photosynthesis and transforms most
of the remaining fraction into thermal energy.
Therefore, the daytime microclimate of shaded plants is characterized by:
(a) Lowertemperature of air and soil;
(b) Lower light intensity;
(c) Poorer light quality for photosynthesis;
(ii) Reduced evapo.tran~piration.
During the night, ·however, the same Shading will restrict the net outflow of
.terrestrial ·radiation, giving rise to:
(a) Higher temperatures of air and soil;
(b) Increased evapotranspiration.
In addition, shading will tend to reduce the vertical and horizontal move-
ment of the air. Depending on local conditions and the type of shading, the daily
mean temperature of air and soil may be somewhat lower, but the diurnal varia-
tion of the daily temperature is always smaller. For soil moist_ure, tpe overall effect
of shading is complicated because shade trees evaporate intercepted rainwater and
transpire soil water. If rain is plentiful and regular, the influence of shading on
moisture conservation is not likely to be important; if rainfall is marginal, ·living
shade can often be harmful.
The "idea that reduction of the light intensity is required to create optimum growth
conditions for perennials, ·such as cacao, coffee and tea, has been contradicted .by
~perimental evidence. Shading is shown to be only advantageous'in permitting modest
yields to be obtained-under marginal climate and soil conditions.
In lower latitudes, where shadin_g i-s a widespread practice, temporary
shading may be _provided Jor individual plants or for crops. For individual plants,
this usually consists of a few leafy 'branches or palm leaves stuck in the ground,
occasionally of material such as wooden slats or baskets. Temporary shading can

:86
CHAPTER 6 - AGROMETEOROLOGICAL MANAGEMENT AT MICROSCALE AND TOPOSCALE

also be provided for crops by means of a roof made of locally available materials
such as palm leaves, grass, sticks, bamboo slats, or industrial products such as
Saran netting.
Permanent shading is achieved either by means of shade trees providing
overhead shade, or by hedges providing lateral shade. This applies both to
shading of annual crops such as vegetables or tobacco and to perennial crops.
When the cover is in the form of bushes, trees (or even forests), it is a significant
factor in land use.

The complexity of the changes in microclimate following the use of shade or

cover become more marked as we pass from simple devices, whose initial
objective is to reduce the intensity of short-wave radiation reaching the foliage,
to complex devices with more purposes than shading only - ranging from
removable low glass or plastic structures (cloches) to the controlled environment
of fully equipped glasshouses.
Some results of microclimatic changes under the relatively simple shade
tents of tobacco plantations are instructive. The shade is provided by white cotton
cloth erected some 2.5 metres above the soil surface on wooden poles. The actual
fabric covers about 45 per cent of the area. It is reported that the shade tent
improves the colour, yield, and burning index of the leaves. Among findings of a
study by Waggoner, et al. (1959) are the following:
(i) Due to the interception and distribution of radiant energy, daytime soil
temperature is slightly lower inside than outside the tent; the reverse case
occurs at night except in overcast conditions;
(ii) The moisture comparisons depend upon which measure, of humidity -
absolute (represented by dew-point) or relative values - is examined. At
noon relative humidity was about the same, but the dew-point was approx-
imately 1°C higher in May, and 2.5°C higher in June under the tent. During
early morning and late evening the inside air becomes saturated. The low
wind speeds and higher vapour pressure (the result of higher inside temper-
atures) lead to a reduced evaporation compared with outside.
Wind speed differences were not reported in a way that was useful for analysis as
no good specification was given of the observation situation outside the tent.
As already mentioned, a bush or tree canopy also reduces the net loss of
long-wave radiation to the sky from the shaded crops. In frost-prone situations,
shading may be used to reduce possible damage to seedlings, young crops and
low crops generally.

Cover crops are much used in the tropics to protect the soil from potential erosion;
almost any vegetative cover is better than none. In addition, the cover crop is
used as a 'nurse' crop to protect young seedlings, other vulnerable plants, etc.,
against some of the rigours of the external climate. Finally, the widespread prac-
tice of multiple cropping results in the tallest crops being used as cover for the
others (Baldy and Stigter, 1997).
The principal meteorological effect of a cover crop is a reduction in the
diurnal variation of temperature of the top soil. Covered soils are cooler during
the day and warmer during the night. Consequently, the microclimate near the
soil is cooler during the day and warmer at night. However, cover plants compete
with main crops for water and if their growth is insufficiently controlled they may
even compete for light. Water consumption by the cover crop is usually of the
same magnitude as that of the main crop. This is, to some extent, counter
balanced by the reduction of surface runoff and the improvement of the water-
holding capacity of the soil.
With regard to water erosion, the cover first intercepts rainfall and so
reduces, or almost eliminates, its impact on the soil (splash erosion by falling
drops is an important first stage in the erosion cycle). Secondly, it reduces the
speed of runoff water and hence its carrying capacity. Cover vegetation is equally
important in relation to wind erosion. The surface wind speed is reduced by the
frictional drag exerted by the ground surface, and the trapping of moving soil by

87
WMO LECTURE NOTE.S - AGRICULTURAL METEOROLOGY

surface obstructions and irregularities inhibits the build-up of an increasing load

of airborne particles. Furthermore, the abrasive effect of such a loaded stream on
downwind objects is reduced. Considerable progress has been made in
quantifying the physics of wind erosion by using expressions involving the
percentage of the area susceptible to erosion that is covered by vegetation.
Permanent covers planted between rows of perennials may consist of natural
regrowth controlled by means of cutting and selective weeding, or of one or more
of the numerous cultivated cover plants. Temporary covers in a rotation crop are
usually leguminous creepers or grasses.

Weeds have very much the same effect on the soil climate as cover crops, but
because of their rapid and luxuriant growth they compete strongly with the crop
for light, water and nutrients. Their influence on the microclimate is very similar
to that of cover plants and perennial crops, as they reduce the air movement and
the diurnal variation of temperature within the crop stand. Sometimes weeds are
allowed to grow to act as a cover or nurse crop.
In seed beds, nurseries and plantings of annual crops, however, competition
from weeds is too strong for the main crop, in particular when it is young and may
easily be smothered. When removed by means of pulling, cutting, cultivation or
chemical means,. weeds may be left on the soil surface to act as a mulch. In drier
regions, another purpose of weeding is conservation of water. Weeds use soil water
and the deeper and more extensive their root systems are the greater the water use.
In some areas, the alternating cultivation of fields to destroy weeds is known as
fallowing. By this means water below the immediate surface, which otherwise
might be transpired, remains:
On the Canadian Prairie, for example, where the practice of fallowing is
frequently followed, land is allowed to lie idle for 21 months between the harvest
of one crop and the planting of another. During this period about 20-25 per cent
of the precipitation during the period is conserved as crop-available soil water.
This conserved water is necessary along with the scanty growing-season rainfall to
produce an economical crop.
However, in many areas the dry soil is exposed to the risk of wind erosion,
which is strongly inhibited by vegetative cover of any sort. A balance has to be
struck between loss of water due to the weeds and retention of top soil due to their
presence. "In (many) locations, fallow may be used safely only in longer rotations
with sufficient grazing or hay crops to protect the soil and keep it in good physi-
cal condition. Water conservation may be enhanced by the practice of surface
mulching which reduces evaporation, mostly during the short period after rain
when the soil surface is moist." (WMO-TN No. 138)

6.4 WIND SHELTER The basic properties of windbreaks and shelterbelts were discussed in Chapter 5.
In this section some particular aspects relevant to agriculture will be
6.4.1 Effect of wind shelter considered, in particular the effects of windbreaks on microclimate, erosion and
on microclimate deposition.
References: WMO-TN No. 59;
Vegetative growth at 'zero' wind (e.g. as experienced in glasshouses or under low
Rosenberg, et al. (1983); Geiger
glass cover) is luxuriant. There is typically a reduction in vegetative growth as the
(sections 37, 52/53); Marshall
wind increases to small values, that is, 1 or 2 m s- 1 . The temperature of irradiated
(1967); Grace (1977);
plants varies according to net heat exchange by convection between plant surface
Oke (1987).
and air. It is lowered . by evaporation and raised by condensation. Gates (1968)
found that in microclimatic investigations on the leaf temperature, as distinct
from the temperature of surrounding air, there was clearly a critical difference
between still air and a minimal level of air movement of the order of magnitude
of 1 m s- 1 inside the crop.
At a slightly larger scale, an adequate degree of air motion ('ventilation') is
necessary within a crop stand and around plants, if transpired water vapour is to be
removed and an adequate supply of CO 2 is to be maintained for photosynthesis.
What's more, in humid regions damp, stagnant air frequently provides favourable
conditions for diseases and pests to flourish and to attack the vegetation.

88
CHAPTER 6 - AGROMETEOROLOGICAL MANAGEMENT AT MICROSCALE AND TOPOSCALE

Optimal wind speeds vary according to the general climatic regime and the
type of vegetation. The ventilation must be adequate, not more, since
plants can only pass water at a limited rate. If wind speeds are too fast, the
stomata close as a consequence of internal water deficiency caused by excessive
transpiration.
In arid and semi-arid regions, the emphasis is on 'improved water efficiency'
associated with shelterbelts (Rosenberg, et al., 1983). Water loss is decreased due to
a reduction of turbulent exchange in the sheltered area (though this is not true at
every point behind completely impermeable barriers). The slower evaporation rate
in shelter provides better conditions for e.g. seed germination.
In arid regions, crops close to a belt may suffer from root competion with the
belt for the available water. Pruning the belt tree roots at a few meters distance
from the trunks may alleviate this problem (Onyewotu, et al., 1994).
Commonly, the level of air temperature in the sheltered zone (up to 10 H or
15 H) during daytime is 2 to 3°C higher in warm temperate regions, and 0.5-1 °c
higher in temperate maritime regions, because it is generally cloudier there.
During clear 'radiation nights', temperatures behind a belt may be lower than in
unsheltered terrain through the creation of a body of still air, partially isolated
from the relatively warmer air which is to be found farther above the surface. On
other nights, no obvious differences are found between sheltered and unsheltered
minimum temperatures (Geiger, section 37 and 52). An interesting dynamic effect
is that the relatively large daytime soil temperatures in shelters imply that at night
more soil heat is available to counteract ground frost. Even so, the ground frost
risk in shelters is larger, except very near the belt (< 3 H distance) because of the
effects of radiation exchange with the belt instead of the sky.
Vapour pressures and, in spite of higher temperature, relative humidities are
higher near shelters, although the increase of the latter rarely exceeds 10 per cent.
The amount of dew observed near shelter is also relatively high, due to more
humidity and less wind.
Soil temperature differences between sheltered and unsheltered plots are
primarily a result of reduced evaporation. Yet, because the heat capacity of moist
soil exceeds that of dry soil, the changes in soil temperature due to wind reduc-
tion are limited and uneven and depend upon type of soil and upon ground cover.
Adjacent to a shelterbelt, additional complications arise due to shading which, in
turn, depends upon type and orientation of the belt. For North-South oriented
belts the net difference is small, since morning shade is compensated by increased
afternoon reflection, or vice versa. Table 6.6. cites daytime increases of between
0.5°C and 2°c at 0.1 m to 0.2 m depth; the greater the wind reduction, the greater
the temperature increase.
Table 6.6 Downwind distance from shelter belt
Mean differences of soil Depth (m) 7H 14 H 21 H
temperature ( C) between
0

0.02 0.6 3.7 1.1

protected and open locations 2.0
0.05 0.7 3.4
(Potsdam, 8 March-31 July 0.10 0.5 2.9 2.2
1953).
The overall increase of crop yields due to the beneficial effect of shelter varies
with the crop. For fruit, yield changes of around 50 per cent are cited; for field
crops, around 20 per cent.
In investigations of turbulent exchange behind shelterbelts, a practical
problem is that some standard micrometeorological relations cannot be applied in
the non-equilibrium conditions of a belt wake. This may be one reason why opin-
ions differ on the existence of shelter influence on CO 2 concentrations.

6.4.2 Windbreaks against Apart from the destruction caused by gales (U > 14 m s- 1) and hurricanes, less
damage and erosion dramatic winds can also produce damaging effects. As wind speed at the standard
height increases to above 5 m s-1 and higher, increasingly severe effects can be
References: Geiger (section
observed. Many of the advantages of wind shelter arise from the reduction in
52/53); Greeley and
mechanical damage and deformed growth.
Iversen (1985).
89
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

Wind speeds around 5 m s-1 can provoke mechanical damage from the fric-
tional rubbing between different plants or parts of the same plant. Sustained winds
above this level can decrease the growth of plants; even lower wind speeds are
damaging to very delicate crops, such as flowers or horticultural crops (including
fruit trees). Higher sustained speeds influence the shape of plants, giving rise to the
characteristic wind-shaping of trees in exposed positions (see Yoshino, 1975).
Under these conditions smaller and thicker leaves tend to develop and bushes and
trees are subjected to natural pruning at the windward side. COz assimilation at
10 m s-1 proceeds at only 10 per cent of the rate recorded during calm weather.
Where average winds exceed 12 m s-1, plants assume a low, dwarf-like form.
Based upon information from the wind-deformation of trees in the British Isles,
Gloyne (1964) has suggested that in regions where the mean annual wind speed at
standard height exceeds about 5 m s-1, wind shelters may be beneficial. They
certainly are if the mean annual wind speed exceeds about 8 m s-1 are essential where
the average exceeds 10 m s-1 if acceptable crops are to be maintained.
Because of the reduced forward speed in the zone of influence of the wind-
break, airborne material (weed seeds, small insects, spores, i.e. agents associated
with pests and diseases) tends to drop out of the general airstream and/or be
deposited by eddies and concentrated in the sheltered zone. Quieter conditions
will often lead to improved microclimate, to better pollination and to more effi-
cient application of insecticides and pesticides, but the disadvantages (higher pest
and spore concentrations) must be recognized (Lewis, 1965).
In addition to effects on crops, wind also influences operations. Overhead
irrigation and the spraying of dust or liquids is more effectively carried out when
the wind speed is low. In coastal regions, shelters can significantly reduce the
damage arising from sea spray and salt particles (see section 7.5.3).
Strong winds blowing over loose soil displace particles, cause erosion if the
particles do not stick together (no cohesion). At a certain threshold wind speed
Un, the upward lift-and-drag action of the turbulent wind on the particle will
exceed the downward pull of its weight. Particles then move forward by saltation
(a short leap, followed by a downward flight of maybe 0.5 m) or by creep (rolling).
At a higher wind speed Utz, the upward turbulent action on particles also exceeds
their average fall speed, so that the particles can fly in suspension for an indefinite
time. Values of the threshold wind speeds Un and Utz are proportional to ✓(Pp D),
Figure 6.7 where Pp is the density of the particle and D its diameter. For sand of D :::::0.2 mm,
Qualitative relationship Un~ 6 m s-1, depending on the cohesion: a few per cent of water can triple Un.
between wind erosion and Generally Utz is more than twice Un (Kind, 1990). As is well known, dust particles
presence or absence of with diameters of 50 µm or less can be lifted in dust storms to great heights and
vegetation at various levels of travel hundreds or thousands of kilometres (Meteorological Department, U.A.R,
mean annual rainfall '1970; Reiff, et al., 1986).
(Marshall, 1972). Wind erosion of exposed soil is much influenced by its moisture status,
which affects cohesion. Regions with low mean annual rainfall, particularly those
with less than about 250 or 300 mm are particularly vulnerable to erosion
(Hudson, 1971). Figure 6.7 sketches the qualitative relation between the mean
annual rainfall and the relative vulnerability of covered and bare soil surfaces to
erosion.
Shelterbelts can trap wind-driven soil and inhibit or prevent mass transport
of material through wind erosion. In arid regions, shelter is used to reduce inva-
sion of wind-blown sand and soil. The effect of windbreaks upon sand deposition
and sand cover varies with the type of belt (Kaiser, 1959; see also Geiger). At dense
windbreaks, wind-driven sand accumulates on the windward side and peaks at the
10 20 30 40 50 (ins)
windward edge of the belt; if the belt is porous, most sand is deposited in the rela-
Mean annual rainfall
tively quiet conditions behind the belt. In order to effectively reduce deposition
in the long run, the belt must be densely planted with a thickness of at least 30 m,
and must consist of trees such as eucalyptus and tamarisk which are able to survive
partial burial (Mohammed, et al., 1996).
Erosion of cultivated arid soil can also be reduced by keeping the surface
covered by leaving crop residues in the field as a mulch. In the African Sahel
region, where wind erosion damage to a large extent contributes to loss of

90
CHAPTER 6 - AGROMETEOROLOGICAL MANAGEMENT AT MICROSCALE AND TOPOSCALE

nutrients from the top soil layer (Sterk, et al., 1996), mulching possibilities are
limited by low biomass production, combined with many other uses for crop
residues. Experiments have showed that for efficient topsoil protection at least
1 500 kg ha-1 of millet stalks are required. Coverage with less residual material
can even increase erosion because it increases turbulence and therefore saltation,
while the surface is insufficiently protected (Sterk and Spaan, 1997). Due to the
limited availability of mulch material, regeneration of natural tree and shrub
vegetation is the most promising erosion control method in the long run.
In non-arid regions, unnecessary wind erosion damage can be prevented by
delaying certain soil-loosening actions when the forecast predicts an unfavourable
combination of strong wind and large evaporation.
In temperate regions, shelterbelts can also influence soil moisture through
their effect on snow deposition, giving rise to accumulation in drifts on both the
windward and lee side of a belt. Drifts are usually deeper on the windward side,
but given sufficient snow, drifts will form in the protected zone as well; the more
open the barrier, the less the build-up. In dry regions, such snowdrifts can provide
a useful addition to soil moisture reserves. However, in humid regions large snow-
drifts usually have disadvantages, because they are a hazard to free ranging
livestock and a hindrance to transport. Also cultivation of the underlying ground
in spring can be significantly delayed by the concentration of meltwater. In
protected areas behind a belt, snow deposits are often more even, with consequent
advantages.

6.5 IRRIGATION AND The computation and scheduling of water balances was dealt with in Chapter 4,
DRAINAGE and drought is discussed further in section 7.1. Here we comment on some
practical aspects of irrigation. The basic principle of irrigation is to provide
Reference: Lomas and Levin
sufficient water in the root zone. However, the water requirement of a crop is not
(1979).
a steady quantity, but depends highly on the crop and the stage of its growth
cycle. Almost all crops require sufficient water around flowering time. How much
they need before and after depends on the crop, regional soil conditions and the
regional climate.
In section 4.1 we discussed the questionable value of Slatyer's statement: "In
most crops growth and development can only proceed fully unimpaired, and crop
yield is only maximal, when high water status is maintained throughout the life of
the crop." The dangers of too much water are varied, and we will just name a few.
The first is that water in excess of that required to bring the soil to field capac-
ity will seep to levels below the root zone and/or drain away, taking along
('leaching') soluble soil nutrients, including those previously applied as fertilizer. Soil
degradation through excessive irrigation is another danger. In sub-arid and arid
regions (such as Pakistan), widespread irrigation has led to 'salinity' when the water
table is near the surface (at most a few metres below). The increased water available
dissolves salt from the soil and some of the saline water moves upwards to the
surface as a result of the high evaporative potential - only to evaporate and leave a
salt crust. In addition, in these regions both the risk and extent of erosion by runoff
are increased if the rare, but intense, rainstorms fall on recently irrigated ground.
In both temperate and tropical regions the pores of certain soils with a high
loam content, or other fine-structured soil types, can become blocked with smaller
soil particles in conditions of excess water. The resulting undesirable imperme-
ability of the soil cannot be undone until after harvest time. In the meantime, it
blocks diffusion of oxygen and carbon dioxide in the soil and the absorbing roots
of the crop may die off. Also, excessive irrigation at the time of root development
may lead to the development of a too shallow root system. This makes the crop
highly vulnerable to drought at a later time.

Actual techniques of irrigation influence its effects. The classic low-cost technique
of using ditches involves lateral transport through the soil. On level terrain this
can lead to inhomogeneity in water provision, and a useful variation is to irrigate
the furrows in ridge-and-furrow cultivation (section 6.2.3). The extreme case is
flood-irrigation, which is necessary for rice.

91
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

Figure 6.8 A Irrigation period Following irrigation B Irrigation period Following irrigation

Average effect of above-canopy

0
and below-canopy irrigation
30
on: (A) cooling below ,,.....
2 0~
unirrigated dry-leaf
~
temperature; and (B) on .a
Ill

relative humidity (Lomas and va. 4

E
Levin, 1979). ~
Irrigation
,r-r-1--~,+
6 above-canopy
below-canopy 0
I,

012341234 0 1 2 3 4 1 2 3 4
Number of hours during and following irrigation Number of hours during and following irrigation

Sprinkler- or spray-irrigation, also called above-canopy irrigation, has rather

high application costs. On fast-draining soils it may be the only effective
technique. As a side-effect, it temporarily increases the humidity of the air.
Another important side-effect is that water is applied directly to the leaves, with
two consequences: first, the direct evaporation from leaves produces cooling, and
second, induced leaf wetness gives a risk of fungus infection. Effects of above-
canopy and below-canopy irrigation are compared by Lomas and Mandel (1973)
(see Figure 6.8).
In subtropical climates, dry leaf temperatures may reach 40°C (see
section 2.3) and therefore cooling by 10°C is beneficial. However, spraying should
not continue too long, thus preventing harmfully long wet-leaf periods. At subtrop-
ical temperatures, fungus spores do not need much time to germinate (see section
8.3), and therefore it is better to sprinkle for a few hours only and not much later
than noon. This allows leaves to dry well before nightfall and germinating spores to
be killed by solar radiation.
Spray irrigation is used in the fully automatic mist propagator for rooting of
cuttings. Here an intermittent mist reduces the average temperature of the air and
its diurnal variation, keeps the air almost saturated, and reduces the surface
temperature of leaves by covering them with a film of water. In temperate regions
the temperature reduction of the rooting medium may be an undesirable side-
effect.
On well-drained soils - and these are the only soils suited for irrigation -
irrigation decreases the diurnal variation of soil temperatures. Depending on the
temperature of the applied water, irrigation may temporarily decrease average
daytime temperature of the soil to some extent, and may increase the soil temper-
ature at night.
A most important effect of draining a soil is improved aeration of the subsoil,
leading to an increase of the oxygen content and a decrease of the carbon dioxide
content. Even on heavy soils, drainage is effective in this respect, for although it
does not appreciably increase the air space of the subsoil, it does remove stagnant
water from the crevices and cracks in the soil and allows oxygenated water to
penetrate rapidly into the subsoil. Other consequences of improved aeration are
greater root development, more intense bacterial activity, and promotion of oxida-
tion processes, leading to better granulation.
A bare, well-drained soil will exhibit a larger diurnal variation of temperature
to a greater depth than in the same soil without good drainage, and its average
temperature will be higher. Moreover, the drained soil will have a larger capacity
for storing rain or irrigation water (and consequently suffer less from erosion
caused by surface runoff).
A beneficial result of draining may be the increased ability of a soil to stand
up to compression by farm machinery - a very important side-effect. In many
areas, e.g. maritime North-West Europe, the cropping programme is severely
constrained by the need to minimize, if not altogether avoid, the use of mechan-
ical vehicles either for cultivation and sowing at the beginning of crop year, or for
harvesting and post-harvest activity at the end of the season. The importance of
the 'state of ground' in this connection is discussed later in section 9.3.1.

92
CHAPTER 6 - AGROMETEOROLOGICAL MANAGEMENT AT MICROSCALE AND TOPOSCALE

In livestock farming, excessive trampling by animals can also destroy the structure
of soils containing excessive water.
Draining, of course, increases leaching from soil. A procedure to quantify
leaching in the humid British climate calculates:
excess winter rain = rainfall - evaporation - soil moisture deficit at the end
of summer.
The excess varies from an average of 75-100 mm over the drier parts of South-East
England to over 500 mm in the higher and wetter regions of the north and west.
The variation about these averages from year to year can be as much as SO per cent
and, "in the driest areas about one year in twenty is so dry, that there is no
complete replenishment of the summer soil moisture deficit and there is little or
no drainage and leaching" - a statement which serves as a further illustration of
the necessity in all climatological analyses of considering variability as well as
mean values.

6.6 FROST, AND A variety of theories and criteria exist on the occurrence of low-temperature
PROTECTION AGAINST damage to plants. The question is complicated by the fact that during winter
FROST DAMAGE dormancy plants can withstand temperatures below -20°C without injury.
However, once growth has commenced, temperatures a few degrees below freez-
References: WMO-TN No. 157;
ing point may be fatal in certain states of development of the plant, particularly
Geiger (sections 42, 53/54);
at flowering time.
Rosenberg, et al. (1983).
At present, the most-accepted theory of the injury and death caused by frost
is the formation of ice crystals in and outside the plant cells. The cell sap is a solu-
tion under pressure and freezes at a temperature somewhat below 0°C. But water
in the spaces between cells may be pure or nearly so and therefore freezes at a
temperature only slightly below 0°C. Extra-cellular ice formation is therefore more
common. Details of biological processes damaging protoplasm, cell structure, etc.,
are of less concern in this context than the existence of the phenomenon. Frost
hardiness is a process involving the slow building of protective measures within
the plant and this requires time.
Obviously there will not be one specific critical temperature for all types of
vegetation. In practice, damage due to low temperatures is determined by a
combination of the low temperature and the length of time it persists. Even the
bark of mature forest trees can be damaged by prolonged low temperatures
(Geiger, section 41) . When cooling occurs rapidly, the damage may be greater
(Scharringa, 1976); but the often-used term 'killing frost' is not a term with any
general validity - frost damage depends on temperature, but also on the vegeta-
tion and its recent history.
The occurrence of below-zero temperatures in the growing season of crops
has a variety of meteorological causes. Apart from high altitude, which may cause
frost even in the tropics, we have to consider latitude. In subtropical climates, say
around 30° latitude, large-scale subsidence rules with generally weak winds.
During relatively cold weather, e.g. in springtime, strong clear nocturnal inver-
sions can result with large radiation heat losses at the surface. The same radiation
frost weather may occur at temperate latitudes, say around 50° latitude, because
large blocking highs can occur around the break-up of winter circulation, and also
in a steady westerly circulation travelling high-pressure ridges may cause short
cold spells. But in temperate latitudes there is also the possibility of advective frost,
caused by incursions of arctic air. At even higher latitudes one more variable is of
importance, the variability of day length. At all latitudes, the continentality of a
location must be considered as well in climatological analysis to determine the
risk of frost. Methods of short-term forecasting of frost probability are briefly
discussed at the end of this chapter.

6.6.1 Passive methods of It is not necessary to accept radiation frosts as an unavoidable hazard to
protection against frost agriculture. In areas where both the frequency and intensity of such frosts are
limited, it is possible to eliminate or minimize damage by judicious soil and
Reference: WMO-TN No. 133
plant management practices and, under actual extreme frosty conditions, by
(1974).
efficient action.

93
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

Methods of frost protection can be described as either active or passive.

Active methods are artificial measures applied immediately before and during the
period of actual frost danger. Passive methods are concerned not so much with the
fight against frost itself, but rather with how the damage during the frost period
can be significantly reduced. These methods are applied well in advance of the
immediate danger of frost, and include precautions at the time of planting. The
best time for protecting a crop against frost damage is before planting it.
The main passive frost-prevention approaches are proper site selection and
judicious soil and crop management. The third preventive action, plant selection
and breeding of frost-resistant crop varieties - late-flowering apple varieties,
hardier winter wheat and oat varieties, and more frost-resistant strawberries - is
very important but falls outside the scope of this book. The same is also true of
farming techniques such as pruning to delay flowering, or applying chemicals and
plant hormones to promote frost hardiness.
Of the two types of frosts, radiation frosts and advective frosts, the former
has drawn almost exclusively all attention in frost research. The idea that the
advection of cool arctic air cannot be helped is widely believed. Yet at the time of
site selection, the probability of advective frost should still be considered. If a
climatological study indicates that in the frost-sensitive seasons most damagingly
low air temperatures are associated with certain wind directions, shelter from that
direction is important. Purely advective frosts are reduced near large water bodies,
with their power of smoothing climatic extremes. Radiation frosts, on the other
hand, can still kill cranberries in the dunes at a few hundred meters from the sea.
Even so, the advantage of the proximity of, for example, lakes or warm city areas
should not be overlooked. Even weak radiation frosts can produce quite damaging
results when combined with cold advection.
At a smaller scale two factors merit study, the soil type and terrain elevation
effects. The soil type should be regarded in the context of soil heat flux (see
section 3.2). If for frost-sensitive agriculture a choice exists between a loose peaty
ground and a soil with a more compact structure, choose the latter.
The other factor, terrain effects, can be static (elevation and orientation) or
dynamic (cold airstream behaviour). The difference in frost risk between a valley
bottom and the thermal belt higher up (section 5.5.4) is typical of the effects to
look for in hilly terrain. Also, cool katabatic winds may be obstructed by trees and
shrubs and rough pasture growing along a downslope or a river bank, creating
local accumulations of cold air. At very small scales, so-called 'frost hollows'
(section 5.5.4) should be avoided at all costs for growing crops or fruit. When
hedges need to be planted in sloping country~ e.g. for protection against cattle,
adequate provision should always be made for cold air to drain away when
'descending the slope - and one should be aware that a slope of 1:100, or even
less, can make a large difference to frost risk. An example is given in Figure 6.9
showing the effect of a slope of less than 10 m elevation difference; at the high
end of the site (205 m) no frost damage occurred, while at the lower end
(195-200 m) the crop was totally damaged (Lomas, 1977)-.

Figure 6.9
Effect of terrain elevation (- - -
in.m) on fruit damage (-in%)
at Amiad, 1973.

::_-_->:~
t
N
,
I
.... .... .. .. 0

<04 - 205
0
(lo
.. - - __2QO_
O - - - 195
- - - - - - - -: j iQ
- - - - - --------
' •190
',200 ' . . 195

94
CHAITER 6 - AGROMETEOROLOGICAL MANAGEMENT AT MICROSCALE AND TOPOSCALE

Mapping frost-liable areas is a necessary preparation for introducing a new

crop in any region. It should start with a desk study of topographic maps and
climatological data. Then hypotheses based on such preliminary studies should be
verified by field observations of varying extent - the details depend upon the
availability of instruments and the length of time (preferably a number of frost
seasons) which circumstances permit. The main features of the pattern of
minimum temperatures on 'radiation nights' can already be determined from
observations over a small number of such nights, in particular when local station
networks are supported with mobile temperature surveys (Lomas, et al., 1969;
Zemel and Lomas, 1976). Additional support can be provided by numerical
mesoscale modelling of minimum temperatures (Avissar and Mahrer, 1988).
Results of frost risk analysis must be mapped in great detail to be useful, for
regional planning at scales of 1:50 000, and maybe down to 1:1 000 for individual
plantation siting.
The length of the frost-free growing season, between the average dates of the
last heavy frost in spring and the first frost in autumn, should be estimated locally
with some care. Such knowledge will assist in determining average safe planting
dates for areas subject to frost risk.

The conclusions that can now be drawn after this discussion of soil and crop
management are that it is advisable to optimize the soil situation so that the soil
heat flux can make a maximal contribution to the energy balance in frosty nights.
This may require various measures, some of them long-term, because to serve as a
heat source the soil must be compact and must not be isolated from the air above
it. Ploughing in the autumn, instead of the spring, gives the so~l the opportunity
to settle down in winter to a relatively compact structure. Winter mulch should
be cleared away before crop flowering time. Mechanical weeding can loosen soil
appreciably, and if it is needed in springtime, chemical weedkillers should be
considered as an alternative. If the soil is too loose for any reason, then it is advis-
able to use equipment such as rollers to compact it.
An interesting way of increasing the thermal conductivity and capacity of
soil is irrigating it prior to a threatened frost. This is particularly effective in the
autumn because it is a time of year in which the soil still stores a lot of summer
heat. Furrow irrigation is quite effective for this. The more extreme practice of
flooding has also been adopted. Possible disadvantages are the slowing-up of the
temperature rise in daytime following a frost, and the risk of waterlogging of the
soil and its attendant damage to plant roots and soil structure.
In addition to frost-reducing soil management, crop planting should be opti-
mized. In particular on slopes, they should be planted so as to allow maximum cold
air drainage. Climatologically speaking, the planting date should be chosen (if possi-
ble) so that the probable time of flowering does not occur at the time when frost
risks are largest.

6.6.2 Active methods of Frost fighting methods (the term is correct, a farmer wants to protect his crop),
protection against frost which can be used in frosty weather, can be classified as follows:
(a) Thermal insulation;
(b) Interception of outgoing radiation;
(c) Air mixing;
(d) Sprinkling;
(e) Direct air and plant heating.
Some of them are quite costly. Their application depends on the value of the
crop. References given here are very few, since there are so many that it is better
to refer the actively interested reader to WMO-TN No. 157 (1978) or to Geiger
(1995).

(a) Thermal insulation Any covering roof of a crop (if such a measure is feasible at all) should be of insu-
lating material - a straw mat is a simple example - then the top of the cover will
be cooled during the frost instead of the crop below. Also, the roof should be large
enough to allow development of a 'microclimate' below it with large input of the

95
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

soil heat flux; and that microclimate needs some ventilation. Tying a small plastic
bag over a plant during a frost is an effective way of killing it. These factors
have to be considered when using small structures like paper shields, protective
cones, etc.
A new method is to use a non-toxic protein foam to smother the crop. The
method has the potential advantages of being mechanized (so avoiding the
manual labour common to many other methods). The foam can remain on the
crops for a period if successive frosty nights are anticipated - providing, of course,
that no associated damage is caused to plants.

(b) Interception of outgoing Radiation can be intercepted by creating artificial clouds of water mists or smoke
radiation screens. The efficiency of such a cloud cannot be judged by its visual density as
the particles must be able to trap long-wave radiation from the ground. In theory,
water droplets in an artificial cloud should be more efficient than the particles in
smoke screens, but mist production techniques remain in a development stage.
Significant thermal effects (of a few degrees) are only observed in weak radiation
frosts. Also, the clouds are blown away by the wind, so need to be produced conti-
nously. The pollution resulting from continuous smoke generation makes this a
rarely-applicable technique.

(c) Air mixing During cloudless radiation nights, the air in contact with the ground is cooled and
a cold layer of air is formed near the surface. From the surface upwards the temper-
ature increases, so that the air layer at about 50 metres may be appreciably warmer
than the surface layer of air. This is known as temperature inversion and is the
reverse of the normal daytime situation where temperature decreases with height.
By mixing the warmer upper air above the inversion with the colder layers
below, the temperature of the air layer-near the ground can be raised. This has
been done by the use of wind machines, fans and even helicopters. The success
attainable to combat frost in this way depends on the strength of the inversion.
Moreover it has been shown that for winds of more than 2 m s-1 natural mixing
is already so strong that a wind machine is of little additional use. The price of
usefully large wind machines is high.
A necessary prerequisite before introducing air mixing is a survey of inver-
sion strengths on frost nights. Unfortunately, climatological statistics on
inversions are very scarce and difficult to generate, because standard radiosondes
provide no useful data on surface layer temperatures. In California, in an area with
very strong inversions and rather weak frosts, wind machines have been used to
protect the very sensitive citrus crop, but that does not mean that the method will
have much value in other climates.

(d) Sprinkling Overhead irrigation or sprinkling as a method of frost protection has definite
merits and deserves careful attention. This practice requires a large water supply,
suitable irrigation equipment and good soil drainage. It is applicable only if the
plant can stand the ice load which forms on branches, twigs and leaves.
In this method the latent heat, released when water freezes, compensates for
the heat loss by radiation. When one gram of water freezes, sufficient energy is
released to heat 80 grams by one degree. If a leaf or fruit-bud is covered with a thin
film of water, the heat set free when the water freezes is sufficient to prevent the
temperature of that part of the pla,nt from falling below 0°C. As long as this film
of water is maintained continuously, the temperature of the plant does not drop
below freezing point, even though a layer of ice is steadily being formed and the
air temperature drops several degrees below freezing.
To be effective, sprinkling has to start early enough. A safe practice is to
commence when the wet-bulb temperature approaches freezing point and to
continue as long as the critical temperatures for plant tissues prevail - in practice
until the air temperature has risen well above zero, and ice starts to fall from the
plants.
Minimum sprinkling rates to give basic protection are 1.5 to 2.0 mm h-1 for
low-growing plants; 2.0 mm h- 1 for fruit trees, and 2.0 to 2.5 mm h- 1 for vines.

96
CHAPTER 6 - AGROMETEOROLOGICAL MANAGEMENT AT MICROSCALE AND TOPOSCALE

Under conditions of low relative humidity, these rates must be increased to

provide for extra loss and cooling due to evaporation. On the other hand, the rate
of application of water needs to be kept to a minimum. The findings of Rogers
and Modlibowska (1954) for apple trees in full bloom in England illustrate the
degree of frost protection to the application rate, in an airflow of 1 m/s.
Rate of application (mm per hour) 2 4 6
Temperature (in °C) to which protection is -3.3°C -4.8°C -6°C
given (based on screen minimum reading)
As mentioned above, the sprinkling system must ensure a complete and
continuous cover of water, as a break or interruption ·of only a few minutes can
cause the loss of the crop. Testing of the sprinkler system beforehand is essential.
To illustrate these points, Davis (1955) reported that small sprinklers turning at
the rate of one revolution per 12 to 20 seconds gave satisfactory results, while
larger sprinklers turning one revolution per 90 seconds or more did not adequately
protect a tomato crop at -5°C. Higher application frequencies result in less
temperature variations of the leaf surface.
On the economical side, the advantage of a sprinkler installation is that it
has many other uses at the farm, such as pest control and irrigation.

(e) Direct air and plant heating A classic method of preventing frost damage is by supplying radiation emitted
from suitable heaters or small fires. In this way the temperature in the orchard or
field may be maintained above the critical temperature for the particular crop, and
in many parts of the world heating is a regular practice in highly-valued orchards.
Diesel oil is the most widely used fuel for heating, followed by solid fuels, gas and
all kinds of waste.
The actual benefit to the plant is partly due to the increase in temperature of
the air, but also to a significant degree to the heat radiated by the heaters. Oil and
gas heaters heat the air primarily by convection but, depending on the type of
heater, some 10-30 per cent of the heat conveyed into the orchard may be radi-
ated by the hot surfaces of the burners (Rosenberg, et al., 1983).
Heaters should be distributed fairly uniformly over the area to be protected,
but it is advisable to have a larger number along the upwind edge and in the
colder parts of the area to be heated. Typically, heaters should be placed at inter-
vals of 5 m. A large number of small fires is more effective for heating the air than
a few large fires. Large fires may cause columns of rapidly rising air which will
break through the inversion ceiling without spreading and providing the neces-
sary circulation for the proper mixing of the warmer and colder air. Large fires may
not only disturb the ceiling layer, but also by drawing in cold air rapidly they do
more harm than good. Heating appears effective only if the inversion layer is
shallow, less than 10 m high, and the frost is not very heavy (Crawford, 1964).
A typical figure for fuel consumption is 300 litres of diesel oil per hour per
hectare (see WMO-TN No. 157). If the frequency of damaging frosts is higher than
once or twice per season, frost fighting by heating is usually uneconomical.

(fl Summary of active frost-fighting A strategic approach to combat the danger of frost should start with a study of the
methods local topoclimate. If heavy frosts occur regularly in seasons when the crop is at a
sensitive stage of its growth, only sprinkling and thermal insulation may be
useful. For climates with only weak frosts, the other methods can also be cost-
effective enough to be considered.

6.6.3 Short-term frost Frost forecasting is the only type of agrometeorological forecast which is available
forecasting in nearly all countries. The simplest forecasting method, over a century old, is to
forecast evening observations of wet-bulb temperature as expected minimum
References: WMO-TN No. 157;
temperatures, often with a locally-determined constant adjustment of a few
Ventskevich (1961).
degrees (Kammerman's rule) and some synoptic corrections. There are many vari-
ations of this approach, such as frost warning when the dew point falls below 2°C.
Often the value of other relevant parameters (cloudiness, or length of night) is
invoked by way of regression analysis.

97
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

A second approach is to compute radiational cooling of the surface, usually

by way of Brunt's formula for long-wave irradiation (published in 1932 - there
are operationally more reliable estimation formulae now available, see section
2.2.1). Early models forgot about advection, but that has been remedied
(Knighting, 1950; Georg, 1970).
However, very little verification has been done on such forecasts of surface
minimum temperature. One problem is lack of verification data, since the repre-
sentativity of, for example, grass-minimum temperature is questionable (section
2.2.2). The usual forecasting approach is to estimate expected minimum tempera-
ture at shelter height over some standard surface, then making subjective
adjustments to surface or orchard conditions, and then issuing the forecast.

6.7 ARTIFICIAL A glasshouse has two complementary functions - it is a form of weather-proof-

CLIMATE IN ing, preventing precipitation from reaching plants and cutting out the external
GLASSHOUSES AND wind. At the same time, it aims to create higher ambient and soil temperatures,
STABLES reduce risks of low temperature damage and provide facilities for control of soil
moisture and air humidity. Equipment can be installed to supplement natural illu-
References: WMO-TN Nos.
mination, affecting the intensity and duration of available light, or equipment to
122, 131; Geiger (section 20);
modify humidity and carbon dioxide content.
Keane (1986); Oke (1987).
Ordinary glass transmits short-wave solar radiation, but absorbs the long-wave
(heat) radiation. These facts underlie the 0ften-quoted 'greenhouse effect', in which
the glass is regarded as a one-way road for radiation. However, Wood and Van Gulik
(see Businger, 1963) found that it made no great difference in daytime or night-time
glasshouse temperatures, whether the cover was made of glass or of (IR-radiation-
transparent) rocksalt. Therefore reduced heat loss and resulting maintenance of
temperature in the glasshouse must be considered to be caused only partly by oper-
ation of the traditional 'greenhouse ·effect'.
The natural climate of a glasshouse is comparable to a desert climate. There
is no precipitation, so temperature in an empty 'Dutch light' glasshouse
(Figure 6.10) can attain 80°C through lack of evaporative cooling (Businger, 1963),

Figure 6.10
Approximate dimensions of a
(movable) 'Dutch light'
glasshouse unit .

. 08m

t I
I I

~:◄
--------3m------~►-:
' ·1

the same level of heat as in solar boilers! The reason for this is not the greenhouse
effect, but rather the lack of convective heat loss; a volume of air is confined in a
more or less airtight box, which is heated and in which heat is transferred to the
ground and the structure. Heat loss from a glasshouse is partly due to inevitable
leakages through any structure, even if well constructed, as well as to losses by
convection and by radiation from the outer surface of the glass.
Radiation transmission into a glasshouse depends on several factors:
(a) Influence of window material: spectrally, glass transmits very little radiation
of wavelengths A.>2.5 µm. On the other hand, polystyrene transmits about
40 per cent of the IR-waveband from 2.5 µm to 15 µm and thin polyethyl-
ene about 70 per cent (Hanson 1963);
(b) . Structure and orientation of glasshouse construction: short-wave transmis-
sivity of glass varies from :::.8 7 per cent at near-perpendicular angles to only
35 per cent at glancing angles of about 10°. Taking this into account, the
glasshouse orientation (N-S? E-W?) and its roof angle (often 27°) must be
chosen so as to profit most from the local radiation climate - the annual
variation of the Sun elevation and the ratio of direct and diffuse radiation,

98
CHAPTER 6 - AGROMETEOROLOGICAL MANAGEMENT AT MICROSCALE AND TOPOSCALE

which depends on factors such as local cloudiness and turbidity (section

2.1.1). The overall transmission of short-wave radiation into the glasshouse
is typically 70 per cent± 15 per cent if the windows are clean. Roof beams, and
screens used for radiation management, reduce the transmission, and
condensation on windows can block radiation;
(c) The long-wave radiation budget will be small under glass because it also
blocks incoming IR radiation. Under polythene, the situation is slightly
different (Silva and Rosa, 1985).
Heat is conducted through glass and is dissipated from the outer surface,
partly by radiation and partly by convection. Compared with metals, glass is a
poor conductor, but it conducts better than many insulating and building
materials. Some relative magnitudes of thermal conductivity are given below
(in W m-1 K-1):
Copper= 300; iron= 70; glass= 0.71; insulating board= 0.19.
The fact that conduction through the glass is not negligible is shown by the
finding, that inside a glasshouse night temperatures can be maintained at a higher
level if mats, designed to provide shade during the day, are rolled out at night.
King (1970) reported an average reduction of some 12 per cent in heating require-
ments through use of mats.
Solar energy is absorbed by the surface layers, by the structure and its
contents. In the absence of convective exchange, heat is retained and the amount
stored increases above that which occurs in the open. Temperature differences will
depend on the size of the glasshouse. Seemann (1973) gives for a spring month the
following averages:
Open air 290m3 40m3
glasshouse glasshouse
Month-averaged T 16.5°C 1s.1°c 23.6°C
Average daily T-variation 8.3°C 14.0°C 20.0°c

In summer the soil is heated relatively evenly throughout the greenhouse.

In winter, when external temperatures are low, there may be a horizontal
fall-off in the temperature of the upper soil layer from the middle to the edge of
the greenhouse, of the order of 1 °C m- 1 depending on the size of the
greenhouse.
The excess of inside temperature compared with that outside during the day
would be expected to be greater on sunny days than on overcast days. Seemann
(1973) found in a glasshouse with a 75 m 2 ground area, that on two successive
cloudy days in April the temperatures ranged from 11° to 14°C inside, compared
with 3° to 6°C outside, while on a sunny day, corresponding figures were 15° to
23°C (inside) and 6° to 10°c (outside).
Soil temperatures at 0.1 m below the bare soil surface, measured a few hours
after sunrise, also benefit. Gloyne (1954) observed that the gain in the critical early
season months of March and April in England was about 2°C in a dull month, 2.7
to 3.3°C when normal sunshine amounts were experienced, and 4°C or more in
years with sunny months. If such results are interpreted in terms of their effect on
the timing of commencement of the growing season (defined by a 5.6°C thresh-
old), a gain of three to four weeks is indicated. Such temperature gains may be
achieved in early season during germination and emergence of annual crops when
the soil is effectively bare.
The earlier growth stimulated under glass will, however, be exposed to spring
frosts, and the performance of protective devices against frost is of great practical
importance. Additional trials with movable covers involved the additional treat-
ment of sprinkling the outside of the glass with water whenever outside
temperature fell below 1°C. A series of observations over the December-March
period showed that the:
(i) Average gain in night minimum surface temperature on radiation nights was
5°C, and occasionally exceeded 7°C;

99
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

(ii) The average extra gain in temperature arising from water sprinkling was
2.5°C and on the coldest nights approached twice this value;
• (iii) Despite the fact that, in the open, frost prevailed up to mid-afternoon, on no
occasion did a temperature below freezing point occur inside the sprinkled
frame;
(iv) Some incoming solar radiation on the following day was required to melt the
ice which formed on the outer glass surface; nevertheless, the maximum
temperatures inside the frames were not effectively lowered, unless the
external temperature exceeded about 15°C.
Plant leaf temperature generally runs parallel to and rather above air temper-
ature, although with low irradiation and high levels of transpiration the leaf
temperature may fall below environment (i.e. the leaf is acting as a wet bulb).
Seemann (1973) quotes the following example for 'bright weather': "By day, leaf
temperature was from 5 to 7°C above air temperature inside the glasshouse; a
comparable excess did occur outside, but was confined to a few hours in the early
afternoon. Leaf temperature inside began to fall below air temperature by about
sunset and continued to remain below, or at the same level as, its environment
until some hours after sunrise. Outside, the temperature decrease started in mid-
afternoon, and leaf temperature remained S°C or more below the environment
during the night. In overcast weather, temperature differences at night were small
(say up to 2°C)."

For absolute humidity, fluctuations in non-ventilated greenhouses are closely

related to soil temperature. Humidity rises if the soil warms up (increasing evapo-
ration) and falls if the soil cools down. The absolute humidity is reduced by
increasing the air exchange (ventilation), even though soil temperature is kept
constant and evaporation is increased. When the temperature is held constant, an
equilibrium develops between the water vapour supply in the greenhouse and the
absolute humidity in the open air (due to ventilation). At this point, the absolute
humidity and the relative humidity in the greenhouse both remain constant.
Dry air (with low relative humidity) can therefore be obtained in the green-
house either by increasing the air temperature or by ventilation. Moist air can be
obtained by increasing evaporation from the soil in a closed greenhouse with rela-
tively low air temperature. This usually requires radiant or soil heating systems
provided, of course, that the soil is sufficiently moist. Where air heating is used,
however, the air is usually drier than is the case with other heating systems since,
under these circumstances, the soil usually remains somewhat colder.
The evapotranspiration process under glass is dominated by the saturation
deficit of the air, since air movement is restricted and therefore plays a relatively
·small part. The deficit is a function of the temperature of the plants and surfaces
and this, in turn, is governed by the level of irradiation. With increasing intensity
of solar radiation there is a rise in ·the transpiration rate, and consequently in
consumption of water by the plant. During sunny weather the saturation deficit
is further increased if the greenhouse is ventilated, so that dry external air lowers
the vapour content in its atmosphere. The evapotranspiration of the moist soil
. surface increases, but the transpiration from the plants cannot rise to an unlim-
ited degree. When the plant water balance reaches a specific level, the plant's
stomata close, limiting the transport of water.
Transpiration and irradiation are closely associated with periods with high
solar irradiation. If both are expressed in the same units, they are closely similar
in absolute terms, i.e. transpiration utilizes all the radiation. At lower intensities
the two move in parallel, but transpiration only accounts for a fraction of the
radiant input.

Livestock stables In addition to the climate in glasshouses, it is instructive to briefly consider

another important agricultural artificial climate, namely the environment in live-
stock stables. In a glasshouse (or its less permanent counterpart, the plastic
tunnel) we try to create artificial climates which are most favourable to crops. In
stables, we try to make the animals comfortable. There are some differences:

100
CHAPTER 6 - AGROMETEOROLOGICAL MANAGEMENT AT MICROSCALE AND TOPOSCALE

(a) In glasshouses, particular attention needs to be paid to radiative energy

management as plants require light for photosynthesis and animals do not.
In stables, light is mainly needed to ease mobility of animals or of their
caretakers (with the exception of fooling chickens into producing eggs by
appropriate light management);
(b) Animals are continuously adjusting their sensible and latent heat balances,
and trying to avoid thermal and moisture stresses. In the field, animals
actively seek out places where they feel more comfortable (their 'preferen-
dum', for example shade). They also adjust their behaviour in response to
climatic conditions in heat by extending their limbs and sweating or panting,
and in the cold by curling up, increasing thickness of their coat and shiver-
ing. The artificial stable climate must be adjusted to the animal's size, its age,
and the time of the day, since heat requirements are larger in the evening
(Verstegen, et al., 1987). In conclusion, it is important to take into account
climatic factors such as heat and humidity in the construction of a stable.
Ventilation is important for maintaining comfort in several ways. In humid hot
weather, thermoregulation by sweating becomes difficult. Often this effect is esti-
mated by defining an effective temperature Teff = 0.4 (T + Tw) + 10°c, and stating
that comfort is impaired when Teff approaches internal body temperature.
Ventilation can then aid heat loss. In stables, during windless and humid anticy-
clonic weather, the problem of getting rid of excess heat and moisture created by
animal breath is increased, but in such weather temperature falls rapidly at night.
Intermittent cold draughts imposed by air sensors inside the stable are bad for
animal health. In cold weather, extra heat loss is undesirable and strong wind
leads to wind chill - unfortunately, this is not well defined for humans, let alone
for animals (Kessler, 1993).

Carbon dioxide: CO 2 is needed for photosynthesis, and efforts must be made to ensure that the
supply is adequate to meet the demands of the plant. In an open field, the ground
level atmosphere contains about 320 ppm of CO 2 (the concentration unit
parts per million, for admixtures with molecular weight M, corresponds to
12.2 (M/T) mg m-3).
Typically, CO 2 concentration exhibits a diurnal variation. The Primary
sources of CO 2 in an open field are from:
(a) Respiration (minus assimilation) from plants and soil, in daytime and night-time;
(b) Combustion of sea and land organisms and combustion of fossil fuels;
(c) Secondary sources including volcanoes and mineral springs.
Because of the strong dependence on plant and soil respiration, CO2 content
varies according to the surface (land or sea, vegetated or bare soil) and with the
season and time of day. The seasonal variation of ground level concentration
(north of 30°N) varies from 332 ppm in September to 311 ppm in April. With
respect to diurnal variation, Geiger (section 17) gives the example of a potato field
where the variation went from about 400 ppm at night to 320 ppm by day.
Rosenberg, et al. (1983) report a diurnal variation over a sugar beet field from a
minimum level of 280 ppm by day (270 ppm on individual days) to a maximum
of 330 ppm at night (over 360 ppm on some occasions).
In the enhanced growing conditions created in the glasshouse (especially if
additional radiant energy of appropriate wavelength is supplied by lamps), CO 2
demand of the plants increases. However, being isolated from an external source,
CO 2 concentration within the glasshouse may fall - Seemann quotes a fall to as
low as 200 ppm. Thus it becomes desirable to artificially inject CO 2 into a
glasshouse to meet the potential need of the plants. Practically speaking, a concen-
tration of CO 2 exceeding about 600 ppm cannot be maintained for any length of
time under normal conditions of air-tightness of the structure and ventilation
even when the glasshouse is surrounded by similarly CO 2-emitting glasshouses.
It should be emphasized that most of the data quoted above are from meas-
urements obtained in humid temperate regions. Glass is increasingly being
employed in lower latitudes, for example in Southern Europe and the
Mediterranean regions. The physical processes already described remain valid, but

101
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

the magnitude of the various effects may be substantially different and the asso-
ciated management problems very different. Excessively high temperatures, too
much solar irradiation during parts of the year and low internal humidities are the
likely hazards rather than low temperatures, insufficient light and high humidity.

It is sometimes assumed that users of glasshouses need no meteorological infor-

mation, as they make their own climate. This is not entirely true. For one, the heat
capacity of a glasshouse is limited, and heating or ventilation adjustments have
typical time lags of an hour. This means that glasshouses are very sensitive to
sudden radiation increase - in summer unventilated temperature can exceed
70°C. Another major requirement is information on existing and expected humid-
ity levels, in order that adjustments to the climate management can be made in a
timely manner. When such management is based on local observations without
any forecast information, the risk exists of severe maladjustment.
A good example is a bright cold night, followed by a sunny morning. In
order to keep the crop in a healthy state, the temperature of the glasshouse has to
be turned down. The glancing Sun of the first morning hours will heat mainly the
topmost crop leaves, causing vigorous evaporation up there, while the lower
leaves still are cold. Water vapour will diffuse down into the crop and condense
there, causing fungus problems. The farmer can prevent this risk by heating the
glasshouse a few hours before sunrise, followed by extra ventilation when the
outside air is sufficiently dry. Automatic regulation can only do the same if
provided with a forecast for a sunny morning.

Review questions on Chapter 6

A. Questions on basic understanding:

6.1 Compare and contrast the effects on soil temperature and soil moisture
and the practical implications of:
(i) A cover crop;
(ii) A mulch of organic material;
(iii) A mulch of some artificial materials (such as plastics, etc.).
Consider these processes in different climatic regimes, for example cool
temperate, humid; hot, arid, etc.
6.2 Geiger (section 19) cites a case in which at dawn in August (Northern
Hemisphere), the surface temperature of a freshly ploughed field was
2°c colder and, at a depth of 10 cm, 1°C colder than in an untilled field.
At 1500 hours, the surface temperature of the ploughed field was 5.5°C
higher. Explain.
·6.3 At the equinox (21 March, 21 September), the Sun rises in the east and
sets in the west respectively at 0600 and 1800 hours local time. Between
the spring and autumn equinoxes it reaches these cardinal points respec-
tively later and earlier in the day. Use these facts to explain the form of
the dashed line (- -.), i.e. fore= in Figure 6.4.
00

6.4 Evaporation in sheltered zones has been measured by a variety of

methods, for example by small pan evaporimeters, by Piche evaporime-
ters, and by measurement of soil moisture. Discuss the factors which can
explain any differences in the patterns of water loss so revealed in the
sheltered zone. Discuss also the need for constructing buffer zones or
guard-rings around instruments designed to measure potential evapo-
transpiration.
6.5 Why, and under what circumstances, could a thin layer of snow be a
greater hazard to young seedlings than no snow at all?
6.6 A gardener is reported to have attempted to protect young plants against
frost damage by covering them with upturned metal cans only a little
larger than the plants. By doing so "he condemned his 'protected' plants
to death by frost" (Geiger, section 54). Explain.
6. 7 Occurrences of frost are usually categorized as 'advective' frosts or 'radi-
ation' frosts. Describe the characteristics of each and the physical

102
CHAPTER 6 - AGROMETEOROLOGICAL MANAGEMENT AT MICROSCALE AND TOPOSCALE

processes which give rise to the phenomena. Why do radiation frosts

(ground frosts) occur more often than advective frosts (air frosts)?
6.8 State briefly the physical reasons which render the following practices
potentially effective anti-frost measures, and in what circumstances are
such measures likely to be most or least successful:
(i) Flooding of the ground;
(ii) Consolidation of top-soil;
(iii) Erection of up-slope wind breaks;
(iv) Smudging by smokes and vapours;
(v) Using wind machines.
6.9 What do you understand by the latent heat of freezing? How much heat
is involved when a gram of water freezes? Briefly, how are these
processes utilized in one method of fighting frost? What can some disad-
vantages of the method be?
6.10 In some circumstances, a wall or similar barrier will increase the risk of
frost damage, in other circumstances it will reduce the risk. Discuss this
apparent contradiction.
6.11 Rosenberg, et al. (1983) state that 'the reduction in turbulence demon-
strated in these studies provides a better explanation of the unique
microclimate and plant growth in shelter ... than does the reduction in
wind speed'. Discuss the issues raised in this statement.
6.12 What do you understand by the 'greenhouse effect' which is often
mentioned when discussing long-wave balance? To what extent is this
an adequate explanation of the heat economy of a real glasshouse?

B. Questions on extended application:

6.13 Geiger (section 20) cites an example of a case where the blackening of an
exposed surface did not give rise to a greater absorption of radiation by the
ground. On two adjacent rice fields the surface of the shallow water cover
of one was covered with black powder, the other remaining clear. Midday
global radiation was 942 W m-2, utilized as follows (all units W m-2):
Blackened surface Clear water surface
Evaporation 548 486
Heating air, and long-wave radiation 229 200
Heating water 94 66
the balance heating the underlying soil surface. Discuss this, using the data.
6.14 Consider a uniform fence of height L, running E-W. Show that:
(1) The depth of the shadow cast is L cot h cos A, where h = solar alti-
tude and A = azimuth;
(2) At, say, <I> = 32° show that at the equinox, when the solar declina-
tion a = 0° (i.e. at latitude a = 0° the noontime Sun is directly
overhead), the depth of the noontime shadow is the same as at any
other time of the day, e.g. 9 hr.
Prove that, for all practical purposes (for a = 0) the shadow depth from
an E-W barrier is constant throughout the day.
6.15 If the sky is partially obscured - as would be the case from the bottom
of a depression - radiation flux is only to a part of the sky. If the rim of
the depression is circular and subtends an angle z from the zenith, then:
when 10 b = net outward long-wave flux in depression;
10 = net outward long-wave flux from a horizontal plane (i.e. an
unobstructed site)
then l0 b/I0 = (1 - cosY+2 z).
Confirm the following values of the ratio:
y 0.3 0.7
z
30° 0.283 0.332
60° 0.797 0.857

103
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

Note: The screening produced by a circular wall subtending 30° to the vertical is
substantial, but that with a lower wall subtending only 30° to the horizontal the
flux is about 80 per cent of that from an unobstructed site.
6.16 The following data are quoted from Geiger (section 19):
Monthly mean temperatures and humidities 5 cm above an uncultivated
bog (B) and a reclaimed meadow (M). The season was high summer and
the grass in the meadow grew vigorously. Describe and explain the
Table 6.7
processes which result in the differences listed in Table 6. 7.
Month Air temperature Air humidity
Maximum (°C) Minimum (°C} Diurnal (°C) Relative Vapour
variation (%) pressure (hPa)
B M B M B M B M B M
- May 23.6 21.4 2.9 3.4 13.2 12.4 65 79 11.2 13.7
1 June 22.4 20.8 5.6 6.2 14.0 13.5 63 77 17.1 18.9
July 23-.5 20.5 8.0 8.2 15.8 14.4 71 84 19.7 20.4

6.17 · One approach to the study of the effect of weather on farm livestock
may be illustrated by an empirical study of the relative strain (RS) on a
man in a light business suit walking at 0.9 m s-1 in a wind of 0.5 m s-1
in hot weather:
8.0 + 0.56 (Ta - 35)
RSS = ------=----
3'3 - Pa
0
where Ta= air temperature ( C) and Pa= vapo~r pressure in hPa.
(i) Construct a psychrometric diagram with Ta on the horizontal axis
and Pa vertically, with isopleths relating to various levels of relative
humidity;
(ii) Superimpose on this diagram isopleths of RS5 (the isopleths will be found
to run downwards from left to-right). See WM0-1NNo. 123, p. 7.
6.18 Heat loss from ventilated bodies maintaining a constant temperature
(specifically humans and large mammals with a deep-body or 'core'
temperature of 36.5°C), when v refers to speed of air passing over the
body and T and Tw are the dry-bulb and wet-bulb temperatures, was
traditionally expressed by the formulas:
• Dry surface: H = (a + b v0-5) (36.5 - T) meal cm-2 s-1
where coefficients a and b vary with wind speed; typical values are:
for v<l m s-1 : a= 0.105 b = 0.485
for v~l m s-1: a= 0.205 b = 0.385
• Wet surface: Hw = (0.37 + 0.51 v0.63) (36.5 - Tw) meal cm-2 s-1
Rewrite the old formulae in S.I. units (1 meal cm-2 s-1 = 41.87 W m-2).
Investigate the values of H and Hw for a range of conditions existing in
your area.
Note: Even the surface temperature of exposed surfaces (e.g. face) will be less than
deep-body temperature. With clothed humans and animals with a hair-coat, fur-
coat or fleece, the temperature of the outermost surface could clearly be much less
0
than J6.S C. For further information see WMO-TN No. 123 (1972), Kessler
(1993), and especially Keane (1986).

6.19 The reduction in wind speed due to shelterbelts has been found to
reduce the heat loss from houses in the sheltered area.
For a particular structure, Wang (1967, quoting Woodruff) expressed the
heating load on a sheltered house as:
Q = 0.69 L\T (IQ0.04 u tt0.07) Watt
L\T = difference between inside and outside temperatures (0 C)
U = wind speed in m s-1
H = number of multiples of shelterbelt height of house from the barrier

104
CHAPTER 6 - AGROMETEOROLOGICAL MANAGEMENT AT MICROSCALE AND TOPOSCALE

Check the following values for H = 6:

U 4.5 8.9 13.4 m s-1
Qjl:!T 1.4 2.2 3.5 Watt
Work out values of Q}l:!T for other combinations of U and a range of
values of H. Plot log Qjl:!T against U.
6.20 The data in Table 6.8 were obtained by Kreutz (see WMO-TN No. 59) for:
u = windspeed (as a percentage of the unobstructed or free wind);
e = evaporation (as a percentage of water loss in the open field)
in the lee of a wind break of H = 2.5 m.
Table 6.8
Upwind of Above Distance to lee of barrier (in m)
barrier barrier 1 4 7 10 13 16 19 22
u 100 122 52 59 63 61 74 80 82 91
e 100 115 70 80 84 80 86 95

Plot u and e against distance, and plot ✓u and e against distance, and
confirm that ✓u and e agree more closely than do u and e. The depend-
ence of evaporation on ✓u is frequently mentioned in shelterbelt studies.
6.21 The data in Table 6.9 (below) are quoted in WMO-TN No. 59 and relate
to the water loss (in grams per 100 minutes) from trays of moist soil on
Table 6.9 either side of a barrier erected in a wind tunnel.

Distance downwind of barrier (multiples of H)

-10 -6 -2 +2 +6 +10 +14 +18 +22 +26 +30 +34 +38
Water loss:
Without barrier 6.3 5.8 5.9 5.7 5.6 5.6 5.5 5.4 5.2 5.2 5.2 5.1 5.2
Dense wall 6.2 6.2 6.5 5.5 6.8 8.1 8.2 7.7 7.0 6.7 6.6 6.3 6.3
Permeable wall 6.0 5.9 5.5 5.0 3.5 3.0 3.7 4.7 5.1 5.9 5.3 5.1 5.3

Establish a graph and discuss the results. Note the contradiction between
the results for the dense and the permeable barrier. The higher water loss
with the dense barrier is attributed to the undoubted higher turbulence
generated by the wall. This situation contrasts with Rosenberg's general-
ization mentioned in 6.11.
Note: Investigations in smooth wind tunnels show a more extended protected
area than occurs in practice. The discrepancy is progressively reduced as experi-
mental arrangements are made more realistic by roughening up the wind tunnel
floor aensen, 1954).
A very rough rule of thumb to apply to much of the earlier laboratory work is to
reduce distances by SO per cent, i.e. 26H (laboratory) ➔ l3H (in the field) .

C. Questions involving practical work or local data:

6.22 The vegetative season (and certainly its commencement when germina-
tion and emergence are of major importance) can be said to be linked
more closely to soil temperature than to air temperature.
Using published or other data for your area, examine the difference
between a vegetative season (both its commencement date, date of
termination and duration) defined by reference to air temperature, and
by reference to soil temperature. Take into account the points
mentioned in section 3.4.
6.23 (a) Assemble data from your area on the dates of first and last frosts for a
series of years. Adopt the following criteria:
• Minimum in the screen of 0°C
• Minimum in the screen of -2°C
(b) Using data for at least 20 years, compute the mean dates of the first and
last frosts;
(c) Assuming that the frost-free season is defined as the length of period
between the average date of the last frost in spring, and the first frost in

105
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

autumn, examine the extent to which the frost-free season so defined

could be regarded as useful for making decisions on policy.
6.24 Ventskevich (1961) states "in western parts of the [former] Union of
Soviet Socialist Republics (USSR), frosts cease as soon as mean daily
temperature .exceeds S°C or very rarely 10°c ...... " Examine data from
your area/region to ascertain how satisfactorily it is possible to deduce
the date of the last frost of the frost-season from the course of mean
daily temperature.
Define· the course of mean daily temperature:
(i) On the basis of individual days and S°C threshold;
(ii) On the basis of three-day running means and S°C threshold,
attributing the mean to the middle day of a group of three, and also
to the last day of a group.

106
CHAPTER 7
WEATHER HAZARDS ADVERSELY AFFECTING AGRICULTURAL
OUTPUT

The occurrence of meteorological disasters places meteorologists under public

scrutiny. They are alternately criticized for not being able to prevent disasters, or
asked to comment on proposed schemes to prepare for them, some of which will
be more appealing than sensible.
The more common meteorological phenomena, such as gales, hurricanes
and thunderstorms, are dealt with at length in other compendia; they do not need
to be discussed here. Instead, attention is given to some other hazards with partic-
ular importance for agriculture. The purpose is mainly to provide an acquaintance
with the most important aspects of these weather hazards. More extended cover-
age is given to such hazards in various WMO Technical Notes and other books (see
Appendix IV).

7.1 DROUGHT The term 'drought', generally understood to signify a period of dryness due to lack
of rain, is variously interpreted according to the experience of individuals,
References: WMO-TN
communities or nations; and according to the needs of a particular branch of
Nos. 138, 184, 186.
industry or activity for which a supply of water is an essential component. The
difference in views is illustrated in WMO-TN No. 138, in which some definitions
are given:
"To the meteorologist drought is a rainless situation for an extended period
during which some precipitation should normally have been received depending
on location and season. The agriculturist considers drought as a shortage of mois-
ture for his crop. The hydrologist views drought as being responsible for
depression of surface and underground water levels or diminution of stream flow.
To the economist drought means a water shortage adversely affecting the estab-
lished economy of the region."
Another reason why it is virtually impossible to arrive at a single, simple set
of criteria for defining 'drought' is stated as follows (WMO-TN No. 138):
"Many definitions of drought are based on the adaptability of husbandry to
the 'average' conditions and the importance of this is stressed in this report. For
example, a pastoralist raising fat lambs on improved pastures with a uniformly
distributed rainfall averaging, say, 1 000 mm a year, might be troubled by the rela-
tive 'dryness' in a year producing only 750 mm, irrespective of its temporal
distribution. To another pastoralist in semi-arid marginal agricultural country,
normally receiving 300 mm a year, this total of 750 mm would represent a record
wet year, bringing with it the troubles associated with excessive moisture, namely
water-logged soils and pastures, lodged crops, untrafficable roads and ploughing
or harvesting difficulties. The agriculturalist, especially in the drier areas, has
assessed the nature of local rainfall and, through years of long and sometimes
bitter experience, has learnt to adapt his operations to rainfall characteristics of
the area. This is particularly the case in those areas with a long history of land use,
in some instances extending back beyond historic times."
This section gives only a brief overview of some agricultural issues and ques-
tions which are important during the occurrence of drought. This may provide
some insight into the extensiveness of drought problems and prevent some
aspects getting exclusive attention while others, maybe equally important, are
ignored. The question of water for irrigation, including an introduction to
methods of assessing the amount needed, has been considered in Chapters 4
and 6.
Water shortage is a basic component of drought. It is a relative rather than
an absolute condition. When a stable agricultural system is found to exist, e.g. a
particular crop rotation or a pasture management practice, it will have developed

107
WMO LECTURE NOTES - AGRICULTIJRAL METEOROLOGY

in such a way as to be consistent with the regular pattern of rainfall. This implies
that insufficient or excessive rainfall will not result in a higher failure frequency
than can be absorbed by the system.
In applied statistical analysis of rainfall, it has been shown that the relevant
statistical 'average' for social absorption of extreme situations is not the familiar
arithmetic mean, but the median value, i.e. that which is exceeded on 50 per cent
of occasions, or equally, not reached in 50 per cent of occasions. When a distri-
bution is negatively skewed, as is normally the case with rainfall, the median is
less than the arithmetic mean. Failures are linked with rainfall amounts signifi-
cantly below median, rather than below mean rainfall.
In agricultural drought, the principal complicating factor (which also arises
in the context of hydrology) is the need to take account of the soil moisture
reserves at different depths, where critical levels necessarily vary with type, growth
and stage of development of individual crops. In the context of agriculture,
'drought does not begin when rain ceases, but only when the plant roots can no
longer obtain soil moisture', and it may be defined on the basis of soil moisture
rather than upon some indirect interpretation of rainfall records. Since available
soil moisture depends upon the character of the soil and its effects depend on the
requirements of the crop, agricultural drought exists when soil moisture in the
root zone is at a level near the wilting point. Direct soil moisture measurement
and/or the techniques of water budgeting, based upon basic climatic data and
specific assumptions as to the extraction of water by plant roots, are necessary for
any realistic evaluation of drought risk. They are also needed as useful guidance
for warning systems and for taking effective counter-action.
In regions which experience well-defined wet and dry seasons, agricultural
practice will have developed in accordance with the seasonal march of soil mois-
ture and soil temperature. However, in areas where drought is only occasional, or
variable in time of occurrence, it can disrupt planned sowing dates. Even in the
relatively humid climate of the British Isles, inadequate spring rainfall can lead to
delays in sowing. In semi-arid tropical areas, the scheduling of sowing dates in
conformity with soil moisture (and soil temperature) is common practice.
Germination is only the first phenological phase of development. It is gener-
ally valid to state that the distribution of rain over the wet season is at least as
important as the total amount of rain received. In Senegal, it was found over the
period 1932-1964 (Popov, 1990) that there is no correlation between total seasonal
rainfall and final crop yield, but that crop yield is well correlated with a locally
derived index I, that weighs the rainfall according to its timing (Figure 7.1).
Similar examples, given below in Chapter 8, show that a Koppen-style climato-
logical precipitation average is insufficient information, while special
·agroclimatological summaries can be very helpful to predict yields.
The effects of drought on plant communities are various. Severe water short-
age brings about a real change in the natural vegetation. Drought-sensitive species
die out; annuals, which do not reach their reproductive stage, are very likely to

Figure 7.1 y Comparison yield data and seasonal

y Comparison yield data and index I
Forecast power: groundnut kg/ha rainfall for 32 years (1932-1964) kg/ha for 32 years (1932-1964)
Groundnuts Groundnuts
crop yields as related to season- 2000- 2000
average rainfall (not e 59
36
35
38
significant, just a structureless e58 43
64. 48 63
cloud of points) and to a well- 1500
devised water requirements 41

satisfaction index 34046 33

1000 56
(Popov, 1990).
e51
•
55
32
55

050
500- 500

I I I I I

400 600 800 1000 1200 70 75 80 85 90 95 100

P (mm) I

108
CHAPTER 7 -WEATHER HAZARDS ADVERSELY AFFECTING AGRICULTURAL OUTPUT

suffer, and perennials (especially grasses) - which in rainless periods survive at the
expense of sub-surface organs - may die off completely during extreme drought.
When drought-stricken areas are used for grazing, many complications arise.
Among these is the sheer reduction in the mass of vegetative material, and the
effect of serious defoliation by the grazing animals on the development of plants.
This depends on relationships between seasonal grazing patterns and seasonal
plant growth, and consequential effects upon the soil. All will be further influ-
enced by the intensity of grazing. Soil cover changes, in turn, have a direct
influence on the possibility of erosion by water and winds, since vegetation cover
is the best defence against erosion. In particular, soil erosion by wind is critically
affected by the moisture status of the exposed soil, as dry soil is easily blown away
(see section 6.4.2).
Drought also impedes the execution of agricultural field operations in two
ways:
(a) Drought influences the workability of soil, the feasibility of field operations
(e.g. clays compact into a near unbreakable cement-like structure - see
section 9.3.1);
(b) If rains are late, soil moisture may be below the level which permits good
sowing or planting of seedlings. After sowing, an occasional dry spell is less
harmful.

Use of water in agriculture The principal uses for water supplies in agriculture are:
(i) Drinking by livestock;
(ii) Crop spraying, irrigation;
(iii) Cooling of e.g. milk; buildings (air conditioning);
(iv) Washing of livestock, vegetables, buildings; also the dilution and trans-
port of waste products (excreta);
(v) Frost protection (using spray irrigation - see section 6.6.2);
(vi) Access to supply for fire-fighting.
Of these, (v) and (vi) are specialist uses and not perhaps of first importance
in the general problem of drought. Water for (iii) and (iv) is a feature of intensive
commercial agriculture, but (i) and (ii), the watering of livestock and crops, is as
basic to common subsistence farming as it is to specialized intensive systems.
A very great amount of water is needed by some livestock (notably dairy
cattle). This varies from region to region and with the moisture derived from
herbage which is consumed. Under favourable climatic conditions (e.g. as in the
British Isles) a dairy cow will consume about 140 litres a day, of which it drinks
about 50 litres.
In extensive livestock (range) farming the location of drinking holes or tanks
is a matter of vital importance. In the arid areas of Australia, stock will not graze
much outside a radius of about 15 kilometers from a supply of water. Since
drought affects pastures, an upper limit to stocking rates is set by the above-
mentioned condition. The more abundant the pasture, the greater the number of
stock which may find sufficient grazing within a circle of such radius.
A rather unexpected example of shortage of drinking water can occur in the
winter in many areas owing to the freezing of streams, ponds, drinking troughs,
etc. In the relatively infrequent severe winters experienced in the hilly areas of the
British Isles, the substantial loss of sheep which can occur is due at least as much
to a shortage of liquid water as to a lack of fodder.
Up to 70 litres per cow per day are required for hosing down houses and
yards and during droughts, when streamflow is low, the discharge of insufficiently
diluted effluent can pollute and poison streams and rivers.
The use of water for air-conditioning and cooling is mainly considered in
relation to urban areas; however, milk yields decline in hot weather and air-
conditioned barns are used in some parts of the world. In the absence of
adequate air-conditioning, Hahn and McQuigg (1970) estimated that shortfalls of
possible production per cow for a summer season ranged from less than 25 kg per
cow per season in the northern United States up to 340 kg per cow per season in
Southern Texas.

109
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

While excess precipitation leads to leaching (see section 6.5), a minimum

amount of water is needed if fertilizers spread on a large scale over the soil surface
are to penetrate to the roots. Dedicated local cultivation can, of course, result in
the material being applied locally near the root zone.
The effictiveness of application of some insecticides and pesticides (dusts and
sprays) depends upon a covering of moisture on the plant tissue. A water film is
needed to make dust particles adhere, to spread spray droplets and possibly to
dilute the droplets as well. For some dusts, the condensate from dew may suffice,
and this can be deposited during the night even in periods of drought.
An associated consequence of a lack of rainfall is the increasing concentra-
tion in the soil of undesirable residues from fertilizers, insecticides and pesticides.
Pests and disease agents are subject to multiple interactions with drought
conditions. As well as the environmental conditions which allow the pest or
disease organism to survive and grow, a successful attack by pests or disease
frequently requires an intermediate host plant or animal to be available in suffi-
cient quantity and condition at a particular time of year. Moreover, the health and
state of the animal or plant at risk may determine whether a disease attack is
serious or not. A film of moisture on vegetation is a necessary condition for the
spread of many plant diseases, especially those associated with fungi (see
section 8.3).
Integrated biological pest management calls for explicit knowledge of the
lifecycle of the pest or pathogen and of the sensitivity of each phase in the cycle
of weather conditions. The active collaboration of biologists and meteorologists is
required for this, as independently-working biologists tend to regard phenological
information as the only understandable input of 'weather' data.
The basic philosophy of drought preparedness should be to restore and
develop the natural resource base in a region to a sustainable level, in order to
minimize the vulnerability to drought. The following objectives seem essential for
this:
(a) Establishment or improvement of an early drought warning and monitoring
system;
(b) Compilation of a natural resources inventory, including improvement of
meteorological data dissemination, and climatological and hydrological
data-bases;
(c) Introduction of new technologies, such as remote sensing and Geographic
Information Systems (GIS) to enable meteorological forecasts to be more
area-specific;
(d) Restoration of the natural resource base by rehabilitating and reclaiming
degraded forest areas and community lands, and making available water
more usable by suppressing surface and ground water pollution;
(e) Improvement of information exchange and collaboration between govern-
ment departments, and between nations in the region.

7 .2 ARTIFICIAL Since time immemorial, mankind has tried to change the weather to their advan-
STIMULATION OF tage. A number of magical and religious procedures along with more
PRECIPITATION technological approaches have been tried. Many people do not seem to realize the
magnitude of atmospheric masses and energy exchanges involved and become
References: WMO TN Nos. 105,
genuinely angry when actions are said to be ineffective. This is particularly a
154; Braham (1986).
problem when governing authorities are involved. When they have promised
their subjects to take action about an aggravating problem, they do not like to be
told that the actions they undertook are useless. Meteorologists need a great deal
of tact when dealing with these matters, and sometimes a little bit of humour.
In disastrous situations people want to do something - be sparing with your
criticism!
The artificial stimulation of rainfall has drawn much public attention ever
since the first experiments to attempt it were conducted in the middle of the last
century. To understand the approach taken requires a basic understanding of the
precipitation formation process (see GM II, Chapter VI and Chapter XII, and
Mason, 1975). Water vapour in the air does not simply condense into droplets

110
CHAPTER 7 - WEATHER HAZARDS ADVERSELY AFFECTING AGRICULTURAL OUTPUT

whenever its vapour pressure is high enough to be in equilibrium with a

hypothetical plane water surface (commonly, and misleadingly called, 'saturation
level' - see Bohren, 1987). Because a single water molecule does not have the
same attraction to another single water molecule as does an entire plane water
surface, condensation of individual water molecules requires either
supersaturation of over 700 per cent, a temperature below -40°C, or the presence
of a suitable condensation nucleus, such as an aerosol particle with a diameter of
more than 0.5 µm. Atmospheric condensation, in brief, requires condensation
nuclei to form drops and freezing nuclei (with an ice-like surface structure) to form
ice crystals.
The basic assumption in rainfall stimulation is that precipitation does not
occur in a particular saturated situation because of a lack of appropriate nuclei. By
introducing such nuclei into non-precipitating clouds (called 'seeding the
clouds'), it should be possible to trick nature into making rain.
The assumption that seeding can help requires two synoptic conditions to be
met:
(1) The air must have reached saturation point, usually because it has risen
enough to be cooled sufficiently. The air can ascend due to large-scale
convergence, or to local-scale cumulus convection, or to orographic lifting.
(2) If rainfall were to start, it will not continue unless the depletion of water
from the cloud is compensated by advection of more water vapour - either
locally from a moist atmospheric boundary layer, or through larger-scale
advection processes.
An interesting general example where conditions failed to meet the second
condition occurred on 25 February 1989, when a deep depression was located in
the southern part of the North Sea, inducing ascending motions of 400 hPa in
24 hours. Nevertheless, calculated and observed precipitation was only 4 mm,
because the depression ascent occurred over such a large area that there was
almost no humidity convergence. Significant precipitation occurred only on the
edge of the low-pressure area, in Spain.

Assuming that non-precipitating clouds are present, any seeding operation must
be preceded by an estimate of the condensation nucleus situation inside the
clouds. We must first distinguish between two major alternatives:
(a) Temperatures in the cloud are above 0°, except maybe at the top ('warm
clouds');
(b) The upper part of the cloud is below freezing temperature ('cold clouds').
For the combination of too-small cloud droplets into larger precipitation
elements, the most effective process is the Bergeron process (GM II, Chapter XII):
the differential evaporation from supercooled droplets onto ice crystals. This
occurs only in cold clouds and requires ice nuclei. Artificially, these can be
provided either by seeding with solid CO 2 ('dry ice', T = -78°C) and converting
droplets into ice pellets, or by using crystals of lead iodide, silver iodide or anti-
mony. For every km of seeding track, we typically need 10 g of such salts or 1 kg
of dry ice.
In warm clouds, cloud droplet combination occurs only by the coalescence
process, requiring the presence of some relatively large droplets (over 20 µm).
These can be produced by seeding with large hygroscopic salt crystals (e.g. NaCl
and (NH 4 )zSO4 ), typically 25 kg per km of seeding track. The basic action mecha-
nism of such large-nucleus seeding, widening the droplet spectrum, would also
have a precipitation-stimulating effect in cold clouds.
Assuming effective means of delivering such seeding material into some
non-precipitating cloud, the chief question is: does it work? Any observed rain
might also have been produced naturally by the cloud, without any seeding. Proof
therefore requires large experiments, where some regions are seeded while other
similar control regions are left alone. This procedure should go on for some years
in order to account for the large natural variability of rainfall. Several projects of
this type have been carried out in Australia, Canada, Israel, the United States and
elsewhere.

111
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

Results have been contradictory, but the overall conclusion (Cotton and
Pielke, 1995) is that seeding requires much more preparation than one would
expect. There is a time window of opportunity which is determined by the
availability of enough supercooled water before depletion by entrainment.
Moreover, a rough knowledge of the actual nucleus and droplet spectrum inside
the cloud is essential, but a non-routine observation. If there are already too
many droplets inside the cloud, they will all be small, and no rain will ever reach
the ground. The natural presence of an excessive amount of (dust) nuclei can
even contribute to desertification (Bowen, 1967), an unfortunate form of
feedback when the desert is the. source of the dust.
Moreover, the local climate determines the available type of clouds. During
drought years, in the absence of cloud and of naturally occurring rain, cloud
seeding can have no effect: On the other hand, in warm-based maritime clouds
the natural probability of rainfall may be so great that there is little opportunity
or need to seed suitable clouds. For seeding to succeed, it is necessary to have a
type of cloud which has enough water content, but an ineffective droplet spec-
trum due to deficiency of a particular type of nuclei. If such nuclei can be
delivered in the cloud economically and at the right time, then artificial stimula-
tion of precipitation may work - but only in long-term climatological terms, with
possibilities of precipitation increases of the order of 25 per cent in the annual
average, which is not enough to eliminate aridity (WMO-TN No.105; Changnon,
1975).
In conclusion, it can be said that artificial stimulation of precipitation is
unlikely to be of use in alleviating temporary drought problems. It may be that in
certain semi-arid regions the regular climatological situation is such that, at the
beginning or the end of a too-short wet season, a short period of cloud-seeding
action might be viable and even economic. However, this would require careful
advance research, along lines set out at length in WMO-TN No. 154. If such action
seems warranted, then its average success or failure will only be known after many
years.

7.3 HAIL Hail is formally defined as an icy precipitation with a diameter of at least 5 mm.
Smaller icy balls are called graupel if spongy, or ice pellets if solid. In a thunder-
Reference: Mason (1975),
storm cloud, the formation of such smaller ice particles takes about 10 minutes in
Rakovec (1989).
a cloud region with sufficient supercooled water drops and a moderate updraft.
The small ice particles have fall-speeds of the order of 1 m/s, and if the updraft is
much stronger, the particles will be blown up and out into the fully glaciated
cloud anvil before they have grown much. This is one of the many reasons why
hail reaches the ground from only a small percentage of all thunderstorms.
·Melting of hail while falling is one of the others.
Only if the ice particles, while growing, remain in a cloud region where the
updraft speed nearly matches their (increasing) fall-speed, can the particles
continue to collect cloud droplets at an increasing rate. Some types of thunder-
storm, particularly the so-called multicell and supercell types, have sufficiently
large updrafts (> 20 m/s) to keep a hailstone suspended for a sufficient period of
time to allow it to grow to large size (Cotton and Pielke, 1995). This process takes
place in a cloud volume of limited horizontal extent of less than 1 km across.
Therefore hail usually falls in strips of several hundred meters width, directed
along the upper-air motion direction of the thunderstorm.

7 .3.1 Distribution of hail in For the world as a whole, hailstorms are responsible for roughly the same annual
space and time economic damage to crops as hurricanes and tornadoes taken together. The latter
do more damage, but are less frequent. In mid-latitude inland regions, hail is by
far the most destructive weather phenomenon for agriculture. For the entire
United States, WMO-TN No. 132 gives an annual estimate of US$ 500 million for
hurricane damage in coastal regions and more than US$ 300 million for hail
damage.
The geographic distribution of hailstorms is somewhat unexpected.
Conditions for strong vertical currents, which are a prerequisite for hailstorms, are

11.2
CHAPTER 7 - WEATHER HAZARDS ADVERSELY AFFECTING AGRICULTURAL OUTPUT

generally best developed in tropical or subtropical areas, but the high tempera-
tures in those areas make the formation of ice unlikely. Therefore, the largest
number of hailstorms takes place in latitudes between 30° and 60°. In polar
regions, where freezing temperatures are most frequent, intense vertical convec-
tion is missing. Nevertheless, some hailstorms are observed when
upper-atmosphere temperatures are very low.
The best conditions for hailstorm formation are over inland mountains and
plains, where intensive heating creates instability in the air. Wide stretches of bare
land are easily heated, and hailstorms are more frequently observed there than over
forested areas. The recurrence of hailstorms over a given area is striking. Time and
again the conditions seem to be favourable for hailstorms in specific regions, while
surrounding regions only experience ordinary thunderstorms.
In view of the fact that the occurrence of hail is a relatively localized
phenomenon, it has been reported irregularly. Consequently, there are very few
places in the world where a proper climatological analysis of the occurrence of hail
has been carried out.
The annual distribution of days with hail is not quite uniform in various
regions. The usual type shows a maximum in the local summer. In the monsoon
Table 7.1 regions, however, spring and autumn have most hailfalls, and the Pacific Coast of
Variation by month of in the the United States shows even a winter maximum. Some of the characteristics of
occurrence of hailstorms over a average annual variations and how the maximum frequency shifts, averaged over
IO-year period (as a percentage a ten-year period, are shown in Table 7.1 below. The table shows an early spring
of all storms L) maximum in Arkansas, a late spring maximum in Kansas and a later summer
(Landsberg, 1968; Benincasa, maximum in both Montana and Ferrara in Italy, both in or near mountainous
et al., 1991). regions.

Area
Month Jan. Feb. Mar. Apr. May Jun. Jul. Aug. Sept. Oct. Nov. Dec. .E
Montana, United States 3 23 35 36 3 560
Kansas, United States 2 6 29 34 14 10 4 1 0 440
Arkansas, United States 3 3 30 13 29 7 3 5 2 2 3 60
United States 0.2 0.9 3 6 20 25 22 18 3 0.8 0.4 0.6 3200
Ferrara, Italy 4 21 21 24 24 5 0.5 130

The diurnal variation of hailstorms is very pronounced. Generally, less than

Table 7.2 15 per cent of all hailstorms occur from midnight to noon, while the six-hour
Diurnal variation of hailstorms interval from 1500 to 2100 hours generally provides more than 60 per cent of all
in the United States (in percent hail observations. Typical diurnal patterns, averaged over the years 1944-1953, are
of ten-year total L). shown in Table 7.2.

Area Hours (local time in UTC)

0000- 0300- 0600- 0900- 1200- 1500- 1800- 2100- .E
0300 0600 0900 1200 1500 1800 2100 2400
Montana, United States 3 1 1 1 13 45 29 7 560
Kansas, United States 6 5 3 2 7 26 37 14 440
Virginia, United States 1 0 1 2 23 43 24 6 100
United States 4 3 2 2 14 37 28 10 3200

Hail damage in agriculture varies greatly according to the crop and to its
growth phase (Changnon, 1971). Damage is sometimes limited, e.g. if the hail
destroys the leaves at the end of a first growth phase, when a second phase with
new leaves is about to begin. Apart from the direct physical effect, hail-damaged
crops are much more susceptible to disease and fungal infections, which may
provoke further losses. The amount of damage resulting from light and heavy hail
on various crop types is illustrated in Figure 7.2.
Because hail damage is usually very localized, some countries have more or
less extensive hail insurance schemes. Preventive action can only be the

113
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

Figure 7.2
Crop damage as related to the
kinetic energy of hailstones /
/
(Svabik, 1989). /
I /
/
I
/
I /
I /
.peach
/
vine
/
apple
- - - -corn

100 500 1 000 1 SQO

·Energy (Joule/m2)

covering of sensitive structures (such as glasshouses!) with mats, and in this

respect timely meteorological warning of hailfall risk can be of immense value.
Placing anti-hail nets over cro_ps is a disputed measure, since it modifies the
microclima te.

7.3.2 Active suppression Proposals to fight off hail l!Y technot0,gical means have always been around.
of hail Around 1960 for instance, the theory was put -forward that explosions would
make air enclosures in hailstones cave in ana so destroy them. A review of this
References: WMO-TN
idea by an Italian team concluded that an explosion of several kilograms of TNT
Nos. 105, 154.
woµld achieve this, but only at distances of a few metres. A Russian investigator,
Khadziev, phrased a similar conclusion differently: for a moderate effect, one
should fire 100 000 rockets at a single cloud. Careful experiments, carried out in
Austria and Italy as early as 1905 had already shown long ago that shooting explo-
sives at hail clouds produces no noticeable effect. In discussion with gunners it
was said, that such action looked "as if an ant was trying to stop a stampede of
elephants by biting their feet" (Bordreuil, 1979).
In 1978, a new type hail cannon was proposed and was said to modify hail-
producing clouds by means of sound waves (see Changnon and Ivens, 1981). The
'inventor' of this new cannon did not, apparently, consider that far more intense
sound waves (thunder) are produced by the electrical discharges in the cloud with
no noticeable diminuation in hail. It was simply proved by Italian and French
investigators that cloud modification with sound is an illusion. One just wonders
which method to fight cumulonimbus with brute force will be proposed next.
A more meteorological and less mechanistic approach is to use some of the
·seeding techniques discussed in section 7.2, to influence the hailstone formation
process. The basic hypothesis has been called 'beneficial competition': presence of
an excessively large amount of small ice pellets or graupel particles ('hailstone
embryos') produced by 'overseeding' can prevent the formation of large hail-
stones, because too many embryos are competing for the limited amount of water
available in the cloud. Smaller hailstones will do less damage and may even melt
entirely while falling.
Success or failure of this approach depends on using the right technique at
the right place at the right time, and on the right type of cumulonimbus. Practical
use of 'overseeding' methods was first developed in the USSR by Sulakvelidze and
colleagues (see Battan, 1969). Their starting point was that a maturing cumu-
lonimbus may have a region just above the updraft maximum where a large
amount of supercooled raindrops _is accumulating. Detection of the occurrence of
such hail-production centres, typically at an altitude of around 5 km, is possible
with radar. Russian investigators then seeded these small centres immediately with
silver iodide, using anti-aircraft guns and special rockets and claimed a success rate
of more than SO per cent.
Attempts to reproduce this success in the United States failed for several
reasons: they used no rockets, but sprayed silver iodide solution from aircraft
flying below the updraft. Delivery of the nuclei into the hail-production centre

114
CHAPTER 7 - WEATHER HAZARDS ADVERSELY AFFECTING AGRICULTURAL OUTPUT

may thus have been too imprecise, too little and too late. Ground-based spray
generators would have been even less effective. Another reason may very well be
due to the different structure of hailstorm: in the United States graupel particles
rather than freezing drops are the main hail embryo types in the central updraft.
Silver iodide is not appropriate in these situations.
However, another major investigation of hail suppression in Switzerland was
not really successful either, though in this case rockets were used and local cloud
types were more similar to those in Russia (Federer, et al., 1986). The question
therefore remains unsolved, when and how hail suppression by seeding tech-
niques is effective. Experimental investigations are still going on (e.g. Rakovec,
et al., 1990).

7.4 FIRE IN Forest fires and wildfires in bush and grass are a contant hazard in dry seasons.
VEGETATION Prolonged drought obviously increases the intensity of fires that do arise. Usually
the fire starts in lightweight dead dry organic matter, low vegetation or forest
References: WMO-TN No. 42
litter. From these surface fires develop the larger forest crown fires, and the diffi-
(1961), Reifsnyder and
cult persistent fires in dry surface layers. The worst fires generate above their
Abers (1994).
centre a convectively rising column, carrying into the atmosphere burning matter
which can start fires elsewhere. These fires can propagate horizontally at speeds of
up to 10 km/hr.
The most important information needed to assess fire risk is the dryness
measured as water content, of both the lightweight fuels (grass and forest litter)
and the heavier fuels (timber branches and trunks). An index of flammability
derived for one fuel category will not necessarily be valid for the other.
Dryness can be monitored in forests and other fire-prone areas from the
state of some test body, for example a stick of wood of standard dimensions and
known oven-dry weight, placed under a screen 25 cm from the ground; moisture
content then is derived from its changes in weight. However, such variations do
not fully reflect the rapid fluctuations in moisture content that occur in lighter
fuels.
Meteorological data are often used to estimate the moisture content of vege-
tation and wood by empirical locally-determined regressions on humidity,
precipitation, temperature and wind. An example is an index used in Sweden:
B=5R-0.l(T-27)
where R is the relative humidity, written as a fraction (0 $ R $ 1) and Tis temper-
0
ature ( C). If the index Bis less than 2.5, forest fire risk is considered to exist.
Other indices are cumulative; an example from the United States is given in
WMO-TN No. 138. The moisture status of the whole volume of topsoil and
deep litter was estimated daily as follows: if l(i) is the index for day i, and
(i-1 )=yesterday:
I(i) = I(i-l)- 3.94 (P - 5.1) + E
where P = rainfall (mm) of past 24 hours; if it is less than 5 mm, (P-5.1) was taken
to be zero. Unfortunately, in this case the evapotranspiration was derived from the
most recently observed maximum temperature only, resulting in a disastrously
poor reliability as noted in WMO-TN No. 138. However, the sensible part of this
method is that the water deficit was monitored cumulatively. In some countries
this is done by keeping track of soil water reserves.
Next to dryness of vegetation and soil, other useful weather factors to
monitor to assess fire risk are cloudiness (degree of irradiation) and the occurrence
of dry air advection. Such information is needed to decide the necessary degree of
alertness of fire fighters and the public.
Once a fire has started, fire fighters primarily need reliable knowledge of the
wind (Beer, 1991). Other useful parameters are those typical of atmospheric
transport analysis (section 7.5), such as atmospheric stability, and forecasts of
precipitation. A good communication organisation is obviously required.
For assessing climatologically the fire hazards in a region, topography is
an important factor. The effects of slopes on incoming radiation and soil

115
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

temperatures discussed in section 5 .5 are obviously relevant, and moreover

orographic precipitation probabilities are worth analyzing. Also, fires have a
tendency to travel uphill.

7.5 ATMOSPHERIC Airflows carry many things: gases, liquids and solids, dead and alive. Originally
TRANSPORTS most investigation of such transport were made in the context of 'danger advec-
tion': air pollution, locust swarms, fungi and spores, radioactive matter or
References: WMO-TN Nos. 96,
sandstorms. However, also beneficial transport can be mentioned, such as polli-
121, 147, 177; Munn and
nation - even though that carries for many the negative association of hay fever
Rodhe (1985).
(see Smith, 1975). The questions raised are twofold: first "where does transported
matter go?", and second "does it accumulate or disperse?"
These issues are the subject of an entire WMO Compendium (Munn and
Rodhe, 1985) and WMO Technical Notes (e.g. Nos. 121; 177), as well as of numer-
ous books. It would take us too long to do more than give a short review of some
basic aspects of mesoscale atmospheric transport, followed by a few examples from
agriculture.

7 .5.1 Elementary aspects of There are three phases in atmospheric transport:

transport over mesoscale (1) Emission of the transported matter from some source. Usually this is thought
distances to be a point source (a chimney, an infection source, a spray nozzle), but it
can also be a line source (road pollution, sea-salt from coastline), or an area
References: Nieuwstadt and
source (pollen from orchard, sulphur dioxide from industrial area).
Van Dop (1982), Pasquill and
(2) Transport of the matter through the atmosphere. The duration of transport
Smith (1983), and Stull (1988).
may imply changes in the transported matter - a good example is photo-
chemical smog, emitted as pollution by car exhausts and industry and then
modified in the atmosphere by radiation and humidity.
(3) Deposition of the matter on the ground, sometimes by itself ('dry deposi-
tion'), sometimes by rain.
The basic context of most transport models is that of small-scale turbulent
diffusion, superposed upon an underlying regular mesa- or macro-scale airflow.
The scale of the diffusing turbulence is determined by its sources, surface friction
and convection.
The relative importance of convection is determined by the atmospheric
boundary layer stability, which is not quite the same parameter as the stability
used in meteorological forecasting. In the troposphere, from 1 km to 10 km
height, the stability parameter is the vertical temperature gradient dT /dz, deciding
whether atmospheric vertical motions at that scale are enhanced or suppressed. In
the boundary layer, from the surface to 1 km height, the degree of vertical turbu-
lent motion is not sufficiently determined by dT/dz. A striking example is the
·vertical transport on a sunny day. Large convection currents rise from the heated
surface and produce very strong vertical transports - but they mix up the bound-
ary layer so well that the temperature gradient becomes adiabatic, and the (dT/dz)
criterium would imply 'average transport'.
Therefore boundary layer stability parameters for calculating the degree of
vertical transport must combine convective turbulent transport (essentially verti-
cal) with the other source of transport, often called mechanical turbulence, which
is caused by the interaction of average airflow with surface roughness. Mechanical
turbulence is mostly non-directional, and if it is large it diminishes the generation
of strong convection because it cools the surface. Mechanical turbulence also
prevents strongly stabilized flow over cold surfaces, because its vertical transport
counteracts large surface cooling.
The most generally used boundary layer stability parameters are the Obukhov
length L, expressed in terms of fluxes, and the gradient Richardson number Ri,
expressed in terms of vertical differences of wind speed and temperature.
The Richardson number relates to particularly specified heights and requires
profile mast data; therefore Ri is a good stability parameter to use in research, but
less applicable for routine station use.
For analysis the Obukhov length has the advantage that it is approximately
independent of height in the boundary layer, and its ratio with height, z/L,

116
CHAPTER 7 - WEATHER HAZARDS ADVERSELY AFFECTING AGRICULTURAL OUTPUT

accounts automatically for the fact that convective stability becomes more impor-
tant at great heights, far away from surface roughness.
Approximate values of L (in m) are:
Stability Very unstable Unstable Near-neutral Stable Very stable
Obukhov length -2 to-20 -20 to-200 ±200 to 00 +20 to +200 <+20
For small z/L values, z/LI <0.01, the statement that boundary layer stability is 'near-
neutral' means that mechanical turbulence dominates convective turbulence.
The ratio of height and Obukhov stability parameter, z/L, is by far the best
descriptor of the boundary layer, and it can be determined from ordinary obser-
vations of a good meteorological station: temperature, wind speed and radiation
- the latter can be substituted by data on cloudiness at the price of extra uncer-
tainty (see section 2.1.2) - and surface information: albedo, surface roughness,
and crude estimates of soil wetness and degree of surface cover (Van Ulden and
Holtslag, 1985).
A much simpler approach has been used in pollution studies for a long
time: estimate convective turbulence from degree of cloudiness and mechanical
turbulence from wind speed. If roughness information is available, a useful
estimate of Obukhov lengths is even possible (Golder, 1972). Results of such
estimation are simply expressed in six Pasquill or seven Turner classes, and can be
summarized by:
(a) Little cloud and weak wind in daytime: unstable
(b) Little cloud and weak wind at night: stable
(c) Overcast and/or wind speeds >5 m s- 1 : near-neutral
The accuracy of this stability classification approach is decidedly less than that of
Obukhov length application. The point remains that for analysis of atmospheric
transports it is necessary to account for stability in some way, however simple.
Apart from the importance of boundary layer stability, another important
aspect in atmospheric transport is the height of the turbulent boundary layer. In
most practical cases, particularly in daytime, this means the height of the capping
inversion. The daytime convective turbulent boundary layer is typically 1 km
high, and at its top there is a sudden increase in (potential) temperature. Above
this boundary layer, the turbulence is absent.
The inversion over homogeneous terrain in daytime is generated by convec-
tion processes, but there are other daytime inversion sources: large-scale
subsidence, fronts, differential advection at a coast or in mountains, occurrence of
urbanized areas. Also, if there is an overall positive vertical gradient of potential
temperature, mechanical turbulence transforms its boundary-layer section into an
adiabatic profile, resulting in a localized inversion at the top of the turbulent layer.
Atmospheric behaviour below the daytime inversion will only be turbulent below
the capping inversion. At night the situation is different: in cloudless 'radiation'
nights there will be a surface-based inversion with stable flow, only slightly turbu-
lent. Strong wind and/or much cloud cover can prevent the formation of such
radiation inversions, in which case stability becomes near-neutral.
With this knowledge about the basic parameters for atmospheric boundary-
layer transports, stability and inversion height, we can look at applications.

7.5.2 Point sources A typical point source is a chimney stack of height H. For its emission we can
picture two extreme cases. One is a situation with a non-turbulent wind speed U
References: WMO TN Nos. 121,
at height H, in which case the emission will be carried away as a narrow pencil-
177; Scorer (1978); Kramer and
like stream in which the concentration is proportional to U- 1 - i.e. only dilution
Porch (1990).
by U. The alternative is a situation with no average windspeed, in which emission
is spread out by turbulence at stack height into a spherical puff - i.e. only disper-
sion. Usually dilution and dispersion occur together, and are then modelled as
Gaussian turbulent spreading around an axis aligned with the average wind. The
formulae look impressive, but the results are imprecise.
Gaussian plume models assume that wind speed and degree of turbulence are
height-independent (the resulting errors compensate for each other somewhat),
and that at the surface plumes undergo no absorption, but reflection. Degrees of

117
WMO LECTIJRE NOTES - AGRICULlURAL METEOROLOGY

turbulent spreading are estimated from the Pasquill stability class by empirical
relations, which have been validated for distances of at most 10 km. Horizontal
changes such as coasts or inversions can only be handled with empirical tricks, so
that daytime averages are not more accurate than ±50 per cent. There are also
Gaussian plume models for surface sources.
Still, these models go some way toward answering the question "where does
it go?" As instability increases, the position of maximum average ground level
concentration moves closer to the stack, and under very unstable conditions can
occur within a distance of three stack heights from the source. In stable stratifica-
tion, nearby concentrations are low, but high concentrations are found at
considerable distance from the source. Also, the surface concentrations are propor-
tional to H- 2, an incentive for building high stacks.
Foot-and-mouth disease is caused by an airborne virus, carried in small water
droplets in the breath of infected animals long before symptoms are manifest. If
relative air humidity is more than 60 per cent, the droplets survive and are carried
away by the wind. The probable locations of new infections can then be found by
transport analysis for weather with sufficient humidity. In plains with a dominant
wind direction the Gaussian plume model can be used for that purpose (Pedgley,
1982; Rumney, 1986), but in more complex terrain a determination of infected air
trajectories requires more insight. Figure 7.3 shows an interesting example of an
outbreak in mountainous terrain, where lee wave effects generated further infec-
tions in three dusters at intervals of about 20 km.
Figure 7.3 400
(I) ........

Foot-and-mouth disease > E

occurrence in Wales in 1967 as ~i

..... > 200
.!: (I)
0)-
related to the mountain lee ·a:; :ii
J:.,,
waves which probably 0
facilitated the transport of the 0 10 20 30 40 50 60
disease (Pedgley, 1982). Distance from first outbreak (km)

Many agriculturally important factors are airborne, including other animal

infections (such as Aujesky's disease), fungus spores and insects. In all cases distri-
pution can be usefully analysed meteorologically; again, it pays to apply more
meteorological knowledge than modelling of stationary plumes.
For instance, aphid swarming is supported by boundary layer convection in
unstable air. ·Observations in England showed that the number of aphids per m 3 of
air depended on the vertical temperature gradient in the lower atmosphere.
Assuming aphid density at 1 m height to be 100 per cent, the following densities
occurred higher up:

Temperature gradient Height 10 m 50 m 500 m

aT/az = -1.0 K/1'00 m (near-neutral stability) 9.1% 1.7% 0.2%
aT/az = -1.4 K/100 m (unstable) 28.2% 11.6% 3.3%
This stability dependence in swarming leads to the occurrence of a diurnal course
in aphid distribution speed.
Another meteorologically interesting case was the spreading of rust disease
spores from an affected lake shore location to trees situated approximately 15 km
inland, without infection occurrence in the intermediate region. Van Arsdel (1965)
discovered that the spores were only released at night, picked up by a land breeze circu-
lation, carried out over the lake and then upwards; from where they returned
landwards at a height of about 500 m to be deposited in the newly-infected area by
the downward part of the breeze circulation.
The most disastrous form of meteorologically-carried 'insect infestations' are,
of course, locust swarms. Their basic development is climate-dependent at every
step. A wet period is necessary for hatching of the eggs, as well as the growth of
new green vegetation needed for feeding the newly-hatched 'hoppers'. Their
swarming is then furthered by the wind, or inhibited by low temperatures which
chill the flight muscles. Extensive discussion of the application of meteorological

118
CHAPTER ?-WEATHER HAZARDS ADVERSELY AFFECTING AGRICULTURAL OUTPUT

knowledge to this problem is given in WMO Technical Notes Nos. 54 and 69 and
by Scorer (1978).

7.5.3 Line sources, e.g. For line sources, dispersion models should differ slightly from point source
sea-salt transport models, but in practice they are often analysed in the same fashion. An interest-
ing and important example of possible transports from a line source is inland salt
transport at sea coasts.
Sea-spray and salt damage are matters of considerable importance in coastal
areas. Tiny droplets result from green spray, torn from breaking wave-tops or wave
impact with shore faces and cliffs. The salt from these droplets can be deposited on
coastal land, either washed down by rain or as solid particles when the water has
evaporated.
Although only narrow coastal strips appear seriously affected by sea salt,
such areas may be highly favourable for cropping. In temperate maritime regions,
these are areas with minimal frost risk and comparatively mild winter tempera-
tures. In lower latitudes, e.g. Mediterranean regions, such areas escape some of the
hazards of the hot dry weather of inland sites and, furthermore, population - and
hence market outlets - tends to be concentrated in coastal regions.
Measurement of salt concentration is difficult; three methods are in wide-
spread use:
(i) The 'wet candle' or 'paper candle' method: a filter paper roll is inserted into
a wide-necked bottle of distilled water placed under a shelter to keep off
direct rainfall; salt is deposited on the paper, which is periodically (e.g. once
a month) analysed for chloride;
(ii) The chloride concentration of rain water is measured;
(iii) The 'particulate' method in which the number of chloride particles deposited
on a chemically sensitised surface is counted microscopically.
As an illustration, Amble (1959) gives data on period-averaged salt concentra-
tion on the regional scale from coastal and inland sites in the United Kingdom and
a few from Nigeria, near sea areas which were rough enough to create spray, using
the 'wet candle' method. Typical values are (depositions in mg NaCl m-2 darl):
Distance inland (km) 0.16 0.32 0.8 1.1 1.6 3.2 8 16 80 160
Deposition 427 253 124 96 74 43 21 13 4 2
As might be anticipated, the concentration falls off rapidly within the first one to
two kilometres (the region of microscale and mesoscale phenomena) and then
more slowly (on the mesoscale and regional or macroscale). A number of places
near the coast were found to return values appreciably lower than those indicated
by these results, and some inland stations gave higher than indicated. Many of the
anomalies for coastal stations could be traced to protection afforded by quite small
topographic features (low hills) between the sea and the measurement site.
Anomalously high inland results were clearly associated with industrial
sources. Airborne industrial effluent contains a very significant proportion of chlo-
rides. Careful siting of equipment can reduce errors from this source.
The effectiveness of shelterbelts to protect vegetation against salt damage is
certainly as much due to the capture of particles by the barrier itself as to the
reduction in wind speed. The effects of windbreaks on the change of salt content
with distance inland from the seashore were investigated by Lomas and Gat
(1967) for the possible benefit to the production of oranges in groves near the
seashore. They measured salt concentration on level sandy ground inland from
the sea edge at some twenty sites up to 6 000 m distance, and also at 900 m,
1 250 m and 1 500 m adjacent to an orange plantation which extended for 800 m
from the western edge inland to some 2 000 m. Salt deposition on the central 10
x 10 cm of a 20 x 20 cm piece of double muslin, kept facing into the prevailing
wind, was measured at 10-day intervals. The following relationship obtained over
the range 10 m to 1 500 m:
y = 7.84 - 2.30 lOlog X with y = mg NaCl (10 cm2)-1 (10 days)-1,
and x = distance from sea (in metres).

119
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

Table 7.3
Distance x Mean yield/tree (kg) Proportion of yield
Yield (height H) (on either side
x/H m Windward side Leeward side Total from seaward side of
of the tree and total) per tree of
trees(%)
citrus crops at distance x from
a windbreak. 7 56 63 91 154 41
10 80 61 80 142 43
13 104 26 61 87 30
16 128 6.7 31 38 18
19 152 4.5 17 22 21
22 176 5 18 22 21

Beyond 1 500 m and up to 6 000 m, the concentration levelled off and varied
usually between 0.3 and 0.4 units, i.e. about 5 per cent of that at 10 m from the
sea. Salt concentration was described as being very high in the first 1 000 m,
decreasing steadily up to 2 000 m. The consequence in terms of yields at about
1 km from the sea are illustrated in Table 7.3.
It was concluded that on the coastal plain most damage was caused to
plants within two to three kilometres from the sea and at plantations that have
slopes exposed to winds from the sea. Windbreaks (height H) can reduce this
damage, but the benefits do not extend much beyond 13 H. In practice the yield
required for economic operation is achieved at a distance of up to 13 H. From
which it can be seen that regular shelter belts are needed at this distance from the
sea.
There is a significant asymmetry between yields on the two sides (windward
and leeward) of the tree. The asymmetry was greater and total yield decreased as
the distance from the windbreak increased. Edwards and Claxton (1964) also
report on wind effects in salt damage.
Similar considerations arise in relation to the application and control of
pesticides, agricultural chemicals and manure. The effectiveness of these activities
depends on a range of meteorological elements: rain dilutes material and trans-
ports it into the soil, either to the desired root depth or further down; wet surfaces
can retain and spread materials. Most chemicals are temperature-sensitive and all
reactions are temperature-dependent; solar radiation can evaporate agents, or
their carrier water, for manure it can produce a crust retarding evaporation. Using
weather information for timing of slurry spreading can prevent undesirable losses
of e.g. NH 3 (Riley and Giles, 1965; Frost, et al., 1990).
With respect to possible spray drift, some distances of horizontal travel during
spraying (at the surface, not from an aircraft) in a steady wind are given in Table 7.4.
Table 7.4
Drop diameter (mm) Distance carried while
falling 3 m in 2 m s-1 wind
Coarse spray 0.4 3m
Fine spray 0.1 15 m
Fog-spray 0.03 250m

The worst weather conditions for drift are when a temperature inversion
exists combined with weak wind (under 3 m s- 1) because in such stable boundary-
layer conditions the downward vertical ·transport is slow. Here the drift residue
downwind can be as much as ten times higher than during suitable weather
conditions (Akesson, 1964).

Review questions on Chapter 7

A. Questions on basic understanding:
7.1 What are the main characteristics of 'drought' as understood by:
(i) A hydrologist;
(ii) An agriculturist?
To what extent are their different requirements for water in conflict?

120
CHAPTER 7 -WEATHER HAZARDS ADVERSELY AFFECTING AGRICULTURAL OUTPUT

7.2 "A state of drought exists when annual rainfall is 75 per cent of average,
or when monthly rainfall is 60 per cent." Criticize this (admittedly)
incomplete specification of drought from the standpoint of:
(i) A farmer;
(ii) A forester concerned with fire prevention;
(iii) A water engineer.
7.3 Write an account of the value of meteorology in the control of forest
fires.
7.4 Write a brief account of hail as an agricultural hazard and explain why
it is so difficult to obtain hail statistics.
7.5 The relationships given in section 7.5.3 indicate a rapid reduction of
airborne sea-salt with distance from the coast. Describe and explain situ-
ations in which this general statement might not be valid.
7.6 Explain why some of the most serious cases of downwind locally
concentrated pollution damage are associated with clear skies at night
and light winds.

B. Questions on extended application:

7.7 The rate of fall of a small spherical particle through a non-turbulent fluid
medium is given by Stokes' Law:
2 2 (CJ - p) -1
v =- gr - - - cm s
9 Tl
where r = radius of spherical particle (in cm); CJ = density of particle
(g cm- 3);
p = density of fluid medium (g cm-3); Tl= dynamic viscosity (g cm-1 s-1);
g = gravitational constant.
Assuming the particle is a water droplet and that p (air density) may be
neglected in comparison with CJ, and ri = 1.8 x lo-4 g cm-Is-I, show that:
v = 1. 21 x 106 r2 cm s-1
If the droplet falls 3 m in a uniform, steady horizontal air current
moving at 5 km per hour, show that for:
d = 20 µm the horizontal travel of the droplet is about 0.3 km
d = 100 µm the horizontal travel of the droplet is about 13.5 m
d = 400 µm the horizontal travel of the droplet is about 1 m
How do you expect these results to change when the air is turbulent?

C. Questions involving practical work or local data:

7.8 For a yield of corn in Southern Iowa to be defined as satisfactory,
G.L. Burger and H.C.S. Thom (1949) postulated that a minimum amount
of rainfall is required in the course of the growing season. This
minimum value R (mm) varies from zero for two successive weeks to:
R = -15.2 + 9.27 n + 0.706 n2
where n = number of successive weeks (1 ~ n ~ 16).
Examine seasonal yields of an appropriate crop in your area against the
total rainfall in successive weeks and investigate the possibility of estab-
lishing a similar type of relationship.
(Note: Statistical procedures such as the one above were common prior to the
extended use of water balance methods; these earlier approaches continue to be
valuable especially when climatological data are sparse or incomplete.)

121
CHAPTER 8
OPERATIONAL AGROMETEOROLOGY

Throughout this compendium, discussions of theory and principles have been

undertaken with their practical application in mind. In this chapter our attention
will be focused on how meteorological knowledge and expertise can be mobilized
and utilized to tackle specific problems of agriculture. The basic subdivision of
these matters is given in Chapter 1: meteorological assistance in this field can be
either strategic or tactical.
Strategic agrometeorological action operates on time scales of months or
years. It deals with planning and commitment of resources for future activities,
either on the farm (e.g. land-use selection, material investment, alternative crops
or livestock) or in society around it (e.g. harvest expectations, planning process-
ing of farm output). Climatological studies and crop-weather modelling are
appropriate methods, and the action is off-line and therefore can be in the hands
of any group with meteorological professional knowledge and adequate data -
not necessarily in the weather service.
Tactical agrometeorological action is geared to the optimal handling of all
available resources on the farm, for example the choice and timing of actions for
the next week, starting from the actual crop and weather situation and using avail-
able committed resources. It requires good three- to five-day weather forecasts and
knowledge about weather influences on farming. Since the action is on-line direct
counseling to farms, the counselors must be either in the weather service or
actively cooperating with it.
Logically speaking, there is a third alternative, namely that no special action
is taken. This assumes that agriculture is suitably served by the standard meteoro-
logical products for the average citizen. However, experience in many countries
has shown that progress will be slow if the Meteorological Service interprets its
role as a passive supplier of information and data. Even in this capacity it can
often fail because, if counseling is to succeed, the right amount of material has to
be available in the right form, issued at the right time and valid for the right
periods. The data must be correctly described, processed and presented to meet the
needs of users, or at the very least in such a form that the user can process his
material efficiently (Wang, et al.,1982).
On the other side, the meteorologist must not appear to take over the author-
ity and responsibility of the user. For example, warnings relating to disease must be
carefully worded. For instance a warning in relation to potato blight must clearly
be a warning of weather conditions that are suitable/favourable to the initiation,
incidence and spread of potato blight. It must not be open to misunderstanding as
a 'warning of potato blight'. There are a limited number of weather warnings which
are practically directives; e.g. warning of wind conditions likely to lead to the drift
of poisonous sprays outside the area of legitimate application, or to storm damage;
but these types of warning are rarely given. Agrometeorological advice is generally
decision-supporting: the farmer makes the final decision, as he knows the state of his
crops and his soil, and the means that he has at his disposal.

8.1 ALTERNATIVE A simple example of operational advice using weather factors (McQuigg, 1968;
FORMS OF WMO-TN No. 132) relates to the scheduling of the sowing of cotton in South-East
AGROMETEOROLOGICAL Missouri (United States). The basic question of "should we sow cotton today or
DECISION-SUPPORTING not?" also implies "is the soil dry enough to get on to the land?" The decision
ACTIVITY sequence is based on the following series of questions and answers assigned.
It is helpful to analyse this case to uncover the various topics and stages in
the argument leading to the apparently straightforward chain of decisions:
(i) It has been established that the process is weather-sensitive. This is a vital stage in
the process and obvious when stated in this case, but it is often overlooked, given
too little emphasis, or assumed to be self-evident.

123
WMO LECTURE NOTFS - AGRICULTURAL METEOROLOGY

Is the soil dry enough?

No -Yes
I
Was the soil temperature at a depth of75 mm (3 inches) at least 20°C yesterday afternoon?

No
·:I
Is the five-day forecast of soil
temperature favourable?

No Yes
I
Sow

(ii) Practical experience and field investigation will have established criteria for decid-
ing whether or not the soil is in such a condition as to allow the use of machinery
without unacceptable risks to the soil structure; this may well be at least one or
two days after the moisture .content of the soil has dropped below 'field capacity'.
The question will probably be judged by inspection.
(iii) It will have been established from practical observation (preferably confirmed by
experiment) that soil temperature in the first period - in this case 10 days -
after sowing is critical for the germination and initial establishment of a healthy
seedling. However, forecasting beyond about 5 days is not regarded as sufficiently
reliable, and a calculated risk has been taken in basing advice on a five-day
forecast.
(iv) It will have been agreed that the soil temperature can be specified by its value at
a depth of 75 mm (3 inches). However, temperature at such depth usually shows
a large diurnal fluctuation in the spring (except if the weather is persistently over-
cast), and a procedure has to be adopted which is both within the skill of the
forecaster and operationally convenient for the farmer.
From a forecasting point of view, estimating the diurnal maximum presents fewer
difficulties than estimating temperatures at other points in the diurnal cycle, a late
afternoon reading is convenient for the farmer. Accordingly, the temperature crite-
rion is expressed in terms of 'not less than a maximum reading of 2O°C at about
1700 local time at about 75 mm depth'.
(v) Given a 'yes' verdict, the farmer still has to decide on whether to act upon it. All
·forecasts are subject to some uncertainty, and only the farmer can balance the
chances of a correct forecast against all his other farming priorities.
The above-mentioned example of an agrometeorological operation has quite
a simple structure; even so practical complications can easily arise:
(a) The relationship between the weather and the crop characteristic or the agri-
cultural operation may be continuous and not of the 'yes-or-no' type; then
more gradual quantitative information is needed to support decision-
making;
(b) Relationships between the biological (crop) data and the meteorological data
are subject to appreciable error. Three basic forms of error exist, namely inap-
propriate or faulty data, erroneous forecasting, and uncertain weather-crop
relationships.
(c) Other factors, or the sequence in which they occur, influence the outcome of
a weather-based decision. One example is the market value, when this varies
with terms of trade (internal and external), rates of exchange, changes in
interest rates, etc. Another example is the influence of labour availability.
However, with or without complications, it is obvious that weather services
can support agriculture explicitly and efficiently in this manner. Nowadays the
possibilities for such assistance are much larger than 20 years ago, because of two
developments:

124
CHAPTER 8 - OPERATIONAL AGROMETEOROLOGY

(a) Reasonably reliable forecasts for more than one day are presently available and
these are necessary for many such actions. Examples are given below.
(b) Radio communication has been available for a long time, but with present-day tech-
nology its costs and energy requirements are so small that it is available at the most
remote locations. Moreover, the availability of fax machines and the advances in
communications technology, including the Internet, allow transmission to a large
public of more extensive information (e.g. maps, datasets) than is feasible by radio.
As a result, tactical agrometeorological counselling can now be made fully
available to the farming community. It is therefore unfortunate, that in many
countries nowadays the political philosophy of running down public services has
led to many weather services not being able even to provide agrometeorological
special data at an acceptable cost (Wieringa, 1996b). With such public service, the
farming community is either unaware of potential possibilities, or has to be
content with low-grade information.
Strategic application of agrometeorological knowledge has less practical
restrictions for various reasons, mostly related to the fact that its urgency is less
immediate than is the case for tactical action. There are two main channels to
deliver strategic weather and climate advice to the agricultural community, to
farmers and to whoever else has to deal with the agricultural working situation or
with agricultural products. The classical approach is to use printed matter -
climate atlases, methodological reports and such. This climatological approach is
illustrated further in section 8.4.
The present-day alternative is the use of a computerized model by whoever
has to apply agrometeorological knowledge - commercial or governmental
organization, agricultural counselor, or the farmer himself. This is discussed at
length in section 8.5.
It should be noted that in (sub-)tropical countries agrometeorological action
tends to be mainly of a strategic nature. First, the weather there is steadier than in
temperate latitudes, so small-scale climatological information may be sufficient to
explain some weather effects on agriculture; then, day-to-day weather forecasts are
often not broadcast in such areas; and third, such forecasts might not even be
useful for agrometeorology at present, unless the farmers have resources for short-
term modification of their activities (Haverkort, 1987).

Forecast for farms For a temperate-latitude weather service, agrometeorological forecasts should be
an explicit activity. These need not be on prime-time television as long as they are
rapidly communicated to users, by radio, fax or e-mail. The content of such fore-
casts should make it possible to estimate for the next few days the progress of
evaporation, of soil state, of favourable circumstances for pest outbreaks and other
relevant parameters. If possible, the information should be 'translated' into agri-
culture-relevant terms - some examples are given here below. At least, solid
quantitative data should be given on subjects important to agriculture; even if to
the non-farming citizen, they are only interesting qualitatively, or not at all.
The types of forecast presentation which are possible depend on the active
two-way cooperation with agricultural consultants and an adequate communication
network. One illustration is the following extract from an 'Agricultural Aviation
Advisory for Mid-South (United States) issued at 5 a.m.', which is quoted in full by
Riley and Giles (1965). After describing current weather the notice continues:
"Skies will be clear today except for one-to three-tenths of cumulus clouds at
1 200 m. Dew will dry by 9 a.m., winds will be less than 2 m s-1 till 9 a.m., will be
south 3-5 m s-1 by 9 a.m. until noon, south 5-7 m s-1 this afternoon but less than
2 m s-1 by 6 p.m. No rain today or Tuesday but about a third of the area will have
showers Wednesday.
Good dusting and spraying conditions today till 9 a.m. and after 6 p.m. Dew
will be moderate to heavy tonight. Temperatures will be 21 °C or higher and condi-
tions are favourable for applying defoliants. The Extension Service advises that
bollworms are still prevalent in many young, green cotton fields and should be
controlled. Some older, dry fields do not need control. If in doubt, call your
County Agents."

125
WMO LECTIJRE NOTES - AGRICULTIJRAL METEOROLOGY

A second example is quoted by Callander (1990), and was evidently issued

during a drought when irrigation, not crop protection was the main agricultural
problem:
"Free-water loss during past 24 hours average 0.6 cm. Expected free-water
loss today is 0.6 cm and tomorrow is 0.8 cm. Rainfall probability will remain low
for the remainder of the week, and crops will begin to suffer from moisture stress
unless supplementary water is applied."
This example was already directed to a longer forecasting term than one day
and the third example from the Netherlands (Radio Gelderland, 28-11-1994),
broadcast on a Monday, was meant to help with farm planning for about five days:
"The humid and mild weather is at its end, and the air has already become
drier with dewpoints around S°C. Advection of dry and increasingly cold air
continues. After today frost is possible, and from Thursday onwards it will freeze
every night. Wind will become easterly, moderately strong, Beaufort 3 to 5, and
the maximum temperatures go down to 3° to 5°. Cloudiness decreases and will be
negligible after Thursday. Evaporation will be about 0.5 mm per day. Global radi-
ation in the next few days will be between 400 and 500 J cm-2, afterwards between
500 and 600 J cm-2 . For the week afterward the preliminary preview is quiet
weather, only light freezing at night, but again increase of dewpoints and higher
probability of fog.
Keep covers handy, both for stacks of sugarbeet, etc. and for the field,
because in sheltered spots it is quite possible the surface temperature will fall to
-7°C. In clear weather, plant temperature can easily be half a dozen degrees lower
than screen temperature at a height of 1.5 m. As soon as minimum temperature
forecasts approach 0°, one must consider covering. Cattle farmers should ensure
that water pipes and field pumps are drained to prevent freeze-up damage before
Wednesday night.
In the course of the week the soil may freeze to a few centimeters at night.
Early in the week this will thaw in the afternoon, but it will remain frozen later in
the week. Harvesting of leeks and similar vegetables will therefore require careful
planning. Unploughed land can be left as it is this week, let the frost do its useful
work."
None of these three forecasts would interest the general public.

8.2 OPERATIONAL Many models exist for the practical handling of weather influences on crop
MODELLING FOR behaviour, these vary from simple rules-of-thumb to intricate computerized
TACTICAL models. Their output can be divided into two groups. The tactical type is used by
AGROMETEOROLOGY the farmer or the agricultural counsellor to optimize the tactical activities at the
farm. The strategic type, which is discussed further in section 8.5, is used by
References: Lowry (1967 /1989),
· farmers and by managers at various levels of agricultural economics, trade and
Perarnaud and Raynal (1991),
politics in order to optimize resource-committing actions.
WMO-TN Nos. 151, 174, 180.
For historical reasons, tactical models are usually quite simple and include
only temperature and precipitation, averaged over periods of at least 24 hours. The
basic reason is that these data are available from all weather stations in the world,
and can also be measured locally with relative ease. The basic problem is that these
two parameters are often not the correct input for modelling crop behaviour.
Water availability should refer to soil moisture availability, but for practical
reasons precipitation data are generally used. Radiation is the real input for photo-
synthesis, while temperature is important to growth rates and respiration losses.
Replacement by more sophisticated models has little chance, unless the
agrometeorologist can do two things. First, he must prove that a 'better' model
indeed gives significantly better information, leading to markedly improved
results. Second, he must be able to access the necessary input data. The second
point indicates a vicious-circle problem: better observations will not be made
unless they are proved useful, and the proof cannot be given without observed
data ....
For tactical purposes, much use is made of temperature sums. The basic ideas
behind this are that development and growth do not really take place if the
temperature is below a certain threshold temperature T0 , and that growth can

126
CHAPTER 8 - OPERATIONAL AGROMETEOROLOGY

Figure 8.1 development ,3

Three different linear fittings of rate
the same typical crop
development curve, using data
from temperature ranges S1 to
S3 . Note the differences in
corresponding threshold
temperatures T1 to T3 !
(Delecolle, 198 7).

therefore be usefully approximated by a simple (usually linear) function of DT =

(T-T0 ), where Tis usually a 24-hour average temperature. OT is set to zero if T<T0 •
Starting from the time of seeding to take one example, the OT-values are summed
day by day, and at some value of LDT it is assumed that a particular development
stage has been reached and a particular action, for example soil fertilization, irri-
gation, or the use of herbicides is necessary.
The existence of temperature thresholds for germination of particular crops
was encountered earlier in Table 6.1, and in the cotton-sowing example of section
8.1. The first uncertainty lies in the assumption that thermal development time
afterwards is linearly related to temperature. Because the typical shape of a devel-
opment curve is not linear (Figure 8.1), it is advisable to use a specific base
temperature corresponding to the considered growth stage - see Yang, et al.
(1995) for deriving it. Using different data sets may produce significantly different
linear approximations.
The second problem is that weather progress is often very unsteady. If a period
of freezing weather occurs while temperature summation is proceeding, it is debat-
able to what extent the LDT-values accumulated immediately after freezing will
take full effect. If soil temperatures are unusually low, high air temperatures will not
promote growth in the way they would otherwise (Wartena, 1996).
The overall problem in using temperature sums is, of course, that the value for
growth of a 'heat unit' (T-T0 ) may vary from day to day, according to other weather
elements and to what happened in the previous period - apart from the fact, that
for practical reasons the standard screen temperature is used, while we really ~ant to
know the temperature of the plant and of its roots as well (Lowry, 1967).
Figure 8.2 All the same, temperature sums are a useful first-guess method to estimate
Relative yield (in per cent) of the necessity of particular measures. However, for a more accurate forecast of crop
grassland after application of development it is necessary to have real insight into the full range of weather
nitrogen when the sum of non- influence on crop development from germination onwards - and to keep track of
freezing temperatures since weather development in the course of a year, day by day, using that insight
1 January reaches LDT. actively. Use of temperature sums must not become a pretext for ceasing to think.
Soil nitrogen application planning is an interesting modelling subject,
110
% ... . . . . because model purposes vary with the precipitation climate. In rainy climates,
the planning problem is fertilization at the optimum time in spring, when the
. ...7.. ~~·
100

crop really starts to grow. Earlier nitrogen application is wasted, as it will leach
I·.. .. ·~ . :
•• I\,.
90
into the deep soil and not be available when the young short-rooted plants need

.1 . . ' \ .
80
it - later application delays growth and gives less yield. This is a typical subject
"C
cii 70
>= r - ·~
for application of temperature sums. In the Netherlands, the time for applying
fertilizer to grassland is estimated by adding positive day-averaged temperatures
60
. from early January onwards. The result in terms of dry-matter yield is illustrated
..
50 in Figure 8.2; nitrogen is typically applied when LDT is at least 200, taking into
40 account if root action is sufficient after actual late-winter soil temperature
development.
30
0 1 00 200 300 400 500 600 The leaching problem in rainy climates also warrants meteorological support
IDT later in the year. When the forecast indicates heavy rain, followed by a long dry

127
WMO LECTIJRE NOTES - AGRICULTIJRAL METEOROLOGY

Figure 8.3 spell, farmers should be advised to apply nitrogen just around the last showers in
Wheat yield in arid climate, order to make up for leaching loss. If they do it earlier, the fertilizer itself will be
related to the application of leached; if they do it later, the nitrogen will not reach the roots in time.
nitrogen with different In arid climates, a further complication is the quantity of nitrogen (N) to be
quantities of water supply applied related to the availability of moisture. Figure 8.3 shows that the optimum
(Lomas, 1984). level of nitrogen application is 100 kg N ha- 1 if the growing season has only
200 mm rain, but 150 kg N ha- 1 with 500 mm rainfall (Lomas, 1984). So for
Yield in + rainfed wheat, N-application will vary between different regions, and in the same
_ 1 000 kg/ha region from year to year (seasonal rainfal_l forecasts would be desirable). The most
4
sensible approach would be a moderate application of fertilizer at the time of
sowing, and additional top dressings if actual rainfall conditions justify such
3- action. Applying fertilizer with irrigation water is another possibility. In modelling
terms, it makes no sense here to calculate temperature sums, but it is very practi-
cal to add precipitation amounts after the sowing date (see section 4.6).
It deserves stressing here that in the very desirable objective of using mete-
orological data - including forecasts - to optimize production, it is essential to
bear in mind that the sequence of seasonal operations must be completed.
Delaying some particular operations in the hope (or expectation) of near optimum
+ = optimum N rate conditions for their execution can mean that, for example, the effective growing
season is seriously shortened or that a group of farm operations, which should be
Kg N/ha
0 -------,------,----..--- spaced, could occur simultaneously and be beyond the human and material
0 50 100 150 resources needed to complete the many tasks faced by the unit.

8.3 PROTECTION OF A major part of farming deals with the protection of crops and livestock against
CROPS AGAINST PESTS diseases and pests. Most of these are weather-sensitive in their own way, though
AND DISEASES for many of them such sensitivity is insufficiently quantified. Agrometeorological
aid in defending crops can serve to increase the efficiency of possible defence
References: WMO-TN Nos. 10
measures by forecasting weather situations where damage risks are large, or when
(1955) to 192 (1988).
pests or diseases are vulnerable, or when methods for fighting pests and diseases
are most effective.
Simple advisory methods which are pest-specific, crop-specific and often also
pesticide-specific have been developed for, among others, slugs, caterpillars,
beetles, eelworms, aphids, funguses, bacteria and viruses. Sometimes weather
influences on disease occurrence are indirect; e.g. mild winters can cause unusu-
ally high survival of particular insects which transport fungus or bacterial diseases
to plants. Out of the multitude of simple crop protection models reviewed by
Seghi, et al. (1996), four classic examples are presented below.
Fungal diseases are transmitted to crops by way of spores, small granules
which are transported by the wind, by dripwater, and sometimes by insects.
'Generally the spores by themselves are immune to fungicides. When spores are
virulent and present on the plant, they usually can only develop into harmful
fungi in sufficiently wet and warm conditions. If the spore is on a wet leaf, it will
germinate at a temperature-dependent speed, and the germ tube penetrates into
the leaf where it is safe from preventive fungicides. So weather-dependent timing
of fungus treatment is essential for success: too early fungicide sprays will not yet
be on newly-formed leaves, and will have eroded from the older leaves - and too
late application is just environment-harmful wastage.

Apple scab Venturia inaequalis is a fungus causing apple scab. In winter, its ascospores and
conidia hibernate on dead leaves. Spore ripening depends on temperature. At a
central laboratory, regularly selected dead leaves are checked under the micro-
scope. When ripening hibernated spores is observed, farmers are warned by way
of public media. Particularly if farmers had scab problems last year, and therefore
probably have ascospores in their orchards, they are advised to carry out a preven-
tive spray, at the latest just before trees begin to show green buds.
In the early spring, weather-dependent problems for spraying are
insufficient vehicle access to orchards, and the fact that this season may be
windy. For both, medium-range agrometeorological weather forecasts are useful.
After spores have ripened in spring, actual occurrence of infection requires leaf

128
CHAPTER 8 - OPERATIONAL AGROMETEOROLOGY

wetness for a sufficient number of hours at a sufficiently high temperature. In

order to decide if the weather has been sufficiently bad, Mills' scab criteria are
widely used (Machardy and Gadourny, 1989). Mills' published criteria tables can
be summarized by the Scharringa-formula:
(continous leaf wetness in hours) x (temperature in °C) >140.
Ripe scab spores are discharged during rain and are then distributed by the
wind. Cold and dry weather may retard or prevent an outbreak. Later in the
season, either in persistently humid weather or after sprinkling actions (e.g.
against frost), farmers are advised by radio that another application of a scab spray
is required - more so if heavy rain (>30 mm) has removed fungicide from trees.
Ascospore production ceases quite late.
In this fungal disease, as in many others, the occurrence of leaf wetness for
a sufficient time is such an essential condition for outbreaks that WMO-TN No. 55
(1963) devoted much space to reviewing the known methods to observe leaf
wetness. More than a dozen different methods were described, which implied -
as usual in scientific matters - that none of the methods existing in 1963 was
satisfactory. As of 1997, one major reason why weather services do not observe leaf
wetness routinely is that there is still no existing measuring method that is
considered reliable (see section 9.2.6). Advisory work often uses relative humidity
and/or precipitation as a substitute working parameter.

Potato blight Phytophthora is a fungus family which was responsible for the destruction of the
Irish potato crop around 1845. One member causes fruit rot in strawberries, and
indeed the fruits are called 'straw' berries, because covering soil with straw
prevents fungus spores on the ground reaching the plants by rain splash.
In particular, Phytophthora infestans causes potato late blight. Blight spores,
which have been deposed by wind on potato leaves, are vulnerable to direct
sunshine. It was found that spores germinated on wet leaves after a humid night
with a minimum temperature at or above 10°C, followed by a cloudy day. From
this Dutch research various national warning schemes were evolved to warn
farmers that they should carry out preventive treatment for potato blight.
Schemes vary according to local climate and preferred wetness information input,
using observations of either relative humidity or precipitation, or even weather
maps (De Weille, 1964). In those Central African areas where residual soil
moisture is available, potatoes are grown in the dry season in order to minimize
the risk of late blight attack (Haverkort, 1987). For tropical farmers, improved
prediction of the start and the end of rainy seasons (and of dry spells) would be
very valuable.

Insect pests The need for meteorological forecasts of optimal data for crop protection action is
not restricted to fungal diseases. In section 7.5.2 we saw how the development of
locust swarms is dependent on weather conditions. Another insect pest which can
only be controlled by close cooperation between meteorologists and entomolo-
gists is the codling moth (Laspeyersia pomonella) which spends its entire larval life
inside an apple. Pesticides can only reach them in the few days when they hatch
from eggs and eat their way into the fruit. Ovulation and hatching appears to be
reasonably predictable from the catch of male adults in pheromone traps, and
can, therefore, be monitored from early spring onwards. When the catch becomes
significant, 'biofix' day is declared for the area around the station, as starting day
for accumulating degree-days (DT) with base temperature 10°C and ceiling
temperature 32°C. First pesticide spray alert is at 1:DT :::::640 from biofix - then
field scouting and pheromone trap catches are used to decide, when and how
further spraying should take place.
The life cycle of many pests cannot be followed usefully by 24-hour average
screen temperature sums only. Development of the cabbage root fly (Delia
radicum) depends in the larval stage on soil temperatures at depths of 5 to 10 cm,
which are not often measured at weather stations. Soil temperature then must be
estimated from screen maximum and minimum temperatures by way of good soil

129
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

heat flux modelling. Moreover, because the fly population is a mixture of early-
emerging and late-emerging varieties, no particular starting date for a simple
temperature summation can be given.
Delia radicum has three or four generations in a year. To forecast the timing
of new fly attacks, Collier, et al. (1991) developed a quite extensive model which
carefully simulated the temperature-related development of larva, pupa and adult
stages. No assumption of linear temperature dependence was made, because that
.does not apply for insect development at low or high temperatures (Baker, 1980).
The above-mentioned examples dealt with the interrelation between
weather and pest. However, in crop protection methodology the weather-depend-
ent development of the host plant may well determine if the pest has a good or a
bad chance of doing damage. For example, some pest larvae can only eat very
young leaves, and if the host plant has been developing very fast or very slowly,
there may be no young leaves at the time of larva emergence.
Additionally, weather dependent development of predators of the pests must
not be forgotten. If the season has been favourable for aphids, crop damage may
be limited if, at the same time, the ladybird population increased because of good
weather. And in wet warm winters, fungus may thrive, but this implies that many
insect pests will die from fungal diseases (Harrington and Woiwod, 1995). Weather
always exerts its influence on an entire situation, not only on the single process
which an investigator is looking at.

8.4 AGROCLIMAT- The necessity for close working cooperation between meteorologists on the one
OLOGICAL SURVEYS hand, and the 'user interest' and those in other disciplines on the other, is implicit
in a diagram by Duckham (1974), which highlights the interrelations between
climate and food. This is reproduced as Figure 8.4 and shows the various inputs
into an agricultural system and the sequence of events.
Some valuable agroclimatological analysis projects have been undertaken
jointly by two or more international organizations such as FAO, UNESCO and

Figure 8.4 WORLD CLIMATIC SYSTEMS

Climate and food:
interrelationships between
PARENT
climate and associated factors TOPOGRAPHY
ROCKS
- feasible farming systems,
economic development, FARMING DISTRICT
population and human diets
(after Duckham, 1974). CLIMATE - - - - - - SOIL

l ~
i
VARIANCE
NAJRAL
VEGETATION

T PESTS
UNCERTAINTY
DISEASES
A

FEASIBLE FARM SYSTEMS

EXTERNAL
ECONOMIC
INPUTS,
DEVELOPMENT
SERVICES KIND, LEVEL
AND VARIANCE
P = mean annual precipitation;
OF FOOD
T = mean potential transpiration in OUTPUT
the thermal growing season; POPULATION
HUMAN
PRESSURE
A = mean effective transpiration in DIETS
AND INCOME
that season.

130
CHAPTER 8 - OPERATIONAL AGROMETEOROLOGY

WMO. Several examples are reported in WMO Technical Notes Nos. 56, 86, 125,
161 and 179 (see Appendix IV for full titles).
Such analyses must integrate biological and meteorological data - the
thresholds, limits, and variability of the latter being formulated in terms of the
biology of the plant or animal concerned. However, the 'meteorology' of different
species, insofar as it is known at all, differs from species to species. So the classic
practice of attempting an overall classification of climate of a place by means of
'climate analogues' in the style of Koppen - which is based only upon monthly
means of temperature and precipitation - has given way to the preparation of a
real inventory of local climates. The difficulty of presenting the mass of data in a
form which can be readily appreciated therefore remains.

8.4.1 Agroclimatology of the Figure 8.5, from WMO-TN No. 86 (Cocheme and Franquin, 1967), shows some of many
Sahel - an example of ingenious plotting and graphic techniques, which have been developed to present
presentation various important parameter averages simultaneously. More presentation forms are
presented in CL-G. Some of the abbreviations and symbols used require explanation.

Figure 8.5 FAO/UNESCO/WMO Place: Zinder Country: Niger

AGROCLIMATOLOGY SURVEY Lat.: 15°48 Long.: 09°00 Alt.: 480m
Specimen diagrammatic ,---,--~--,---,--~--,--,,..,...-r-,-r----,r~-,-,---r----,---.--~mm
summary of some main Rad ation ____ .. - - - - - - - - - "T • -,- - - - - • - ~ : -15
agroclimatic charac- Rad out -
--- I
1
I
1
I
1
I
1 1
j----- - - •, - - - - - •
I I j I I -10
teristics of a reference Global rad. --f--t---r--t--t--4~'-1--f-"'::t--t--4--_.J__
station in Niger (WMO-TN ·- -·-····-·- -·-:· • .J.. ______ 1,..1 •.L. - - 5
Net rad.-·-·-·•·- 1 : : : : •-■•- •-■•-■•-·-·-·-··
No. 86, p. 122).
-+--__,.--+---+-__,.~-+---+--i-:+--;.-:--+--+:-+:-i:~---+---+----■1----1- 0
40 - Terrperatu·es
36 - 1
v~ _,,._.,_
~-

, - '•
-r-,....._ !
L -......., I
• -. • I N.....__
1 1 1 1
I I I,.----..
I _J..,i'"'...- .--............
32
28-- ~ ~ ax. --- , • - •
_,, .. - ,,,.,· , • • - ' - ' - •-■ •■ • ,. I
_______..... .,.~~-.-- ....l,.--•-•,11, l .,1 • - • • - •-. r----...-.
..... l-a-
Cay-· ,·--~----- - ...,_i-J •-.:..:.7"_:.: ... ,_: _____ .,., . ....... t-
24-: oc ~~ean-•'•--t,_....---i---" --,-io-■...J._-•--- • •-, I - - ... ,'""• .... ,-■~••

20 ~ ight- r - ,_,,,. ,1 ,1 ,1 ,1 ,1 ~ - ... ... ... - ..

16 _ ~ I , .. _.
~~in . .,,,,- I I I I I r--,.....___.
12 - I I I I I
8 - 1.l__J_21oodd jl.._i.l mm
I I I I I
~r
4 -
... 800
~
i.L...! 19,lo dd-11 1 1
0 _.,__ __,.~--+---+---+----+----+-----+--.i---+-----+--+......+----+---+~
400 - Rair !fall (l' 31-19 ~O) : : e--1:: ~
700
360 -
320 -
I
I
I
I
I
I
L-1
-I

I
I I
I I
I I
,__

5451
►

►
~
600
500
280 - ...
240 -
mm
ILJ_
I I !
-I
~! I I
11 ,__
400
200 -
160 -
120 - cr
a= lU
p= 0 ..22 ' -
'
I
I~
I ~
~ I :'
-I
I
'
L-L.L.
11
I I
,__
►

►
~
300
200
::e--;;-; -
-
80-
4
~~-L_l_J__L_J:=diiiit~[IJ__f~r2""~b;-;-;4i,_JlJLJ_I
Water bud Jet
1
►
100
0
225 _
200-
175 - ,
,, ----··---. rwr '\!'it
', I
~.
I
I
I
I
I I
II

150 - , ',• : ' I, I : : • , - '• '

:;:~ m:/1 -E-- ..1924

T =
E
T/2
-·-·•rr; :'••-• :~,r i
• I I •I
'•. •-- •••
75 - p = 548 I ""' ._,,... I ~ - •
WD = 1480 : ,1 : (~...,ci-"~ ;•.
50 -
_
WS = 104
EP = 482 E
~
r
I I , I I\
1 , 2 d
1 r!
1 1
25 1
T/ 0.'" " .. "' • ' 1 10 d \..
0 -+-----1--4-----"■ -..:.P.....,.-=:·-_,,,-:;....4,-_-4-__~--'---•-+--......,_,,.:_,:.-1-........__...;::a..:..P-+----+---1-■-..- m s-1
mb I I I I I ~16
35 - , • ., • • - ' -. I I I I I - 14
30- "----.,,_•• ••,_.---.L.•-1 I I I -12
25 - V-- ■--- , ... _ _ ■---" - - - '•. : ~ - ........... : : : ; : . • : ~ ••: : : . ~ , - - -10
20 - Sat. Def • • • • 't.I , 1I ... r - 1- r■ ... • • - •• - - 8
I I~•
15- I •,, 1,1•1 '-- 6
10-
5
_ Win[j spee~ and iaturaton def cit 11 , 11 I I ••- .. -•r
... 2 I I - 4

o-~·~·~-·-~·-~··~~·~·~-·-~·-~··~-·~·-'~·-l~,~·-·~'-w''-·-1~,-~,-~,,~-~··--·-----·o
1020 1020 1020 1020 1020 1020 1020 1020 1020 1020 1020 1020 1020 1020
Dec. Jan. Feb. Mar. Apr. May Jun. Jul. Aug. Sep. Oct. Nov. Dec. Jan.

131
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

Radiation: Top line - extraterrestrial ('outside') daily totals;

Middle line -global radiation received daily at Earth's surface
Bottom line - net radiation at surface daily.
(All are expressed in mm of evaporable water per day.)
1 mm depth of water= 2.47 x 1Q6 J m-2.
Temperatures: Maximum, mean and minimum; and effective day and night temperature (i.e. temperature
during the daylight and night hours respectively).
Below are given temperature sums, in degree-days accumulated during the moist period
(plus storage), and during the first intermediate plus humid period.
Rainfall: Monthly variability histogram, based on a 30-year record (1931-1960). Successive horizontal
lines (working downward) relate to the maximum monthly amount and that equalled or
exceeded in 6, 12, 18, 24 and 30 years out of the total. (On the right are given yearly amounts) .
All heights are given in mm.
cr = standard deviation of annual rainfall;
(cr/P) x 100 = coefficient of variation (P = 30-year average rainfall).
Water budget: Er = mean potential ·evapotranspiration (mm of water per month);
Ertz= SO per cent potential evapotranspiration (in mm of water);
Ertrn = 10 per cent potential evapotranspiration (in mm of water).
(Br is calculated here by a variation of the Penman method.)
The period when rainfall >E-r is called the humid period (H) and defined by the intercept
HH (50 d = days). In moist (M) period, rainfall >E-r12 ; this period can be extended at end-
season, if supplies (S) stored in the soil can be drawn upon.
WD (water deficit) and WS (water surplus) are respectively the excess of Br over rainfall; and
the excess of rainfall over Er for the months exhibiting these inequalities.
EP is the effective precipitation, namely the mean amount of water deemed to be available
to crops, calculated as:
EP = WS minus (evaporation loss from the ground+ runoff).
Wind speed Wind speed V (broken line, in m s- 1) represents the sum of three daily readings (why?) .
and saturation The saturation deficit (broken line with dots, in mb) is the mean of three daily readings.
deficit:

8.4.2 Irrigation need - a Hogg (1968) investigated the amount of irrigation water needed in England and
climatological case study Wales, if the growth and development of a range of plant species were to take
place uninterrupted by any water shortage (see section 6.5 on the implications of
this working hypothesis).
It was postulated that a plant's progress would be checked, unless evapo-
transpiration could take place at the 'potential' rate. This implied that the soil
moisture deficit (SMD) should not exceed certain stated levels, i.e. the moisture
,content should not fall more than a stated amount below that of 'field capacity'.
In an ideal situation, continuous monitoring and immediate application of water
would maintain the soil moisture at the optimal level. In practice, the moisture
status is assessed at intervals (a period of 15 days was finally adopted), the SMD
being allowed to build up to some stated limits, and irrigation then applied to
restore the soil to some predetermined SMD. The limits are based upon judge-
ments as to the depth of the rooting system.
It was decided to analyse four irrigation 'strategies':
1. Allow SMD to reach 2.5 cm and then irrigate to field capacity;
2. Allow SMD to reach 5 cm and then irrigate to field capacity;
3. Allow SMD to reach 7.5 cm and then irrigate to 2.5 cm SMD;
4. Allow SMD to reach 12.5 cm and then irrigate to 5 cm SMD.
Irrigation in plans 3 and 4 would be practically difficult to implement in the field.
Plan 1 is appropriate to shallow-rooted crops, and plan 2 where deeper roots
already existed or had developed. Plans 3 and 4 were based on the hypothesis that
deep-rooted plants (e.g. wheat, sugar-beet, established fruit bushes or fruit trees)
would be able to tap water supplies from considerable depths (say 50-100 cm or
more) and would not be completely dependent on moisture in the upper layers. If
a plant can grow unhindered when SMD of 7.5 cm or 12.5 cm occur, then it is
obviously not essential to irrigate up to capacity, hence the basis of plans 3 and 4.

132
CHAPTER 8 - OPERATIONAL AGROMETEOROLOGY

A further benefit in these latter cases is that rain which falls just following irriga-
tion will not be 'lost' in runoff or drainage (as in plans 1 and 2), but can be stored
in partially depleted upper soil layers.
The water needs for a given plant will vary with its growth and development
and different irrigation schedules will therefore be appropriate at different stages.
In the British Isles, the best procedure is to assume that the growing season
extends from 1 April to 30 September and to allow for within-season variability
(including different planting dates) in order to compute irrigation need for periods
of 2, 3, 4, 5, and 6 months:

April-May; May-June; June-July; July-August; August-September;

April-June; May-July; June-August; July-September;
April-July; May-August; June-September;
April-August; May-September;
April-September.

In all cases it is assumed that the soil is at capacity at the commencement of

each of the monthly sequences.
Computations were carried out for a 20-year period (1930-1950) for each of
the four plans and each of the periods, and the result was demonstrated in five
different maps:
(i) The total amount of water required in the 20-year period (see Figure 8.6 for
an example of this);
(ii) The adjusted total, taking the need in the five highest years equal to that of
the fifth highest (i.e. total=.'£ for lowest 15 years+ 5 x that in the 16th year);
(iii) Need in driest year;
(iv) Need in fifth driest year;
(v) Number of years within a 20-year period when no irrigation was needed.
Type (ii) and more obviously (iii) and (iv) are designed to take account of
year-to-year variability. Type (iii) indicates the amount of water needed in the
extreme case, while type (ii) recognizes that, in practice, it is seldom prudent to
invest in equipment to cover the rarest of situations.

6°W sw
0
4°W 3°w 2°w 1°w oo l°E 2°E
Figure 8.6 56°N 56°N
Theoretically required
irrigation water needed in
average years (case (i)), when
for a 6-month season (April to 55°N 55°N

September) irrigation to field

capacity is carried out as soon
as SMD reaches 2.5 cm.
54°N 54°N

133
WMO LECTIJRE NOTES - AGRICULTIJRAL METEOROLOGY

It should be recognized that this study assumes homogeneous conditions in

Britain in all non-considered aspects such as temperature and soil, and that water
requirements are the only important aspect for growing a particular crop.

8.4.3 Agroclimatological The differences in crop responses to soil moisture for different climates were
analysis of a rainfed semi-arid discussed in Chapter 4 and illustrated in Figure 4.2, and the point was made that
situation there is generally a good correlation between rainfall and crop yield, in dry
climates where water shortage is such a dominant factor. However, the study
below shows that moisture matters can be slightly more complicated, even in
semi-arid and arid climates.
In the Indian state of Maharashtra, just east of Mumbai (Bombay), the
overall rainfall is not much less than in Britain, but its interannual variability is
larger: at Dhule, annual precipitation is 537 ± 280 mm (Sarker and Biswas, 1986).
Sorghum grain requires 340 to 600 mm of water per seasonal crop, and while
average rainfall could meet this demand, the variability also causes large fluctua-
tions in sorghum yield, more than ± 60 per cent.
However, in a study of precipitation and yield from 1967 to 1990 (Deosthali
and Lomas, 1996) it was found that rainfall in the growing season accounted for
only 4 to 7 per cent of interannual yield variability! This surprising conclusion
proved to result from an over-simplified analysis. Analysing rainfall effects on
crops per week instead of per season, using multiple regression or Fisher's orthog-
onal polynomial method (1924), it appeared that rainfall effects varied greatly
during the crop's lifecycle.
During the first six weeks of the cycle - the vegetative stage - the potential
water demand was calculated at 154 mm, while average rainfall was 270 mm in
that period. This means that above-average rainfall then has a negative effect on
sorghum yields. During the reproductive period of the next 3 weeks, the demand
is 114 mm and average rainfall is 86 mm, and yield shows little response to rain-
fall over that period. The last six weeks of grain filling and maturity require
208 mm for transpiration, while average rainfall supply is only 17 6 mm, with high
variability - so additional rainfall in that final period has a very beneficial effect
on sorghum yield. In summary, the time distribution of rainfall is at least as
important as the annual total.
The Maharashtra study provided a few other interesting crop-weather links.
High maximum temperatures at the end of the grain filling period had a negative
effect on yields, presumably because they shorten the duration of the grain filling
period. Early in the vegetative period, high maximum temperatures slow down
germination; high minimum temperatures in the same period enhance leaf area
development, and so have a positive effect on yield. Sunshine duration does not
·limit sorghum production - as would be expected in semi-arid tropics.
A significant negative correlation was found between yields and average
wind speed, probably because strong winds increase evapotranspiration. It may be
worthwhile to investigate wind protection measures.

8.s COMPUTER Models are developed to understand and handle situations for which we do not
WEATHER MODELLING have enough information. They use available information by adding a set of
FOR AGRICULTURE assumptions about the considered situation. Some assumptions will be uninten-
tional, or signified by 'it is self-evident that', or hidden in some commonly-used
equation, or derived from a limited dataset. A model is valid if it gives accurate
results when its initial assumptions are correct.
Model users, then, should check if model assumptions are appropriate for
their situation. If a model has been developed from data obtained in one particu-
lar region, they should consider to what extent the characteristics of the data region
(e.g. orography, humidity, temperature range, season) could influence an obtained
model. If such influence appears significant, the next question is if the user's region
differs very much. In particular, such a check is needed when the proposed model
is descriptive, i.e. summarizes data into formulas, usually by way of regression
. calculations. Such a descriptive model must not be extrapolated beyond the situa-
tion covered by the data which have been used for model development.

134
CHAPTER 8 - OPERATIONAL AGROMETEOROLOGY

Explanatory models, on the other hand, are based on verified 'laws' of nature
for the considered biological and/or physical processes. From environmental
factors such as radiation, humidity and crop status they deduce the behaviour of
a system and its development rate (Baier, 1973). Some of the natural 'laws' that are
used may describe an imperfectly understood or not. exactly describable process,
and will then contain a descriptive component. Various explanatory models are
very good as they include many of the possibly relevant processes, but are also so
complex that they cannot be used except by the model creators and by very
advanced research teams. In that case, summary models tend to evolve from these
comprehensive explanatory models.
The performance of available models is increasing with time, particularly
those developed by interdisciplinary groups. However, if models are to be practi-
cally used, the following requirements must be met:
(1) The model programme must be usable by the people who are expected to
benefit from it. If its actual use requires either a great deal of computer expe-
rience, or advanced academic knowledge of agricultural or meteorological
science, then only a relatively small number of users will be able to apply it
to their advantage;
(2) The programme must require only those data that users will be able to
acquire at a level of effort and expense which is appropriate for the benefit
expected from model application;
(3) The model must provide information of acceptable accuracy, easily applica-
ble to the class of problems or decisions for which it is intended.
Unless the models complies with these conditions:, the agrometeorology user may
be practically better off with good climatological maps and tables, even if he then
has to do without the status of solving his problems by computer model.
Regression and correlation agrometeorological models can give good results,
when crop growth is dominated by a single weather factor: net water balance in
subtropical regions, radiation and freezing temperatures in temperate climates.
Classification of details according to season, or local factors (fertility, soil state),
may improve model performance. All the same, models of this type are limited to
the range of the original database, and the assumption of linear interdependence
for regressed variables is a weak point.
A more fundamental objection against many regression-type models is that
they are purely diagnostic, describing present developments, but are not designed
to use prognostic weather information. For tactical usefulness, a model must have
input parameters which are either local constants or weather which can be forecast.
An example of a prognostic model approach is given by the harvest drying
models presented by Atzema (1993, 1995). Their input .parameters are five-day
forecasts of surface temperature, dewpoint, cloudiness (to estimate radiation, see
section 2.1.2) and precipitation. The procedure follows an evaporation model
along the lines of the one contained in section 4.4, operationalized to the point
that it can be handled by a private weather bureau. Output is the development of
dry matter content of grass after mowing (Figure 8.7), enabling the farmer to
Figure 8.7 50%
dry
Modelled dry matter content of matter
cut grass on a Dutch farm. content
40% - ..! - -
Based on the weather forecast I

of 8 May, it was estimated that

a dry matter content of 40 per 30% - ..! - -
I

cent would be reached on

11 May at 1400 if grass were 20% --r- -7-- -,-
mowed (M) on 8 May,
assuming some additional
_J __
10% --L- --L- - ....J_ - - .l -
tedding (T) and windrowing I
- .J. - -
I I I I

(W) of the cut grass. Resulting N-1I T'T

I
'W
I
I

advice: do not mow later than

9May. Hour 0 12 0 12 0 12 0 12 0 12 0 12 0
Date 8 9 10 11 12 13

135
WMO LECTIJRE NOTES - AGRICULTIJRAL METEOROLOGY

decide if in the coming days it is worthwhile to start haymaking. The prognosis

shows, that making silage (dry matter >35 per cent) is feasible when starting on
day 8 or 9, but that making hay (dry matter >70 per cent) is not likely to be possi-
ble right now.

8.6 MODELLING OF All plants grow and produce within a range of temperatures. The metabolic
HEAT STRESS FOR processes responsible for organized growth cannot function outside these limits.
AVOCADO - A CASE Most processes achieve a maximum r_a te at a temperature towards the upper end
STUDY of the range~ i.e. at about 25°-30°C (temperate crops) or 30°-35°C (tropical crops).
During excessively high temperatures, desiccation is not the only damage to
crops, but flowering and fruit set also suffer from heat. One example is Helianthus
tuberosus, a member of the artichoke family, which does not flower at all if temper-
atures are steadily above 27°C.
During excessive heat stress, avocado pear trees may produce many flowers,
but thes.e gradually are sterilized as temperatures rise above 33°C, and when they
exceed 39°C pollen becomes inactive. The heat stress effect is amplified when the
relative humidity is below 15 per cent. In the Mediterranean region such condi-
tions are possible, but fortunately not too frequent (Figure 8.8).
Figure 8.8 12
Frequency distribution of v, 10
maximum temperature (front >..
to rear) and minimum relative
humidity (at one of the three
daily observations - left to
-v
~
0

.0
8

6
right) for Bet-Dagan, Israel, E 4
:::J:
between April and May z 2
1973-1983. Shaded sections are
heat stress conditions (Lomas, 0
1988).

The effects of high temperature, its duration and its timing related to crop
phenology, were modelled jointly into a single climatological heat stress unit
(Lomas, 1988). Such a simplified representation can be useful in assessing the
climatological suitability of specific agricultural regions for growing avocados. It
can also be used as an indicator for irrigation practice for the reduction of occur-
ring heat stress damage. Also, early assessments of expected avocado yield can be
made for the agro-industry. Finally, it was used to compare heat stress sensitivity
of different avocado varieties (Lomas, 1992).
(1) The critical threshold value of maximum temperature was experimentally
reported to be 33°C. Increasing maximum temperatures proved to have
damaging impact on the fertilization process and the degree of pollen
sterility. Maximum temperatures exceeding 42°C were considered to be cata-
strophic events. A relative weighting factor was introduced, linearly
interpolated between the experimental threshold values 33°C (factor 1.0)
and 43°C (factor 4.5).
(2) A long duration of excessively high temperatures can increase the amount of
inflorescence sterility. Lomas and Shashoua (1973) showed that the longer
the period of hot and dry days, the lower the wheat yield in Israel. However,
the effect of the duration of heat spells on avocado yields had not yet been
quantified.

136
CHAPTER 8 - OPERATIONAL AGROMETEOROLOGY

Therefore, the duration of a heat spell was defined as the number of days
(n) that the daily maximum temperature was over 33°C. Sterility may
become a function of the percentage of open florets x the number of days
that temperatures are over 33°C, but especially over 36°C. A series of empir-
ical weights for heat spell duration, with exponentially increasing value, was
developed through trial analysis by computer. The finally chosen weights
were those that gave the highest correlations with yields. For a temperature
range of 33°-35.9°C the relative spell duration weighting factor was (n-1), for
the 36°-38.9°C range it was 1.4 n, and 2.0 n for the 39°-43°C range.
(3) The timing of occurrence of heat stress in relation to the flowering process of
the avocado pear is also an essential factor in the resulting damage. Avocados
begin flowering in Israel in March and complete the flowering process in
May. The duration of the flowering period is temperature-dependent: in the
coastal plain, the flowering process lasts about 70 days, whereas in the inte-
rior valleys it is completed in about 35 days. The flowering process starts
relatively slowly due to low temperatures early in the season. The percentage
distribution of open flowers follows a near-normal distribution over the
entire flowering period, peaking a few days after the period centre.
The effect of heat stress on sterility relates to the percentage of florets open
at the time of heat stress. Heat-caused fruit drop was reported to occur also in early
June, so a flowering and early fruit period of 80 days duration was considered. A
weighting factor was apportioned to follow the distribution curve of open flowers.
At bud swelling and bud burst during March, weighting factors gradually increased
from 0.1 to 0.4, and reached 0.9 at the end of April when the percentage of open
flowers approaches its peak. At 1.3, weighting factors are highest by mid-May
when early fruit set occurs. Then the weighting factors decrease rapidly to 0.3 at
the end of June.
The Heat Stress Index (HSI) was the weighted product of the above-named
three components. Daily values of the HSI were compared with 19 years of histor-
ical avocado data (1965-1984) on the whole flowering and early fruit set period
and fruit yields. The data were obtained from a well-run plantation, located about
5 km inland from the coast, and where agrotechnology had not changed in
20 years. Climatic data were taken from an agrometeorological station which was
some 60 km distant, but which had an excellently consistent observation history
and which was also about 6 km inland.
Avocado producers claimed that possible heat stress effects were, to some
extent, predetermined by temperatures of the preceding winter. Therefore, from
the whole 19-year data period six years were selected with more or less the same
intensities of heat stress, and of their preceding winters (December to February),
and the mean minimum temperatures were compared with yield. In this small
sample, these effects proved insignificant.
Two non-modelled climatic factors which might influence the predisposition
of avocado trees to heat stress are winter temperatures and rainfall (soil moisture).
In this case, variable rainfall during the winter is supplemented by irrigation, thus
soil moisture is unlikely to be a predisposing factor. The number of flowers (0.8 to
1.5 million per avocado tree, with 300 fruit trees per hectare) is unlikely to limit
plantation yields. But mean minimum winter temperatures will affect both the
timing and duration of the flowering process, and not only the stage at which
fruitlets are exposed to heat stress, but also the number of fruitlets which are
exposed.
In a follow-up investigation (Lomas and Zamet, 1994) it was found that
though the number of single nights where the temperature fell below 10°C
showed no significant correlation with yield, occurrence of 5-day periods with
such low minima were clearly related to avocado yields.
The original HSI-model thus allows for improvement. Even so, the original
unimproved model is simple and accounts for some 72 per cent of the interannual
variability in yield during frost-free years. Also, it provides 3 months of lead time
for making a yield assessment and consequent decisions on agrotechnology and
on commercial matters.

137
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

8.7 AGROMETEOR- It will be obvious that meteorological expertise and advice can make a valuable
OLOGICAL WEATHER AND contribution to decisions in agricultural management and practice, both on
CLIMATE INFORMATION climatological-strategic matters and on actual weather application for tactical decisions.
However, this cannot be produced by standardized weather service activities,
References: Price-Budgen
notwithstanding their usefulness for non-agricultural purposes such as aviation.
(1990), WMO-No. 134, CL-G
Too often the agricultural community has to ask for raw weather data series,
(Annex A9), Wieringa (1996).
to be analysed by agrospecialists without much real meteorological insight,
because the weather service is not providing the minimal locally necessary
amount of agroclimatological information.
Some guidance on the type of information needed is given in Table 8.1,
taken from the WMO Guide to Agrometeorological Practices. Professional advice
should be prepared with respect to representativity of these and other basic data
for local application. Table 8.1 has not been supplemented with application-
directed parameters, such as temperature sums and risk indices, as their
importance varies with locality.
Not all of these items are necessary everywhere, but the least a weather
service should do is to find out which of these parameters are most pertinent to
local agriculture at the local level of sophistication, and to take care that these can
be made accessible in a timely and economically acceptable way. This may require
making some basic or summarized data accessible by telephone, fax or e-mail. In
summary, the weather service should know and care about more operational
application aspects of their data than is needed for synoptics.
Table 8.1 (a) Air temperature
Presentation of agroclimatic (1) Temperature probabilities
information (based on (2) Chilling hours
section 3.2.4 of the Guide to (3) Degree days
Agricultural Meteorological (4) Hours or days above or below selected temperatures
Practices, WMO-No. 134, 1981) (5) Interdiurnal variability
(6) Maximum and minimum temperature statistics (e.g. frequency of
minima <T)
(7) Growing season statistics
(8) Frost risk
(b) Precipitation
(1) Probability of specified amount during a period
(2) Number of days with specified amounts of precipitation
(3) Probabilities of thundershowers; hail
(4) Duration and amount of snow cover
(5) Date of beginning and end of snow cover
(6) Probability of extreme precipitation amounts
(c) Wind
(1) Wind rose (frequency distribution per direction sector of 30° or 45° width)
(2) Average wind speed (hourly, daily)
(3) Maximum wind (average and gust)
(4) Diurnal variation
(5) Hours of wind less than selected speed
(d) Sky cover, sunshine, radiation
(1) Percentage of possible sunshine
(2) Number of clear, partly cloudy, cloudy days
(3) Amounts of global and downward IR radiation
(e) Humidity
(1) Probability of specified relative humidity
(2) Duration of specified threshold exceedance of humidity
(f) Free water evaporation
(1) Total amount
(2) Diurnal variation of evaporation
(3) Relative dryness of air
(4) Evapotranspiration

138
CHAPTER 8 - OPERATIONAL AGROMETEOROLOGY

(g) Dew
(1) Duration and amount of dew
(2) Diurnal variation of dew
(3) Association of dew with vegetative wetting
(4) Probability of dew formation with season
(h) Soil temperature
(1) Mean and standard deviation at standard depth
(2) Depth of frost penetration
(3) Probability of occurrence of specified temperatures at standard depths
(4) Dates when threshold values of temperature (germination, vegetation)
are reached
(j) Soil moisture
(1) Mean value at standard depth.

Review questions on Chapter 8

A. Questions on basic understanding:

8.1 Many biological phenomena (including pests and diseases) are strongly
influenced by the simultaneous occurrence of at least two climatic
elements. Such joint data may be graphically displayed by climagraphs
(i.e two-dimensional diagrams in which one element, e.g. temperature,
is plotted along, say, the x-axis, and another, e.g. humidity, along the y-
axis. The diagram is completed by joining up the points in sequence.
Construct such a diagram from the data in Table 8.2.

Table 8.2
Station A Station B
Month Precipitation (cm) T(°C) Precipitation (cm) T(°C)

January 13.4 26 5.1 7

February 14.5 27 3.8 4
March 19.6 27 3.5 7
April 19.6 27 4.6 9
May 15.7 26 4.6 12
June 0.8 24 4.1 16
July 0.3 23 5.1 18
August 0.3 23 5.6 17
September 3.1 25 4.6 15
October 11.9 26 5.8 11
November 22.0 26 6.4 7
December 14.2 25 5.1 14

8.2 Examine and discuss the statement: " ... the limit for regular dry-land
farming (in the semi-arid and arid areas of the Near East) is determined
by the amount of rainfall and its variability and possibilities of storage,
rather than by the potential evapotranspiration." (WMO-TN No. 56,
p. 50).
8.3 An index of continentality (C) has been given (expressed in per cent) as:

C = 1. 3 A - 36.3
sin <I>

A = average difference between mean maximum of the hottest month of

the year and the mean minimum of the coldest month; <I> = latitude.
Compute C for your region and map the results. Examine the pattern of
isopleths and indicate areas, where particular features of the pattern do
identify and/or explain characteristics of the local climate, e.g. transition
from a maritime to an inland climate.

139
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

B. Questions on extended application:

8.4 Especially in areas with low annual rainfall, the year-to-year (also, at
times, the month-to-month) variability is important. This variability can
be expressed by the so-called Inter-period Variability [l/(n-1)] Llrn-rn_ 1 1,
viz., the average value of differences between successive amounts, ignor-
ing sign.
The Relative Inter-annual Variability (Rel IaV) is the ratio of the inter-
annual variability to the mean annual rainfall (expressed as a
percentage), i.e:
1
- L lrn - rn-11
Rel IaV =n- l x 100
Average rainfall

Calculate the Rel IaV from the following annual rainfall data (in mm) for
Ankara (Turkey) in Table 8.3.
Table 8.3 Year 1920 1 2 3 4 5 6 7 8 9
Rainfall n/a 272 279 218 337
Year 1930 1 2 3 4 5 6 7 8 9
Rainfall 286 423 270 338 259 388 394 365 501 367
Year 1940 1 2 3 4 5 6 7 8 9
Rainfall 397 335 363 398 268 276 387 423 365 346
Year 1950 1 2 3 4 5 6 7 8 9
Rainfall 264 411 363 334 396 349 250 421 405 333
Year 1960
Rainfall 362 n/a

C. Questions involving practical work or local data:

8.5 The ratio Q*/(LR) , Budyko's 'Radiational Index of Dryness' (see Lowry,
1972-IV, p. 40), is roughly the ratio of net energy annually received
at the surface to the energy required for the evaporation of annual rain-
fall (R). In arid regions there are extensive areas with the index >3; in
humid areas (of Western Europe and parts of Canada) the value is
typical '£2/3.
Obtain values of Q* and R for your region. Examine the relationship
between the-index and types of land-use (see Question 4.9).
8.6 In WMO-TN No. 56, p. 15, the limit of dry-land farming in the semi-arid
and arid region of the Near East is associated with mean annual rainfall
and Rel IaV.
If x = mean annual rainfall (in mm) and y = Rel Ia V then successful dry-
land farming in 8 years out of 10 is possible in areas where:
x >180 mm and y <6.15 x 10-2 x + 23 (%)
Analyse the distribution of cereal crops in a dry region in the light of the
above criteria.
8. 7 As a long-term project, construct - for your area - a diagrammatic
summary of climate on the lines of Figure 8.5.
8.8 The spread of potato blight requires, among other conditions, the
persistence of a film of moisture on the leaves for several hours. This has
been linked with a RH of 90 per cent, as registered by a hygrograph
placed amongst the crop foliage.
Utilizing basic data from the standard climatological network it has been
found (in the United States) that, if:
y = number of consecutive hours with RH of 90 per cent within the crop;
x = number of consecutive hours in the screen with:
[T(dry-bulb) -T(dewpoint)] ~1.5 K
then: y = 0.68 x + 4.01.

140
CHAPTER 8 - OPERATIONAL AGROMETEOROLOGY

(Note: The 1.5 K deficit implies a RH of about 85 per cent).

At a selection of agrometeorological stations, ·install hygrographs within
crop stands and investigate the relationships between humidity condi-
tions within the crop and at the standard site.
8.9 A series of generalizations grouped under Hopkins' Bioclimatic Law
include the following statement: "Other conditions being equal, the
variation in time of occurrence of a given periodic event in life activity
in temperate North America is at the general average rate of four days to
each degree of latitude, 5 degrees of longitude and 400 m of altitude,
later northward, eastward and upward in spring and early summer and
the reverse in late summer and autumn" (Wang, 1967, p. 315).
Examine the validity of this 'law' in relation to selected phenological
events in your area.
8.10 "A good year of snow is an indication of abundant harvest." Analyse the
truth of this statement, indicating the climatic regimes and circum-
stances in which it could be valid, and explain why a 'good' snow cover
could be beneficial. If you live in a region with a temperate or cold
climate, examine the year-by-year gross production of wheat or other
crops in relation to the winter snow cover.

141
CHAPTER 9
AGROMETEOROLOGICAL INSTRUMENTS AND OBSERVATION

9.1 BASIC All aspects of local climate have to be considered in agricultural meteorology.
OBSERVATION RULES However, it is convenient to divide observations into two categories 'ordinary' or
'routine' meaning the standard official synoptical or climatological types, and
into those meant for special agrometeorological purposes (see section 5.2 and
section 5.4 for the difference).
Homogeneity is an essential aspect of all climatological observations. This
means that results of measurements must be comparable from year to year and
from place to place. Broadly speaking, this requires that comparable types of
instruments be used at all stations. Their exposure should be comparable, and
should not be changed significantly. Official climatological or synoptical network
stations do have to conform to instrumentation and maintenance requirements in
the station guides mentioned below.
Moreover, comparability of data between areas requires that observations are
made at regular and agreed times throughout the 24 hours. For synoptic purposes
it is essential that observations be taken at the same agreed times according to
'Universal Time' (UT). The main synoptic observation times are 0000, 0600, 1200
and 1800 UT(= GMT, Greenwich Mean Time) with subsidiary times 0300, 0900,
1500 and 2100 UT.
At stations which make only climatological observations, night observations
may not be possible and readings at three of the principal synoptic hours may be
accepted. In climatology generally, and not least in the context of agriculture,
comparability of data is more likely to be linked with the daily course of the Sun,
and hence with Local Mean Time (LMT) as agreed for the various time zones rather
than UT. So a common practice involves observations at 0700, 0800 or 0900, at
1300, 1400 or 1500 and at 1900, 2000 or 2100 (LMT). Choice among these times
varies from nation to nation.
If only a single daily observation is possible, current practice is to select one
at 0700, 0800 or 0900 LMT. The 24-hour totals for rainfall, evaporation and the
diurnal extremes (daily maximum and minimum temperatures) are read at the
early morning observations. At many stations, sunshine cards are changed in the
morning as well, but it is better to exchange them in the evening after sunset, then
every card refers to a single day only. In agrometeorology the diurnal course of
weather is very important, so for many agrometeorological purposes one single
daily observation is insufficient.
For detailed instructions on formal and practical observational practices and
on instrumental requirements, reference should be made to:
• Guide to Meteorological Instruments and Observing Practices (WMO-No. 8);
• Guide to Climatological Practices (WMO-No. 100);
• Guide to Agricultural Meteorological Practices (WMO-No. 134).

9 .1.1 Agrometeorological GM-II and the various WMO Guides give the site requirements for synoptic
networks and documentation stations and for standard climatological stations. However, as explicitly
Reference: Linacre (1992) mentioned in WMO-No.134, places in which crops are raised, animals live and
grow, diseases and pests flourish and operate, are markedly different from the
open sites such as airports regarded as ideal for synoptic and standard climato-
logical purposes.
Recognizing these differences, it is frequently necessary to set up a network
of subsidiary agrometeorological stations. Assuming that the standard
synoptic/climatological network stations exist, the first stage is to equip some of
them as 'base' agrometeorological stations by addition of suitable observations -
at least global radiation, and surface and soil temperatures. It is frequently the case
that in the subsidiary agrometeorological network some instruments are deployed
in non-standard conditions of exposure or sampling, the most common example

143
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

being that measurements are made at a number of heights both above and below
the standard. In that case it is necessary to replicate this arrangement at the base
station, in order to separate the effects of unconventional observation practice
from those of site differences.
It is usually necessary to supplement these 'base' stations by other auxiliary
agrometeorological observing stations. In these auxiliary stations, the surrounding
terrain may, and typically will, differ from that of the standard 'base' stations.
There are broadly two types of auxiliary stations:
(1) Semi-permanent stations, set up to investigate and monitor conditions
within the very special agricultural situations typical of a region - e.g.
orchards, vineyards, valley floors, hilltops and slopes;
(2) Temporary stations, installed for specific 'probe' investigations over a relatively
short period, e.g. the growing season, or season of spring frosts. Even a few days or
nights in a selection of weather situations, such as clear skies and light winds (i.e.
conditions of high levels of solar input and terrestrial radiation loss), or of overcast
windy days (in studies of exposure), may be adequate for specified purposes.
Adequate documentation of agrometeorological stations is a prerequisite for
the evaluation of their observations - even more so than for stations of the synop-
tic/climatological network, which are presumed to conform to instrumentation
and exposure standards. Often even for these official stations the situation is not
ideal, because an ideal site with good attendance and perfect exposure is seldom
easily available. Some airport observing sites, for instance, are heavily sheltered by
airport buildings. Conditions are still more difficult for agrometeorological
stations, where even the best available site may have severe limitations. Knowing
that the station was on the best possible site does not describe the site.
Therefore agrometeorological networks always require a good archived
description of all stations. Without this the analysis of the observations is uncer-
tain and corrections are impossible to implement. To give a factual example: a
network used several anemometer types, with read-outs in kilometers or in miles
- but nobody had recorded which type had been used at which time at individ-
ual stations. Thus a wind climate evaluation was impossible.
The first item that needs to be described is the site of the station, using a map
with a typical scale of 1:104 , which should show the surroundings: obstacles (and
their height), surface roughnesses, slopes, water surfaces, and directions of distant
hills or of other major terrain features. Any changes (e.g. significant growth of
vegetation, new buildings) should be noted with the date of change. A few photos
may help.
The second action that needs to be taken is to archive the actual (as compared
to the proposed) state of the instrumentation: height of observation, identification
·of instruments, control measures (cleaning, psychrometer wick change, calibration),
as well as malfunctions and repairs. This requires a logbook to be kept at the station.
Minor items (e.g. cleaning times) should be logged, as well as major occurrences.

9 .1.2 Dynamic responses of All instruments require time to adjust their reading completely to a change in the
meteorological instruments element being measured, and usually the adjustment speed is proportional to the
difference between actual and desired reading. For example, suppose that a
thermometer of which the reading at time (t) is called T(t) is plunged at time t = O
into a medium of temperature T0, then the change speed of the reading is
dT/dt = [T0-T(t)]/A. Here the proportionality constant A is called lagtime and has a
value which depends on the thermometer type (and on its ventilation). As a result
of this, the reading adjusts exponentially to instantaneous change of the temper-
ature to a new value T0 :

(T0 -T(t)) ~ (T0 -T(0)) exp(-±)

written out in the following table:
t/A 1.0 2.3 3.0 4.6 6.9
To - T(t) = 0.368 0.10 0.05 0.01 0.001
T0 -T(O)

144-
CHAPTER 9 - AGROMETEOROLOGICAL INSTRUMENTS AND OBSERVATION

So after three lagtimes the adjustment has been completed to within 5 per cent of
the difference.
If the sensed variable T changes periodically with period-rand amplitude A,
then the result of the instrument lag is that the read-out will show a diminished
amplitude AA and a phase lag q> in time. In formal notation this is represented as:
if T(t) = A sin (2nt/-r) then A/AA= ✓1 + (2n)J-r,) 2 and q> = arctan (2tl.lt)
Evaluation of these relations will show that fluctuations with a period -r =
S).. will only be indicated with 62 per cent of their real amplitude, and for -r =
1Q).. this will be 85 per cent. For observing agrometeorological processes, when
these are caused by fluctuations with short duration, it is therefore necessary to
choose instruments with short lagtimes.

9.2 AGROMETEOR- Descriptions of conventional instruments, installation and operation are

OLOGICAL INSTRUMENTS discussed in GM-II, and also in WMO-Nos. 8, 100 and 134. Most of the apparatus
described in these publications is simple and almost fully mechanical, partly
because advanced methods of measuring atmospheric parameters usually require
electricity. This is not available in a regular way at many places where meteoro-
logical measurements are desirable. Therefore classic instruments such as bimetal
thermographs or counter anemometers have been and will continue to be the
solid foundation of weather observation.
However, technological developments in the past 20 years have made it possi-
ble to replace writing recorders, requiring either a connection to a mains power
supply or substantial accumulator capacity, with small digital recorders which need
only a few small dry batteries to run for months. The availability of batteries may
still be a problem in some cases, and repairing the instruments requires technical
skills and spare parts which are not always available - but it is clear that for elec-
trical instrumentation in meteorology the problems have diminished by an order of
magnitude. Also in view of the increasing availability of electric power from field-
usable solar cells, there is no longer any reason to omit or minimize discussion of
well-tried types of electrical measurement, or to classify them as 'only for research'.
In addition, this section will discuss some instruments and some observation
methods, which are not routinely used at standard synoptic or climatological
stations, but which have particular importance for agrometeorology. Here the
main criterion for discussing instruments is that they must be usable in an opera-
tional context, like an agrometeorological station network. For specialized
instrumentation, one should refer to advanced research books like Fritschen and
Gay (1979) or Kaimal and Finnigan (1994). However, please note that some of the
observation methods which have already been adequately covered in GM-II, such
as pressure measurement, are not discussed again here.

9.2.1 Air temperature Standard thermometers with differential expansion as the working principle -
mercury or alcohol in glass, Bourdon, bimetals - have been extensively discussed
References: GM II,
in GM-II.
pp. 185-194; WMO-No. 8,
One non-standard instrument of this type, used in agrometeorology for the
Chapter 4.
sole purpose of determining the range between daily maximum and minimum
temperatures, is a combined 'max-min' spirit thermometer (type 'Six') bent into a
U-shape. The spirit column is divided into two portions by a long 'plug' of
mercury, which occupies the U-bend and extends up each arm of the U. At each
end of the mercury plug is a small metal index, which the mercury pushes along
the tube in response to temperature changes affecting the spirit-filled bulb. The
indexes remain where they are pushed to give temperature extremes reached
during the daily cycle. After read-out they are brought back into contact with the
ends of the mercury column by use of a magnet. This instrument gives useful
information on the amplitude of the diurnal course, but its absolute temperature
readings are not reliable - observations with the 'Six' have errors of at least ± 1°C
because the bore of the bent part of the tube is never entirely constant. However,
one of its advantages is that it is relatively cheap, precisely because it is not a preci-
sion instrument.

145
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

Electrical thermometers have two advantages over mechanical thermometers:

(1) They can be read at a distance, either directly or recorded;
(2) They are smaller than glass thermometers or thermographs, and can there-
fore be used for observation in small spaces or thin layers.
Three types of electrical thermometer are mainly used: resistance thermome-
ters, thermistors and thermocouples - each with different advantages and
disadvantages. Here, we include only brief descriptions to allow a reasonable choice
to-be made for specified purposes. For detailed information the literature can be
consulted, such as Meteorological Office (1981) and Strangeways (2000).

Resistance thermometers Resistance thermometers are based on the temperature-dependent resistance of

platinum or nickel wire. If R1 is the resistance at T°C and Rm at TO°C, then:
RT = Rm (1 + a (T - To) + b (T - To) 2]
The values of the constants are:
. for platinum: a= 3.85 x 10-3 °c-1, b = -5.5 x 10-7 c-2
0

for nickel: a= 5.5 x 10-3 °c-1, b = +7.5 x 10-6 c-2

0

So the linearity is reasonable, certainly for platinum. Usual sizes are lO0Q and
500Q. The calibration of a good Pt-resistance thermometer is quite stable.
Read-out of a resistance thermometer is generally carried out using a servo-
stabilized Wheatstone bridge circuit which requires a stable power supply. The
bridge current is kept small in order to avoid heating the resistance. In order to
compensate for cable resistance, a three-wire cable is used and one of the bridge
junctions is placed close to the Pt-resistance at the observation point.

Thermocouples Thermocouples are junctions of two different metals, e.g. constantan (a nickel-alloy)
and copper. A complete electric circuit will have a pair of these junctions - and if
the junctions are at different temperatures, T and (T + ~T), a voltage difference ~V
Figure 9.1
is generated in the circuit, which is almost linearly proportional to ~T:
A thermopile consists of a series
of thermocouples made up of ~V = 38.47 ~T + 4.66 (~T)2 µV K-1 for copper-constantan at 0°c
copper/constantan junctions
In practice the quadratic correction (<1%) is omitted, and we say ~V =40 ~T µV K-1,
since the proportionality constant increases slightly with increasing
temperature.
Generated voltages are small, but since thermo-electric forces are additive,
we can use a chain of thermocouples with all copper-to-constantan junctions at
the relatively cold temperature, and all the constantan-to-copper junctions at the
Q) Q)
relatively warm temperature. The multiple-thermocouple combination is called a
::c
~
::c
~ thermopile; in Figure 9.1 the thermo-voltage is about 200 ~T µV, if ~Tis the differ-
u u
.ence of TA and T8 in °C.
D C However, care should be taken to ensure that no other thermo-electric volt-
Q) ages can be generated at other places in the circuit, e.g. at the voltmeter
a.
a. connections. If a thermopile is connected to a cable, the two cable junctions (C
u
0 A
and D in the figure) should be at equal temperatures. Also take care to use a
high-resistance voltmeter, in order to avoid cable resistance corrections. When
stable solid-state amplifier chips became available in the 1970s to measure small
thermo-electric voltage, thermocouples became much more useful.
B
Thermistors Thermistors are resistive semi-conductors with a high but non-linear Negative
Temperature Coefficient of resistance - this is the reason why the name 'NTC' is
often used. The temperature dependence of NTC's is highly non-linear. They are
smaller than Pt-resistances, which is an advantage, but their calibration is not
universal and it changes rather fas_t, needing frequent checking. Therefore they are
less operational than Pt-resistances.
If we compare the above alternatives of electrical thermometry, but forget
thermistors because of their poor interchangeability, the first notable point is that
thermocouples measure temperature differences. This is a handicap for use in
absolute temperature measurement, because that can only be achieved by keeping
one (set of) junction(s) at a reference temperature. NTC-based zero-reference-
temperature units are available (at a price), and a simple field trick is to use the

146
CHAPTER 9-AGROMETEOROLOGICAL INSTRUMENTS AND OBSERVATION

stable deep soil temperature as a reference. However, it is much more convenient

to measure absolute temperatures with resistance thermometers.
On the other hand, situations in small-scale meteorology exist where accurate
temperature difference measurements are required, and the absolute temperature is
of secondary importance. Examples of such applications are vertical temperature
gradient observations, psychrometers and radiation balance meters. For such
purposes a thermopile is excellent, as the required accuracy can often not be
obtained by taking the difference of two absolute temperature measurements.
The second point of comparison is size. Thermocouple sensors can be very
small, a useful characteristic in special observations (e.g. leaf temperature meas-
urements), but this also results in negligible radiation error, so that they can be
used without ventilated shielding. Their small size also means a small lagtime (see
below), which is a disadvantage for measuring averages and an advantage for
measuring fluxes. The size of a resistance thermometer is moderate, much less
than that of a mercury thermometer, but not very small.
For the major sensors which are currently available for air temperature meas-
urement, it is worthwhile to make a general comparison in Table 9 .1.
Table 9.1
Mercury Bimetal Bourdon Resistance Thermo-
in glass couple
Size large very large moderate small very small
Output I !lT ± + + +
Linearity + ± ± + +
Lagtime x ✓u 100 s sos 200 s 10 s 1s
Working accuracy (±) 0.2°c 0.5°C 0.S°C 0.2°c 0.1°c
Calibration stability + ± ± + +
Recording option no yes yes yes yes
Repair possibilities + ± +
Electricity requirement none none none high moderate
Price low low high high moderate

The lagtime A of various temperature sensors is proportional to C/✓U, where C is

the heat capacity of the sensor and U is the ventilation speed supporting energy
exchange between sensor and surroundings. The radiation shielding also adds to
the lag: the lagtime of a Stevenson screen is about (600/✓U) sec (with U in m s-1).

9.2.2 Grass-minimum In spite of doubts about the physical meaning of the observation (see section
temperature and radiative 2.2.2), the grass-minimum temperature, the lowest temperature reached in a
surface temperature 24-hour period by an unscreened spirit thermometer exposed 'just above' the
surface, is a routine station measurement (see GM-II, p. 191 and HMSO, 1982).
References: GM II,
Although ideally the thermometer should be placed 'just touching' the tips of the
pp. 190-193; WMO-No. 8,
blades of mown grass, such exposure can only be 'typical' in temperate humid
Chapter 4; WMO-No. 134.
areas. In many regions, continuous grass cover is extremely unusual and it is more
common to find bare soil, sand, rock, etc. Even in temperate areas, seasonal crop-
ping will result in fallow soil for much of the year.
Accordingly, the practice is growing of exposing a duplicate instrument over
bare soil. Apparently the bulb may be put in direct thermal contact with the
surface, or a fraction of a centimetre above it. The practical value of this 'probe
type' observation is evidenced, for example, by the use in the United Kingdom of
a grass minimum thermometer placed on a slab of concrete in order to indicate
the risk of ice formation on highways. In countries such as France and
Switzerland, an exposed spirit thermometer is placed at 0.5 m from the surface for
an observation termed 'indice actinothermique'.
Measurement of temperature at actual surfaces, either of the ground or of
vegetation tops, are difficult to achieve even with extensive arrays of thermocou-
ples. One of the reasons of the success of the evaporation determination method
of Penman (1948) was the fact that it did not require any input of the 'unknown
and unknowable' temperature and humidity at the surface. Only over water

147
WMO LECTURE NOTES - AGRICULTIJRAL METEOROLOGY

surfaces can the surface temperature be easily determined, using a small electrical
thermometer on a little float.
However, currently available remote sensing instruments make it possible to
determine surface temperature TO integrated over small areas, using a hand-held
radiation thermometer. The instrument observes in the 8-14 µm waveband the
radiation entering its frontal opening; if pointing at the ground, that is the radia-
tion emitted by the surface (proportional to T0 4) and the reflected sky radiation at
the same wavelengths. The latter can be neglected in usual cases with IR emissiv-
ity approximately 1.0. So T0 == ( Li/cr)114, where cr is the Stefan-Boltzmann
constant. Neglecting emissivity differences in the types of observed surfaces gives
for this radiative temperature an uncertainty of about ±1 °.

9.2.3 Soil temperature and In temperate regions, the practice has been to measure soil temperatures at depths
soil heat flux to 0.2 m under a bare, consolidated soil, and at greater depths, 0.3 m, 0.5 and
1.0 m under short turf. The custom is growing of duplicating the 50 mm, 100 mm
References: GM II,
and 200 mm depth observations under short turf.
pp. 191-193; Rosenberg, et al.
The traditional mercury-in-glass bent-stem soil thermometers are described
(1983); WMO-No. 8, Chapter 4.
in GM-II, p. 192-193. Continuous records of soil temperature can also be obtained
using a distant-recording mercury-in-steel twin bulb instrument (GM-II, p. 190).
Additionally, all forms of electrical thermometry are usable for soil temperature.
Bringing them at the desired depth with the least possible soil disturbance is done
by digging a temporary pit and inserting the thermometer elements from suffi-
ciently far aside that vertical heat transfer along the cable is minimal - the first
part of the cable close to the sensor should be horizontal.
Since for much of the year (except perhaps in high latitudes) and apart from
the top few centimetres of soil, the daily course of soil temperature exhibits a well-
defined cycle, observations of the range between daily maxima and daily minima
give information which is useful for many purposes. 'Six' thermometers (see
section 9.2.1) can be employed for this. Bulbs are usually inserted into the ground
at depths of 50 and 100 mm.
In certain situations, for example, to investigate the soil flora and fauna in
the uppermost layers of the soil, it is necessary to obtain an estimate of the
'surface' temperature of the ground. Often the actual requirement is a measure of
temperature in the top few centimetres or millimetres. Then a useful estimate can
be achieved by placing a thin-bulbed thermometer horizontally on the ground
surface, pressing it gently into the soil and dusting a layer of soil over the bulb.
Gloyne (1952) summarized results from Penman (1943) and other investigators,
using a range of similar procedures, empirically:
T(sx) == 2 T(x) -10 for T(x) > 10°c
where T(sx) and T(x) are the maximum temperatures in, respectively, the
upper soil layer and the standard screen. The 'skin' temperature estimate was
(T(sx) + 8) C.
0

The vertical soil heat flux density is equal to (-k !1T/!1z), where L\T is the soil
temperature difference over the depth interval !1z, and k is the thermal conduc-
tivity of the soil. But k varies appreciably, not only with type of soil but also with
the soil water content (see section 3.2). Therefore it is useful to support soil
temperature profile observations with direct measurement of the soil heat flux
density.
Heat flux plates are used for this purpose. They are small disks, sensing the
temperature difference between their top and bottom faces by means of a
thermopile. For good measurement, the plate heat conductivity should be
comparable to that of the soil. The small plates are typically buried at a
compromise depth of 0.1 m. Deep burial makes them unrepresentative for soil
heat flux at the surface, but very high placement leaves a very thin covering soil
layer, which then may dry out or crack. The presence of plant roots also has to be
considered. For full information see Fritschen and Gay (1979).

148
CHAPTER 9 - AGROMETEOROLOGICAL INSTRUMENTS AND OBSERVATION

9.2.4 Wind Classic wind instruments, namely rotation and pressure-tube anemometers, are
described in GM-II. Parts of that text are excellent, in particular the description of
References: GM II,
the Dines pressure-tube anemograph - a reliable and useful instrument, which is
pp. 213-222; WMO-No. 8,
the only anemometer mentioned in GM-II which is fully compatible with the
Chapter 6; Mazzarella, 1972;
requirements of good wind analysis. It provides registration of gustiness and
Wieringa, 1980.
azimuth without the need of electricity. Unfortunately, the Dines is being phased
out in most countries (Pike, 1989).
The cup anemometer is an excellent instrument: it is relatively cheap and
sturdy and its calibration is reasonably linear, provided that the cup arms are rela-
tively short compared to the cup size. It does not matter that nobody really
understands why the cup anemometer works so well - it simply does. It is also
easy to repair. Its output is an axis rotation speed, and there are better and worse
ways to read this speed out for use.
The simplest useful read-out form is the counter anemometer which has a
small set of gears activating a four- or five-digit read-out of the wind run. Each
time that another kilometer of air has passed by the anemometer, the read-out
increases by 1 km. By reading the wind run at regular intervals one can obtain
the average windspeed in the interval. Unfortunately, it is not possible to obtain
gustiness information in this way, but this is compensated by the practical
advantage that no electricity is needed to run this type of read-out. Upon
acquiring a counter anemometer, one should take care to get a type that has a
downward-directed read-out window with large figures. Such anemometers can
be easily read when placed on top of a pole of a few meters length - though the
formal requirement of 10 m height can only be met with field-glasses. The
counter types with sideways-directed windows can only be used at heights which
are too low. Typical starting speeds of counter anemometers are at least 1 m s-1
(2 to 3 knots).
The only other type of cup anemometer with a distant read-out, which does
not require an electrical power supply, pays for this by severe loss of usefulness at
low wind speeds. The so-called dynamo anemometer, containing an electricity
generator which provides a distant indicator reading, has an average starting speed
of 3 m s- 1 (six knots). At lower wind speeds the wind has insufficient energy to
drive this mini wind turbine.
Distant continuous read-out of windspeed is obtained from cup anemome-
ters in a reliable manner, when axis rotation is translated into an electric signal
frequency. This is done by making the anemometer generate a number of pulses
per rotation, either with a light chopper or with a magnetic reed contact. An extra
advantage is that this form of rotation sensing exerts no mechanical load on the
axis, so that the anemometer starting speed can be of the order of 0.5 m s-1
(1 knot). Pulse train frequencies are not modified by cable transport, and can be
easily transformed into some form of digital or analog recording. For field work,
battery-driven wind reading units have been developed, giving averages,
maximum gusts and standard deviations of windspeed.
Frequent use is made of hand anemometers which work on the tachometer
principle. A direct read-out in windspeed units is achieved by attaching to the cup
axis a magnet, placed inside a copper cylinder. The rotating magnet entrains the
cylinder, while the entrainment of the cylinder is counteracted by a spiral spring.
Because hand anemometers are only an arm's length away from wind blockage by
the human body, they should not be used for calibration, but rather for a rough
indication of wind force.
Anemometers are often placed at a height of only two metres above ground,
since the Penman evaporation formula (section 4.4.1) uses wind at that height.
However, in rough terrain we need a placement higher than 20 z0 (section 5.3).
Sometimes, for getting records of wind at surfaces of evaporation pans, a measur-
ing height of one metre is used, but such badly-blocked low observation is
absolutely useless for any other purpose (just as Stevenson-screen temperature at
1.5 m cannot be 'replaced' by grass-minimum temperature at 0.1 m).
If studies of airflow over natural terrain are more demanding than simple
broad surveys, available cup-counter anemometers must be supplemented by

149
WMO LECTIJRE NOTES - AGRICULTIJRAL METEOROLOGY

observations from some instruments which provide either a continuous record of

wind speed and direction, or at least a record of frequent instantaneous (i.e. short-
period) measurements. As a rough guide for consideration, four (or even three)
ten-minute observations, equally spaced throughout a 24-hour period (always
including wind direction), may be sufficient to upgrade a coarse survey of the
wind climate of-an area into a useful description for solving problems of wind and
exposure which arise in agriculture.
For agrometeorological research, anemometers are also necessary to provide
information on vertical windspeed fluctuations. Hot-wire flow meters are some-
times used for this purpose, but their fragility and their fast calibration change in
field conditions makes them more suitable for laboratory conditions. Sonic
anemometry, i.e. sensing the transport speed of sound pulses, has now become a
common, useful and reliable (but high-priced) turbulence measurement method.
Triple propellers and propeller bivanes are a slightly cheaper, but slower alterna-
tive. Research handbooks on the subject (Fritschen and Gay, 1979; Kaimal and
Finnigan, 1994) should be consulted.
A sub-standard probe is the 'tatter-flag' (McAdam, 1980). The rate of wind
destruction of a carefully specified textile flag is related qualitatively to the
monthly average wind experienced at the site. It has found some use in Scottish
forestry for assessing patterns of wind run within an area of, say, a few square kilo-
meters. Unfortunately, this probe appears to be useful only in climates where
annual average wind speeds (at 10 m) exceed 7 m s-1, which rather limits its range
of application regions.
A simple sub-standard probe for cases where 'something is better than
nothing' can be made from a table tennis ball (size and weight are standard) on a
thin pivoted rod. Wind drag will make this pendulum move upward, and its angle
of displacement can be related to airspeed by calibrating it from the window of a
moving car.

-□
Figure 9.2
Wind vane shape and quality.

Bad vane ®~t

I Good vane
Good analysis needs azimuth records, and any handyman can make a useful
windvane. For the fin a plate should be used, at least twice as high as its dimen-
sions downwind, light, and just a flat plate; double fins in wedge position can
make the vane misbehave (Wieringa and Van Lindert, 1971). The counterweight
·to balance the vane must have a much shorter arm than the fin-arm's length in
order to obtain steady vane behaviour. If the fin arm is long enough and the vane
axis rotates without excessive friction, a simple vane will indicate azimuth well for
winds down to 1 m s-1. Windbags, and other tricks mentioned in GM-II, do not
perform any better at low wind speeds.
Wind direction registration on a flatbed recorder requires ingenious solu-
tions, either mechanical (see Dines) or electrical, to prevent continuous full-scale
writing when the azimuth varies around North, say between 350° and 10°. This
problem is avoided by the Baxendall-type of recording on a drum which rotates
along with the wind vane, either through direct placement on the bottom part of
the vane axis, or by some form of mechanical or electrical coupling. A
wormwheel-mounted pen then can be moved down the drum in a day or in a
month (see e.g. Meteorological Office, 1981).
The fluctuations of wind velocity, usually called gusts (short maxima) and
lulls (short minima) are such a basic property of wind in the surface layer, that
instruments which are not capable of rendering them should be classified as
insufficient. Regular users of wind data, e.g. in aeronautics or construction design,
ask for information on maximum gusts along with the average wind speed
information. Agrometeorologists and air pollution meteorologists, to take but two
examples, dealing with turbulent transport in the atmosphere, profit from some

150
CHAPTER 9-AGROMETEOROLOGICAL INSTRUMENTS AND OBSERVATION

information about actual degrees of turbulence. Moreover, information on the

ratio of maximum gust and average wind is directly usable to obtain an objective
correction for obstacle shelter errors in the average wind speed (see section 5.3).
There are three alternatives for recording gustiness. The classic method is to
make a continuous recording of measured wind speed on some type of pen
recorder, and to read off the maximal recorded gust along with the average. The
electronic alternatives require sampling anemometer output for 3-second periods
(WMO-CIMO recommendation), and then either to read out the largest gust which
occurred during the averaging period of 10 minutes or 1 hour, or to sum the values
together with their squares and compute a standard deviation at the end of the
sampling period. All three gustiness forms can be usefully employed for roughness
and dispersion analysis (Beljaars, 1987b). Other periods (10-second sampling, 20-
minute averaging) can be used with minor analysis modification. However,
1-minute samples should never be used, because they are too long to be analysed as
gusts, and too short to provide effective averaging of atmospheric turbulence.

9.2.5 Radiation and As was discussed in Chapter 2, the short-wave radiation reaching the Earth's
sunshine surface is derived from the direct solar beam ('direct' radiation) and from solar
energy which has been dispersed by molecules, dust and aerosols, reflected by
References: GM II,
clouds and transmitted through clouds, eventually to reach the ground as diffuse
pp. 275-280; WMO-No. 8,
radiation from all parts of the sky canopy ('diffuse' radiation). As shown earlier:
Chapter 9; Robinson (1966);
Coulson (1975). Total short-wave radiation on horizontal surface (KJ.) =
= direct component (S) + diffuse component (D) = I sin h + D
where I = intensity of the solar beam normal to the beam, and h = solar altitude.
Standard meteorological stations usually measure only sunshine duration.
The traditional instrument to observe this is the Campbell-Stokes sunshine meter
(GM-II, p. 275-280). WMO abolished the world standard status of this sunshine
meter in 1989, as the process of evaluating the bums on its daily cards was both
cumbersome and arbitrary. Instead, sunshine duration has been defined as the
time in which direct radiation (on a plane perpendicular to the Sun's beam) is
larger than 120 W m-2 • This definition makes it possible to now use automatic
sunshine recorders, like those of Soni and of Haenni (Dehne, 1989), at meteoro-
logical stations.
Agrometeorological stations have to obtain full information not only on the
state of the atmosphere, but also on the energy balance at the surface. Therefore, as
said earlier, they must expand the standard station program at least with measure-
ments of soil temperatures and of radiation. Sunshine data are useful for estimating
e.g. monthly radiation averages. But on an overcast day the diffuse radiation is
sometimes strong and sometimes very weak; the difference is quite important for
agriculture, but it cannot be deduced from sunshine duration observations only.
For radiation instruments the following names are customary:
• Pyranometer (or 'solarimeter'): for downward global short-wave radiation;
• Pyrradiometer: for downward global short- and long-wave radiation;
• Pyrgeometer: for downward global long-wave radiation;
• Pyrheliometer: for direct-beam solar radiation; and,
• Net pyrradiometer (or 'balance meter'): for net short- and long-wave radiation.
It is useful for agrometeorological stations to measure separately direct and
diffuse downward short-wave radiation. This requires two pyranometers, one for
measuring global radiation and one for diffuse radiation only. A poor man's alter-
native is a pyranometer for global radiation, combined with some measurement
of actual times of bright sunshine. Whenever other than horizontal surfaces are
involved, the distinction between inputs of direct and diffuse radiation is impor-
tant. This occurs in some agricultural problems such as the operation of
glasshouses, the efficiency of orientation of row crops and of soil ridging, the
seasonal variations of hillslope climate - leaving aside the question of the appli-
cation of direct/diffuse radiation in architecture and civil engineering. Separate
data on diffuse radiation are also useful in estimating Photosynthetically Active
Radiation (PAR) (Stigter, et al., 1987).

151
WMO LECTIJRE NOTES - AGRICULTIJRAt METEOROlOGY

A diffusometer is a pyranometer which is shaded from the direct beam of the

Sun, either by a shadow ring mounted on a polar axis, or (more difficult to
achieve) by a Sun-following disk. If only a single solarimeter is available, useful
rough estimates of the diffuse contribution can be obtained by shading the pyra-
nometer from the direct sunbeam with the hand for the duration of several
pyranometer lagtimes. Shadow rings also block about 10 per cent of the diffuse sky
radiation, requiring a season-dependent correction.
A pair of pyranometers, one of them placed upside-down, can measure the
short-wave radiation balance and determine the albedo of the underlying surface.
Global radiation sensors consist of a flat black thermometer under a dome,
which serves to shield it from sensible and latent heat exchange with the
surrounding aiL If the dome is made of glass, only short-wave radiation can reach
the thermometer. Polyethylene has good long-wave transparency; however, that is
rather weak material for a dome, so that its dome-shape can only be maintained
by filling it with compressed air. Condensation on the inside of domes is
prevented by drying agents (such as silica gel). The outside of domes must be
cleaned very regularly, and after rain they must be dried. For drying purposes, arti-
ficial ventilation is sometimes used.
Great care should be taken to mount the black thermometer face horizontal
and exposed to the whole hemisphere of the sky. Exchanging the dome for a flat
cover is not good, because flat transpar_ent plates reflect radiation at low input
angles. To prevent errors from wind cooling of IR-heated domes, a second dome
can be used. Polyethylene discolours with time, depending on its quality, and
should be replaced every couple of months.
In the Moll-Gorczynski (Kipp) pyranometer, the dome is glass, transparent
up to 4µm, and the thermometer is a black surface with a thermopile underneath;
the other side of the thermopile is connected to a heat sink. The Eppley pyra-
nometer also has a thermopile sensor, but that compares the temperatures of black
and white surfaces under the same glass dome. In both instruments, radiation is
derived from the thermopile voltage. The output of a standard Moll-Gorczynski is
approximately 13 µV per W m-2 incoming radiation.
Two non-electrical pyranometers are also in use. The Robitzsch actinograph
has two bimetal strips under the dome, one black and one white. When this pair is
heated by short-wave radiation, the difference in their bimetal curvature is recorded
in the same way as for ordinary thermographs. The continual curvature changes
make the bimetal surface paint crack, but with good maintenance the resulting
gradual change in calibration is less than 4 per cent per year (Stigter, et al., 1989).
In the Bellani distillometer, the thermometer is a black bulb partly filled with
an easily evaporating liquid, CC14 or alcohol, sometimes water. Heating the bulb-
·by short-wave radiation makes some of the liquid vaporize, and it condenses again
in a glass burette tube attached below the bulb. Lowering the burette into a hole
in the ground provides constant condensation temperature. For reading the
instrument is raised, the liquid level is noted, and the instrument is temporarily
inverted to refill the black bulb.
When choosing a pyranometer, non-availability of electricity rules out the
Kipp and the Eppley. An additional argument is the response of the various instru-
ments: lagtimes are about 8 to 10 sec for a Kipp, about 10 min for a Robitzsch
actinograph, and about 20 min for a Bellani. For Europe, good following of poss-
ible radiation changes needs').. ~15 sec.
In partially cloudy weather, radiation fluctuates widely. Manual analysis of
data from Kipp and Eppley is then extremely time consuming and electrical
recording and integration is then almost inevitable. The sluggish response of the
Robitzsch leads to a substantial 'smoothing' of chart traces, and consequently a
reduction in labour for hand-scaling charts. The Bellani is an integrating instru-
ment. The accuracy of daily totals is ±5 per cent for Kipp and Eppley, ±10-15 per
cent for Robitzsch and ±15-30 per cent for Bellani.
The non-electrical pyranometers receive little coverage in handbooks. More
information can be found in Gunn and Yeo (1951), Pereira (1959), and McCulloch and
Wangati (1967) for the Bellani, and in Esteves and De Rosa (1989) for the Robitzsch.

152
CHAPTER 9 - AGROMETEOROLOGICAL INSTRUMENTS AND OBSERVATION

Good measurement of short-wave global and diffuse radiation at a reason-

able amount of (agro)meteorological stations would be an improvement on the
present situation, in which regular radiation observation is still an exception. For
example, between 1973 and the mid-1980s, the United States weather station
network made measurements only of sunshine duration. Even the present public
interest in global warming has not yet manifested itself in large increases in radi-
ation monitoring. The longest available series of global radiation measurements
recorded at Wageningen dates from 1928 (De Bruin, et al., 1995).
Other forms of radiation measurement, as tabulated above, are even less
common at stations but will be mentioned here briefly. The next most important
measurement is net pyrradiometry. 'Radiation balance meters' with upward- and
downward-directed polyethylene domes are either of the Schulze type (two sepa-
rate pyrradiometers), or of the Funk type (one thermopile for the two domes).
An interesting variation is the net radiometer of Beckman and Whitley (see Suomi,
et al., 1954): it is domeless, and equalizes environmental influences on upward
and downward surfaces by strong ventilation.
Pyrheliometers (e.g. Angstrom, Linke-Feussner) are essential instruments for
a calibration centre, but are not suitable for field work. Further information can be
obtained in the handbooks.
Advances in photoelectric solar cell technology suggest new possibilities to
develop cheap new high-sensitivity pyranometers (Kerr, et al., 1967), but this has
not been done on a large scale. One reason might be their high sensitivity to very
short wavelengths, while UV is not of top importance in many investigations.
Selenium does not really react to IR radiation and silicon does so only to the near
IR. The option might be kept in mind for investigations which require very many
radiation observation points.
The Jordan sunshine duration meter, where sunrays mark photo-sensitive
paper in a can with a hole, is maybe also disappearing because of its UV-preference.
It has sometimes been used as a simple sub-standard probe (Beadle, 1966).
Measurement of illumination differs from radiation measurement.
Luxmeters have a wavelength-dependent sensitivity conforming to that of the
human eye, aimed at giving no response at all outside the range 0.4--0. 7 µm, and
maximal response around 0.55 µm (Platt and Griffiths, 1964). Photographic light
meters only need to represent the sensitivity of film or electronic sensors and have
differing sensitivity ranges. Anyhow, for meteorology all illumination data are of
limited use (section 2.1.4).

9.2.6 Humidity, dew and Traditional instruments to measure atmospheric humidity, namely the psychrom-
leaf wetness eter and the hair hygrograph, are discussed at length in the references given
above. However, it should be mentioned that the wet-bulb-depression (T-Tw) can
References: GM-II,
also be measured directly and accurately by means of thermocouples (Hickman,
pp. 205-212; WMO-No. 8,
1970). The advantages of this are that less, or no artificial, ventilation is needed
Chapter 5.
for full response, and that a thermocouple psychrometer sensor can be built small
enough to be used directly on crops or insect habitats.
Most other methods mentioned in literature, such as Regnault dewpoint
hygrometers or electrical absorption cells, do not survive well in meteorological field
conditions. Infrared radiation absorption analysis is an exception, and it can, in
principle, also be used at meteorological stations to determine CO2 concentrations.
All operational forms of hygrometry require regular maintenance, cleaning
and calibration. For field checks generally comparisons with Assmann psychrom-
eters are used. An alternative is to fit out a simple humidity chamber (Hickman,
1970), in which a few humidity checkpoints can be provided over saturated solu-
tions of certain salts. Over a NaCl-solution we get RH =76 per cent, and over
Mg(NO 3)z we get RH =55 per cent.
Actual measurement of dew has, so far, not been standardized, and only a
few countries carry out routine measurements of dew, usually with specific agro-
meteorological requirements in mind. The very large number of instruments, that
have been developed for the measurement of dew or duration of leaf wetness (see
Nagel, 1962 and WMO-TN No. 55) indicates, that not even a moderately reliable

153
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

method has yet been found. That is unfortunate because this is extremely vital
information for the protection of crops against fungi and diseases (see WMO-TN
No. 192), which cannot be deduced usefully with rules-of-thumb such as 'RH>90
per cent'.
Porcelain plates (Leick), woodblocks (Duvdevani) and other artificial objects
can share in any dew-fall or distillation occurring on a given surface. However,
unless they are more or less flush with the surface and have similar physical prop-
erties (surface structure, heat capacity, etc.), they will not indicate reliably the
amount of dew which the surface receives. If exposed above the general level of
their surroundings, as is normal with Duvdevani blocks and usually appears to be
the case with more refined 'drosometer' devices, their behaviour will diverge from
that of the surface below, and the deposited dew may bear little relation to the dew
on adjacent natural surfaces. Weighing-type instruments, modified hygrographs
with a hemp thread instead of a hairbundle, and systems with surface electrodes
which connect when the surface is wet, all have their problems (Getz, 1992;
Bedson, 1992). Surface-electrode instruments are the simplest to interpret, but
again do not measure real leaf wetness - the sensor is again a fake leaf, with differ-
ent heat capacity. A pioneering new development is an optical sensor to detect
water on actual moving leaves (Heusinkveld, 1998).

9.2.7 Evaporation and Basis features of evaporation measurement at meteorological stations are
evapotranspiration given in GM-II (p. 267-274). In particular, the use of the Class A evaporation pan
is described. The alternative use of atmometers or lysimeters is discussed below.
References: GM II, Ch. XII;
The very large number of existing evaporation pan types (WMO-TN No. 83
WMO-No. 8, Chapter 8;
lists 27 types) indicates that none of them is really reliable for general application.
WMO-TN No. 83 (1966).
The Class A pan, however, is used as a reference: it is an agreed instrument
for purposes of comparison, and as such its use does not imply knowledge
about relations between its output and the 'true' value of evaporation. For compar-
ison purposes, repeatability will do. Calling the Class A pan a 'standard' would be
overrating it, since that would imply some certainty about the measurement.
However, it does have the distinction of being both inexpensive and easy to use.
Pans, in general, may give consistent estimates of potential evaporation in
conditions where advective effects are unimportant, particularly in arid localities
where they therefore will not give realistic results. A minor inconvenience in such
conditions is the attractiveness of the pan to thirsty animals. Howell, et al. (1983)
found that protective wire netting over the pan decreased the pan evaporation by
about 10 per cent, without really changing the correlation of pan evaporation
with calculated potential evaporation.

(a) Lysimeters · Instrumentation for lysimetry has developed gradually over many years and is
now reaching the stage where reliable equipment suitable for basic network use is
References: Molga (1962);
becoming available. A simple lysimeter measures only potential evapotranspira-
WMO-TN No. 83 (1966);
tion, and is not suited to the short-term (e.g. daily) measurement of transpiration
Rosenberg, et al. (1983).
from a crop suffering water stress.
Other more elaborate equipment produces a continuous record of change in
weight, providing data for research purposes. However, high capital and mainte-
nance costs mean that such lysimeters are not ideal for network purposes. More
detailed discussion is found in Slatyer and Mcilroy (1961), in WMO-TN No. 83
(1966), and briefly in Rosenberg, et al. (1983). It seems quite obvious that no stan-
dardized equipment is yet available. The following notes are restricted to a few
types of lysimeter which are often used, but not on a scale approaching network
requirements.
(i) Popoff Iysimeter:
This consists of a circular metal drum of 0.25 m diameter and 0.275 m
depth, which holds about 25 kg of soil. Three instruments are used at each
installation; two identical drums ·are set up at each site in pits with a
minimum wall clearance. The bottom of each drum is wire mesh and this
rests on the undisturbed soil at the pit bottom, permitting free drainage and
capillary action. Drums are weighed daily and interchanged. The third drum

154
CHAPTER 9 - AGROMETEOROLOGICAL INSTRUMENTS AND OBSERVATION

is mounted in a pit over a drainage collector. Thus the water balance is

assessed by measuring rainfall, drainage and change in weight of the soil
containers. This instrument is used a good deal in some countries.
(ii) Thornthwaite lysimeter:
This instrument is square, 2 x 2 m 2, and 0. 7 m deep and is filled with well-
mixed sandy loam. Water is applied to the crop either by sub-surface
irrigation to a fixed water level or by overhead irrigation in quantity suffi-
cient to give some percolation. Potential evapotranspiration is computed
from water added, less drainage. WMO recommended the adoption of this
instrument for the International Geophysical Year in 1957, and a sparse
network exists over parts of Western Europe. A sketch of this drainage lysime-
ter is given in Figure 9 .3.

Figure 9.3 A F C

A simple type of drainage

lysimeter
A Extension of drainage pipe for
cleaning
B Drainage pipe 38 mm 0

1_:,_
b b I
slope 2.4%
bb 6 mm drainage holes in the C
drainage pipe every 20 cm
C Valve and pipe or leak control 25m E
D Drainage tank D
E Graded glass tube
F Soil surface
G Gravel filter
1m

(iii) Weighing lysimeters:

Weighing lysimeters have been developed in a range of sizes and
degrees of accuracy. Generally speaking the more refined instruments, such
as those using the Morris balance principle, are not ideal for network
purposes, particularly in countries with a great range of climate and plant
communities.
(iv) Hydraulic lysimeter:
A soil chamber rests on one or more connected capsules filled with air,
water, oil or other suitable fluid. Changes in weight of the chamber
contents, for example, by evapo(transpi)ration, are reflected by change in
pressure exerted on the capsule. Observation of these changes is usually
made by manometer. The instrument is somewhat cheaper and easier to
maintain than the more accurate weighing types and would serve as a
network instrument. For routine measurement, i.e. in more or less constant
operation, a simple, robust but reasonably accurate instrument would be the
objective.

(b) Atmometers The Piche atmometer has a charm of its own - it is the cheapest instrument used
anywhere in meteorology. It consists of an inverted test tube with water and a bit
of blotting paper, and it tells you about evaporation! Many people do not realize
that, if one adds the cost of effort, the Piche may still not give value for money in
terms of usable results.
In atmometers, a measured supply of water is in contact with some porous
material (unglazed ceramics, filter paper) exposed to atmospheric influences, often
in a Stevenson screen (well-isolated from the evaporation-generating radiation).
Observed loss of water from the small reservoirs gives 'potential' rather than 'actual'
evaporation. However, field experience and elementary theoretical considerations
readily reveal that water loss from such devices is a function of individual design
(size, shape, surface and internal properties) and method of exposure (height
above ground and hence influence of wind-speed gradients, whether screened or
not). As a result these devices can only safely be regarded as 'probes', giving qual-
itative information or possibly semi-quantitative comparisons. Excessive

155
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

sensitivity of the Piche to wind (rather than to humidity) has already led to its
alternative use as an anemometer probe (Stigter and Uiso, 1981).

9.3 OBSERVATIONS OF The term 'state of the ground', couples qualitatively the moisture content in the
'STATE OF THE GROUND' surface layers and the mechanical properties of the soil. If we were able to measure
AND SOIL MOISTURE soil moisture well, or to estimate it using the water-budgeting techniques
described later, the broad descriptive 'state of the ground' observations might be
9.3.1 'State of the ground' obsolete. However, this type o'f observation may, in many cases, be the only data
available on soil moisture. They will be of limited value for analysis unless more
Reference: GM-II, Chapter XIV,
soil properties are .known.
(p. 281-283).
To show the importance of surface soil moisture (or surface wetness) in agri-
culture, a short review based on remarks nf Ventskevich (1961) is illustrative. He
discusses specifically how the soil moisture content affects the possibility of carry-
ing out certain field cultivations and operations with the maximum efficiency in
terms of ,power requirement, and distinguishes five states of humidity (i.e. soil
moisture status):
(i) Excessively moist (flowing state): corresponds to a (moisture content) greater
than the maximum capillary capacity. When all the capillary and "part of the
non-capillary pores are .filled with water, cohesion of the soil _particles disap-
pears and the soil is in a flowing state. Wheel-type tractors and cultivators
sink in, and it is imp0.&Sib'le to carry out any work;
(ii) Very moist (sticky state): when coarse pores me empty of water, soil particles
adhere to each other and the soil is in a sticky state. It adheres to wheels, and
great tractional force is needed to cultivate the soil;
(iii) Moist (soft plastic state): when there is still less moisture, coinciding approx-
imately with 'field capacity'; the soil is not sticky, but soft and plastic.
Cultivation can be done with maximum efficiency;
(iv) Slightly moist (hard plastic state): dryer still, the cohesion between soil parti-
cles increases and soil is no longer plastic. At a moisture content somewhat
higher than the wilting point, clays and loams 'cement';
(v) Dry soil (hard or friable state): when it is more dry, either the structure is like
dust, e.g. for sand, or it is so hard that enormous traction is needed to culti-
vate the soil. This state also applies if over-moistened soils are frozen to a
depth of a few centimetres or more.
An adverse combination of soil moisture levels and weather may not only
hinder field work, but also damage soil structure if cultivation is undertaken. A
particular hazard is that of 'poaching' the soil, i.e. the breakdown in soil crumb
structure due to working it at a too-high moisture content. The critical soil mois-
ture level seems to be field capacity, which is lower for heavy soils (e.g. clay) with
·excess water leading to breakdown.
-In higher latitudes, the <duration and depth of snow cover, whether or .not
the surface 'is frozen underneath the snow mantle, or whether or not there is a
layer of .ice on the snow surface, and the all-important question of the state of soil
during thaw (and subsequent runoff), are very relevant matters. Ventskevich
(1961) cites several topics: snow depth and soil temperature, deleterious effects of
snow on arable and pastoral husbandry - including the fact that "the formation
of ice crusts, or of ice layers one centimetre thick or more in a snow cover at least
eight centimetres thick, impedes the grazing of sheep and camels" (see also
section 4.5.2).
A 'crop weather scale' was formerly used at agricultural meteorological
statinns ·in the United Kingdom for ground dassification, but is now out of -use.
The observational schedule is reproduced below {Observer's Handbook, 1956; s-ee
also 3rd edition 19697 :.PP· '91-:92): ·

156
CHAPTER 9 - AGROMETEOROLOGICAL INSTRUMENTS AND OBSERVATION

Descriptive term Significance

0 Dry (Bare plot) The soil has a dry surface but is not frozen.
1 Wet (Bare plot) 'Wet' like 'dry' refers only to the surface of the soil, the condition of the
soil below the surface being irrelevant to both observations; a difference of colouring
usually indicates which term is applicable. 'Flooded' and 'covered with thawing
snow' are not included under this term (see scale numbers 2 and 6).
2 Flooded (Ground representative of the station). The term is used when water is standing in
pools over a considerable area although the prepared bare plot may not itself be
under water. The term refers only to water, and a frozen or partially frozen condition
is not described by it (see scale numbers 4 to 9).
3 Covered with (Ground representative of the station) The term indicates 'slush', which may be
thawing snow produced by the thawing of fallen snow, a fall of sleet or the mixing of snow with
flood water. Dry or nearly dry snow is described under scale numbers 7 to 9.
4 Frozen hard (Bare plot) This condition is easily recognized; it is produced by the freezing of
and dry moisture in the pore spaces of the soil. (A slippery layer of ice on the surface of the
soil is described by scale number 5).
5 Covered with ice (Ground representative of the station) This observation is made when the whole or
or glazed frost the greater part of the ground is covered with ice, but without snow. The layer of ice
may be produced by the freezing of flood water, or the alternate thawing and freez-
ing of snow; or the formation of glazed frost. The ground itself may or may not be
frozen, and the surface of the ice may be covered by water.
6 Partly covered (Ground representative of the station) The term may be used even when considerable
by snow or hail patches of the ground are free from snow or hail.
7 Covered with (Ground representative of the station) The three terms for snow cover refer to snow
snow less than which is either not thawing at all, or is thawing only slightly. The depth of snow
15 cm deep, denotes the total thickness down to the ground, including any old or icy layers and
ground not frozen layers and not only freshly fallen snow. It may be measured on the bare plot;
provided that the covering there is uniform and representative of the locality.
8 Covered with It should not be measured in places where drifting has brought about
snow less than unrepresentative conditions.
15 cm deep,
ground frozen
9 Covered with If depth varies considerably within short distances, measurements should be made
snow more than at a number of points and the mean value determined.
15 cm deep
Fresh snowfall For this supplementary observation, a mountain is provisionally defined as ground
(hyphen) on the mountains exceeding 610 m above M.S.L. The observation is additional to one of the above
represented by the scale numbers 0 to 9.

As will be seen, greater attention was paid to the all-important phenomena

of snow and snow depth, and a distinction was made between superficial and tran-
sient moisture and moist soil proper. These scales for the 'state of the ground' do
not provide for the occurrence of hoar frost nor dew. When these conditions are
observed, the 'state of the ground' should be determined independently, as
though they were absent. They do in fact affect the ground surface, at least
temporarily, but are regarded primarily as 'weather observations', and are
recorded by means of coded letters in the descriptions of present weather and past
weather.
In actual codes, situations with (possible) snow and/or ice are separated
from the general scale. WMO Code table 0901 on the 'state of the ground' is as
follows:

157
WMO LECTURE NOTES - AGRICULTIJRAL METEOROLOGY

Code figure State of ground

0 Surface of ground dry (without cracks and no appreciable
amount of dust or loose sand)
1 Surface of ground moist
2 Surface of ground wet (standing water or large pools on surface)
3 Flooded
4 Surface of ground frozen
5 Glaze on ground
6 Loose dry dust or sand not covering ground completely
7 Thin cover of loose dry dust or sand covering ground completely
8 Moderate or thick cover of loose dry dust or sand covering
ground completely
9 Extremely dry with cracks
However, if there is snow or a measurable ice cover, WMO Code table 0975
is used; this code table shares some similar features to the 'crop weather scale'
above. Full descriptions can be found in GM-II.
We may conclude that, on the one hand, knowledge of the 'state of the
ground' is of major importance for agriculture - but that, on the other hand, its
scalar description is only descriptive, and not a very favourable starting point for
(agro)meteorological work.

9.3.2 Soil moisture For measurement and assessment of soil moisture, a WMO working group recom-
mended four techniques as useful for practical application (WMO-TN No. 97,
1969):
(a) The 'classic' direct gravimetric method;
(b) Resistance blocks;
(c) Tensiometers;
(d) Neutron-scattering method - although at present it is the most expensive
and requires skilled labour and maintenance facilities.
A fifth technique has since been developed, and is know as:
(e) Time domain reflectometry.
A detailed discussion of most soil moisture techniques has been presented by Black
(1965) and these have also been reviewed by Gardner (1968).

(a) Direct gravimetric method This method involves drying a sample of soil, weighing before and after the
drying and expressing the moisture content in terms of the percentage of water
held by a unit mass of dry soil.
However, as concerns the saturation of pore space with moisture the water
·capacity is more readily visualized if expressed on a volume basis, the 'volumetric
water content', i.e. in terms of the equivalent depth of water per unit depth of soil.
Although in principle the 'volumetric water content' can be determined directly
in terms of grams of water per cm 3 of volume extracted (Coile, 1936), this gener-
ally involves considerable difficulty when samples at depth are required, and it is
more common to determine bulk density separately and then to obtain volumet-
ric water content estimates from gravimetric samples.
Baver (1948) points out several advantages in the volumetric expression of
soil water: "In the first place it helps to picture the relationship of soil to air and
water. A given soil may be visualized as possessing a certain total porosity which
is filled partly with water and partly with air. Moreover, it is possible for two soils
to have the same moisture content on a percentage-by-weight basis, but not on
the volume basis. Second, water retention in the soil mass is within a given
volume, not a given weight. The roots of vegetation obtain water from the volume
of soil through which they penetrate."
Following Slatyer and Macllroy (1961), suppose the moisture profile of a
column of soil one metre in depth is required. The column is divided into strata,
e.g. successive intervals might be 10, 10, 20, 30, 30 cm and samples are taken at
the midpoint of each stratum, i.e. at levels of 5, 15, 30, 55, and 85 cm. From any
sample, w grams of water would be extracted from s grams of soil:

158
CHAPTER 9 - AGROMETEOROLOGICAL INSTRUMENTS AND OBSERVATION

(weight before drying) - (weight after drying)

Gravimetric moisture content (w8) = - - - - - - - - - - - - - - - -
Weight after drying

i.e. w8 = [(s + w-s)/s] x 100 or, w8 = (w/s) x 100

Dry weight .of soil in depth interval

The bulk density is defined as: d =
Volume of soil in depth interval

If total water in the profile is W, then:

W = 10 (w8 x d)s + 10 (w8 x d)i 5 + 20 (w8 x d)zs + 30 (w8 x d)s 5 + 30 (w8 x d) 85

= 10 (wv>s + 10 (wvhs + 20 (wvhs + 30 (wv>ss + 30 (wv) 85

where (wv)z = (w8 x d) 2 at z = 5, 15, 25, 55, 85 cm depth,

(wv)z being the volumetric water content in the column sampled at z
Some guidance on practical procedures follows. Several types of soil auger are
available commercially for sample extraction. Although most are generally satis-
factory, differences in design or in dimensions can involve appreciably different
degrees of physical effort, and can produce samples of significantly different
quality. For sampling within about 0.2 m of the soil surface, however, a simple
tube sampler is usually quick and sufficiently accurate, enabling samples of known
volume to be readily obtained in the zone of greatest soil water variability. A stain-
less steel tube of about 3 cm in diameter and about 35 cm long is quite suitable
for most purposes. Veihmeyer tubes provide more elaborate facilities for this type
of sampling, enabling tube samples to be taken at much deeper levels. Special
heads to be used with this equipment can be produced for best results in either
sand or clay soils, in which sample extraction is sometimes difficult. Care must be
taken in the choice and operation of equipment, so that soil compression during
sampling is minimized.

(i) Gravimetric determinations: Remarks: Given the high degree of soil water variability, large numbers of
samples are required to give reliable average values.
Apparatus: Soil tube or auger; airtight sample cans; balance to weigh to
+0.01 g; drying oven.
Procedure: Using a tube or auger, collect 50-100 gram samples at the
required depth from the location and rapidly eject into sample
cans. Weight the cans to give fresh weight, dry the contents at
10S°C constant and re-weigh to give dry weight.
Calculations: Gravimetric water content is calculated as loss of weight on drying
divided by dry weight, and is usually expressed as a percentage.

(ii) Volumetric determinations: If a soil pit is available, direct sampling can again be used, either with Coile's
method or by breaking off clods and determining their volume and dry weight.
Volume is measured by coating the clods with a waterproofing solution and meas-
uring their apparent loss of weight suspended in water. For most
microclimatological studies, however, a number of sites are needed and the labour
required for pit excavations becomes excessive. In consequence the most
commonly used methods involve either direct or indirect determinations in auger
holes.
Apparatus: Soil auger; rule; soil sample bags and cans; quantity of graded
sand of uniform density, or calipers for measuring hole diameter.
Procedure: Excavate a hole with a soil tube or auger to desired depth inter-
vals. From each interval collect the soil into bags and measure
the depth. Weigh it and place two or three sub-samples (about
SO g fresh weight) into cans and weigh them. Continue the
procedure until the hole reaches the maximum desired depth.
Refill it with sand interval by interval, noting the volume
required in each case. The volume added can be measured

159
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

directly if poured from a graduated cylinder or by weighing the

container before and after filling each interval. The dry weight of
the sub-samples of soil is then obtained, as already described.
Calculations: The dry weight of the soil in each depth interval is estimated as:

(Fresh weight of soil in depth interval) x (Dry weight of sub-samples)

Fresh weight of sub-samples

Dry weight of soil in depth interval

. Bulk density is then defined as:
Volume of soil in depth interval

and is usually expressed as g cnr3 dry soil.

As an alternative, the volume of each interval can be determined by meas-
uring the auger hole diameter with calipers, but Tisdall (1951) considers this
method less accurate.

(b) Soil moisture determination This technique depends on the relationship between soil water tension or soil
from the electrical conductivity water content and the electrical conductance between the electrodes of a buried
of porous blocks porous block. Although the method provides an indirect measurement of soil
water tension rather than water content, it has been used extensively for soil water
storage estimates. In that case the accuracy finally rests on a comparison with
some other method which directly determines soil moisture content, usually the
gravimetric method.
Each block normally consists of a pair of electrodes surrounded by a porous
matrix, generally composed of gypsum, fibreglass or nylon. Units containing fibre-
glass or nylon are sensitive over a wider range of soil water tension, particularly in
wet soils. They also have an indefinite life, whereas the life of gypsum units is
limited.by the time taken for a significant amount of gypsum to dissolve.
The method is not as precise as the neutron-dispersion technique, but can
give satisfactory results with careful use in appropriate situations. With a well-
designed installation, embodying a central switchboard for each major group of
units, soil compaction by trampling is eliminated, and it is not difficult to take
over 150 readings per hour. Furthermore, by running the leads underground there
is minimum interference with agronomic operations.
The effective range of soil-water tensions which can be measured is approx-
imately from field capacity (pF = 2.0, about 1 m water column) up to the wilting
point, which for most plants lies near pF = 4.2 (1 SO m water column). Gypsum
blocks are usually considered unreliable at water tensions of less than 5 m. Nylon
and fibreglass units are normally sensitive down to tensions below 1 m .
The calibration of block resistance against soil water content is best
conducted in the field during the course of the experiment. The technique is quite
simple and similar to that for the neutron-scattering technique for which the same
remarks apply. It merely involves gravimetric soil sampling at times when the
blocks at any one depth are providing uniform readings. Because of hysteresis
effects it is desirable to confine the calibration to the drying curve only, since this
governs water status when it is of greatest significance to plant growth.
The drift in the calibration curve of gypsum blocks, presumably due to the
gradual solution of gypsum with associated change in the pore space geometry,
requires that calibrations should be made periodically so that the magnitude of
the drift does not introduce significant errors to data interpretation.
Since a fixed arrangement must be adhered to (the procedure is not flexible,
e.g. the shallowest porous block must be below the range of cultivation imple-
ments), the decision on the vertical spacing must be carefully considered. Some
practical details follow below.
Apparatus: Conductivity units of desired type; measuring instruments; cable
and switchboard equipment.
Installation: The units are conveniently installed in auger holes dug with a
Jarrett auger~ To avoid unnecessary labour, up to three blocks can
be installed at different depths in any one auger hole.

160
CHAPTER 9 - AGROMETEOROLOGICAL INSTRUMENTS AND OBSERVATION

Prior to installation, a plan should be drawn up, so that the block layout is arranged
as efficiently as possible. Each hole should then be excavated to the depth where the
deepest unit will be placed. The soil from each auger extraction should be kept sepa-
rated. The deepest unit is then lowered to the base of the hole and pressed firmly to
ensure good soil-unit contact. The soil is then returned to the hole in the reverse ·
order in which it was extracted, and pressed firmly into position until the next
depth is reached. This procedure is repeated until the soil reaches slightly below
depth of the plough where the leads are then carried away in a trench, excavated to
the same depth, to the appropriate switchboard. During the refilling of any hole,
care must be exercised so that the soil occupies its original volume. The observer
should obtain practice with one or two 'dummy' holes before commencing the
experimental ones. After the trench is refilled the experiment is ready to commence.
When convenient, the wires can be left buried at the switchboard sites so that culti-
vation can proceed unimpeded. The switchboards can then be wired after sowing.

(c) Measurement of soil The following field methods are available for the measurement of this soil
moisture tension characteristic.
(i) Resistance blocks: Indirect measurements of soil moisture tension can be
obtained from resistance blocks. In general, gypsum blocks are preferred for
this purpose because of their greater uniformity and lower hysteresis error,
although they are insensitive at tensions less than 3 m. For lower tensions,
standard fibreglass or nylon units may be employed, although combination
units are preferable. If the widest possible range is the criterion for selection,
fibre-gypsum combination units are the best;
(ii) Tensiometers: Direct field measurements of soil moisture tension in the
range up to 1 atm (pF = 3.0, about 10 m water column) can be effectively
Figure 9.4 made with tensiometers - tensions exerted by most soils at field capacity
A simple tensiometer system. and at the irrigation-requiring point lie within this range. These instruments
consist simply of a porous ceramic cup buried in the soil and connected by
Stopp water-filled tubing to a manometer or vacuum gauge on which negative pres-
Vacuum sure can be indicated. A device for filling with water and air removal is
attached to this (Figure 9.4). In practice, different sizes and shapes of porous
cups and the other components are employed. It is the most accurate device
for measuring water tensions of less than about 8 m.
Since the instruments are direct reading, no calibration is necessary. Though
of very limited value in naturally arid conditions because of the restricted tension
range in which they operate, tensiometers can be of considerable value in irriga-
tion studies. In such studies diurnal fluctuations in tensiometer readings
frequently occur due to the development of temperature gradients between
porous cup and soil. These effects are minimized if plastic components are substi-
tuted for metal, and the variation is also much smaller in fine rather than in
coarse-textured soils. Alternatively, readings can be made at the same time each
day, preferably early in the morning when soil temperature gradients are likely to
be least steep.
Preferential root growth around the tensiometer cup can also make meas-
urements unreliable. Care should be taken to check on this point, or periodically
to change the location of each instrument.

(d) Determination of soil water This method measures soil water content by counting slow neutrons in the vicin-
content by the neutron- ity of a source of fast neutrons because hydrogen atoms are particularly effective
scattering method in slowing down and scattering neutrons. This is a good measure of water because
virtually all the hydrogen in soil is found in the soil water. Even in soils with high
Reference: Greacen (1981).
root density, or high levels of organic residues, the amount of organic hydrogen
is generally small compared with that in the soil water.
Both this technique and the conductivity method involve fixed installations
so that repeated observations can be made at any one point. This gives less
sampling error than auger determinations.
However, there is a major difference. In the neutron-scattering method,
access tubes are placed in the soil so that the neutron probe can be lowered to the

161
WMO LECTIJRE NOTES - AGRICULTIJRAL METEOROLOGY

desired depth each time a measurement is made. In the conductivity method, the
individual units are usually placed permanently into pre-determined positions in
auger holes. In establishing the layout of installations of either type, several points
should be considered. It is desirable to have at least four, and preferably six, repli-
cates of positions for each agronomic or other treatment involved. The vertical
spacing of the measurements depend on the experiment under consideration, but
for soil water storage purposes it is often helpful to have an arithmetic series so
that volumetric estimates of profile water content are simplified. However, root
distribution is usually concentrated in the top 0.5 m of soil and closer block meas-
urement is required in this zone.
The deepest measurement point should be at a point beyond which it is
expected that neither root growth nor water will penetrate. With the neutron
method loss of neutrons through the soil surface may occur with measurement
points shallower than 0.2 m.
Apparatus: Neutron moisture meter with accessory equipment; access tubes
constructed of _d rawn polythene tubing closed and pointed at
one end; covers for open tops of access tubes.
-Installation: The access tubes are usually 5-6 cm in diameter and can be
conveniently installed in holes drilled with augers of the same
diameter. The holes should be carefully drilled with a minimum
of 'reaming out'. The walls of the hole should be slightly wetted
to reduce friction during tube installations. This can be achieved
with as little as 50-100 ml of water if a small can with several
punctures around the lower rim is raised and lowered inside the
hole. The tube can then be forced down the hole, if necessary by
hammering a rod inserted in it. Approximately 10 cm of tube
should project above the ground. Between measurements this
can be covered with a can to prevent water entry.
Calibration: Holmes (1958) has derived a theoretical calibration curve relat-
ing slow neutron counts to volumetric soil water content and
Holmes and Turner (1958) have presented a calibration curve
based on data from samples of sand, of loam, and of clay, with
pure water.
It has been found that the constitution of the soil solids may affect the
counting rate, and it is consequently necessary to obtain calibration points for
individual soils. Elements such as boron, chlorine, iron and some of the rare earths
are suggested as those which, when present in atypical abundance, may change
the calibration relationship sufficiently to merit individual calibration.
Calibration occasions should be selected to cover a range of soil water
·contents, and occasions selected when the readings at any one depth indicate
uniformity of soil water content at that depth. In practice, although it is quite easy
to obtain uniform block readings in fairly wet or dry soil, it is often difficult to
obtain intermediate values because differential root distribution results in differ-
ent rates of soil water extraction. The number of gravimetric samples required for
each calibration occasion can be calculated and should be adequate to provide a
mean value accurate to not more than + 0.2 per cent water content (g water/100
g oven-dry soil). On each occasion a number of volumetric determinations of soil
water content should be made, in addition to a set of neutron readings. The
number of volumetric determinations should be adequate to reduce sampling
error to tolerable levels. Once sufficient points have been obtained, corrected
meter readings can be readily converted to volumetric soil water content.
Regular readings: Manufacturers include instructions on the recommended
operational procedure. In addition, it is most important that
the soil around each tube should not be disturbed in any way
that might affect plant growth and water entry or extraction.
Some experiments have shown the importance of conducting all measure-
ments from a small platform elevated a few inches above the soil surface. This is
particularly desirable in experiments with natural or sown pastures where cultivation
is infrequent. In crop experiments it has generally been found satisfactory for the

162
CHAPTER 9 - AGROMETEOROLOGICAL INSTRUMENfS AND OBSERVATION

observer to stand on a small platform placed on the soil surface as far as possible from
the tube, while being able to raise and lower the probe to desired depths.
If soil water determinations are being made with a neutron moisture meter,
the same scaler can also be used to obtain density estimates if a different probe,
incorporating a gamma source (usually caesium 137) is used instead of the normal
fast neutron source. This commercially available equipment uses gamma rays to
interact with the orbital electrons of atoms, the resulting energy of these rays is
reduced proportionally with the number of electrons per unit volume of soil. This
can be related to the density of any material of known composition. Therefore
density estimates can be made effectively with this technique once the counts
have been calibrated against soils of known density. Density can subsequently be
determined from the counting rate and the calibration chart.

(e) Time domain reflectometry Next to neutron-scattering, another 'remote-sensing' non-destructive method for
(TDR) determining soil water content is the measurement of the dielectric constant of
soil. The propagation speed v of electromagnetic waves in a medium is according
to Maxwell's laws equal to v = c/✓(Eµ), where c is the speed of light, and the
medium is specified by its dielectric constant£ and its magnetic permeabilityµ.
The magnetic properties of nearly all geological materials, including water, do not
vary significantly from those of vacuum, so in practiceµ= 1. On the other hand,
for water £ = 81 , while for nearly all geological materials £<5. It was therefore
found (Topp, et al., 1980) that for many soils with volumetric water content 9v the
dielectric constant is given by:
2 3
£ = 3.03 + 9.3 9v + 146 9v - 77 9v

This is employed to obtain 9v by measuring in the soil the propagation speed of

alternating current pulses of very high frequency (>300 MHz). Pulses are reflected
at inhomogeneities, either in the soil or at the probe-soil interface, and the travel
time between reflections is measured (Alharthi and Lange, 1987; Nadler, et al.,
1991). From the travel time the volumetric water content of the soil can be deter-
mined. Equipment for this method is now available commercially.
Each of the five methods outlined above have their advantages as well as
their limitations. For guidance on practical use of classical methods to measure
soil moisture, the reader is advised to consult WMO-TN No. 97. Also, a brief
summary and comparison of the first four methods was outlined by Slatyer and
Mcilroy (1961) and is reproduced overleaf (Table 9.2). The fifth method, reflec-
tometry, has practical properties analogous to neutron scattering such as the large
range of water content, but the spatial resolution is better, calibration require-
ments are less severe, and the cost is lower.

9.4 BIOLOGICAL A complete introduction to this subject gives compendium AG-T. The WMO Guide
OBSERVATIONS to Agricultural Meteorological Practices (WMO-No. 134) sets out some recommenda-
tions regarding biological observations, and these are summarized in the
following section.
Biological observations are, at least, a qualitative and, at best, a quantitative
measure of the responses of a plant or animal to weather conditions. They are
made on living material, which may have inherent individual differences; they do
not have the advantage of the uniform behaviour of a standardized meteorologi-
cal instrument. Therefore each user of biological observations must always
remember the limitations to their general applicability. As a general principle,
however, it is essential that the accuracy and extent of the biological observations
should be matched by the accuracy and extent of the meteorological observations
with which they are to be associated.
The biological observations can be conveniently divided into five categories:
(i) Detailed observations of high accuracy or considerable complexity (see e.g.
Idso, et al. (1988) on the measurement of stomata! conductance), required
during a specific experiment at a research station or experimental site;
(ii) Observations of a less complex character, taken in a larger geographical area in
a network with a large number of sites, which are required for operational use

163
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

Table 9.2
SUMMARY AND COMPARISON OF METHODS FOR SOIL WATER MEASUREMENT
(Slatyer and Mcllroy, 1961)

Method Main advantages Main disadvantages Best application

1. Gravimetric auger • Simplest and most • Repeated measurement • Can provide accurate
or tube sampling reliable technique (used at one point impossible, result in most situations,
as a calibration standard). so sampling error high. provided sampling is
• Rapid sampling close to • Physically tiring. adequate.
the surface (0.2 m), • Relatively slow sampling
although the subsequent at greater depths.
processing takes time.

2. Conductivity block • Rapid reading. • Affected by salinity • Most useful for culti-
• With appropriate installation, (particularly fabric units). vated non-swelling soils,
cultivation can be unimpeded. • Limited range of soil water preferably free from
• Suitable for repeated measure- covered (through usually protracted wet or dry
ments at a point. adequate). periods.
• Relative simplicity and • Significant time lag of block
robustness tension behind soil water
tension (more pronounced
in gypsum units).
• Short life of gypsum units
in wet soils (although
indefinite in fabric units).
• Soil-block contact affected
by swelling soils.
• Provides only indirect water
content measure, needs
calibration per soil.
• Calibration drifts with time.

3. Tensiometer • Only instrument which ·• Useful range only up to • Not used successfully
provides direct measurement about 8.5 m tension. for evaporation
of tension. • Needs to be recharged estimates except
• Well suited to calculation when soil dries out, so possibly in moist soils,
of profile permeability unsuited to permanent e.g. frequently
characteristics. deep installation. irrigated area.
• Projecting instrument head
prevents cultivation.
• Preferential root growth
(with water extraction)
occurs near porous cups.

4. Neutron scattering • Full soil water range covered. • High cost and fragility of • Considered to be the
• Volumetric estimates (density equipment. best method available at
correction not necessary after • Large volume sampled, present (i.e. in 1961) for
initial calibration). so unsuitable for point quantitative estimation
• Large volume of soil sampled, measurement or near surface of changes in soil water
smooths out small-scale measurement. storage.
variability. • Need for frequent access
• Permanently installed access causes trampling and soil
tubes enable repeated measure- compaction hazard.
ments at same point. • Projecting access tubes
• No lag of measured, compared restrict cultivation.
with actual water. • In swelling soils, access
• Readings unaffected by salinity. tubes may start cracks.
• Absolute readings affected
by high amounts of organic
matter, Cl or Fe.

164
CHAPTER 9 - AGROMETEOROLOGICAL INSTRUMENTS AND OBSERVATION

or administrative action within a short period of time after the observation has
been made - in other words, observations for immediate operational use;
(iii) A similar network of observations on cultivated plants, farm animals and
farming operations, e.g. ploughing, sowing, harvesting. These data are
required for a detailed assessment of work on the land and of the progress of
the farming year. Such observations are required for long-term analysis, not
for immediate operational use;
(iv) A similar network of observations of natural phenomena, dealing with wild
plants, animals, birds and insects;
(v) Assessment of damage to cultivated crops or animals due to meteorological
causes.

9.4.1 Observations for Research-type observations require skilled, scientifically-trained staff and it is
research and for therefore appropriate to the work carried out at principal agricultural meteoro-
operational use logical stations. Work of this nature may be carried out either under natural
conditions in the field; or in laboratory experiments which can often involve the
use of specialized scientific laboratory equipment; or in climate control chambers,
in which it is possible to study the reaction of both plants and animals to a single
meteorological factor.
In all cases it is important to measure both physical and physiological
responses in the living organisms, such as basal metabolism, pulse and respiration
rate, body temperature of animals, carbon dioxide intake, leaf and fruit tempera-
ture, osmotic pressure and the chemical constitution of plants. Care must be taken
in animal experiments because it is not always possible to obtain representative
responses from small groups in a limited series of experiments.
Observations for operational use are required for administrative decision and
action, for making functional forecasts, or for purposes of technical advice. They
are of a less complex character than for the first category, but they must be of a
standard nature so that observations from different sources can be comparable.
Such observations are obtained from a large number of sites, which form a
national network, and must be made by skilled or semi-skilled observers who have
received adequate training to enable them to attain the desired standard. The
density of the network may be limited by the availability of efficient observing
staff. For any one operational requirement, all areas of the country which are
concerned with that particular problem should be adequately covered by the
network. Ideally, the density of the network will depend on the type of problem,
the distribution of the crop and the variation in soil and climate over the region.
A common aim is to have one observer for every SO km 2, provided that he is able
to visit areas in his immediate locality.
Some examples of operational data use were mentioned in previous sections,
these include:
Forest fires: Information is required on the state of forest litter and its
liability to fire.
Diseases: Information is required on the state of plants, the availabil-
ity and release of spores, and the incidence and spread of
infection.
Pests: Information is required on the hatching of harmful insects,
the build-up of populations of insects and other destructive
animals, or their invasion from other territories.
Weather hazards: Information is required on the state of crops, and whether
they are in such a stage of development as to be particularly
liable to weather hazards such as frost.
Farming operations: Knowledge of the progress made in the farming year is
needed by weather forecasters so that they may be aware of
the operational implications of forthcoming weather.
Under this heading are included observations on yield, which may be
concerned with small experimental plots or with regional or national production.
In this type of observation the meteorologist should always act in collaboration
with statisticians and agricultural experts. Where necessary, he should

165
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

encourage the agricultural authorities to obtain the data in a suitable form. In the
case of regional or national yields he should be aware of the degree of accuracy of
the yield.
The analysis of meteorological and biological data is carried out simultane-
ously by the regional or central authority. Operational decisions are based on such
analysis, and a summary of the season's work should be prepared and either
published or permanently retained for reference, so that the experience of any one
year is always available for subsequent consideration.

9.4.2 Observations of This category includes the phenological observations on cultivations, on culti-
natural phenomena for vated crops and trees, farm animals, and general activities on the land which are
agroclimatological use required to form an accurate picture of the agricultural year. Observations which
concern wild plants and animals, for the most part free from deliberate human
interference, may also be included. Because of this relative freedom, the pheno-
logical data provide a form of integration of local climate, and as such may be
useful as operational parameters.
The observations are made on a wide selection of phenomena by a perma-
nent network of reporting stations. The network can be less dense than that
required for operational purposes, but should cover the entire country and not be
confined to a smaller area of specialized agricultural production. Observations can
be simpler in character than those previously specified, but must follow an agreed
and fully understood standard. They often consist of the recording of the dates on
which certain events took place and should describe any occurence of weather
phenomena which may cause damage to plants and animals. Any occurrences of
weather which may cause damage to plants and animals should be recorded,
including occurrences of: snow, ice, frost, hail, heavy rain, drought, weather condi-
tions leading to serious air pollution, unseasonable heat or cold, strong winds and
floods, sand- and dust-storms, and sea-salt damage in maritime areas. The obser-
vations are subsequently analysed, published or otherwise placed on permanent
record by a central authority, but there is no degree of operational urgency.
Although each country makes its own selection of types of observation, it is
clearly desirable that it should act in conformity with neighbouring countries as
far as possible. With this end in view, consultation between adjacent countries is
necessary. A uniform system of specifying geographical positions of observation
sites is desirable.

9.4.3 Specific examples of A few practical examples may be useful to illustrate the general principles of
biological/phenological biological and phenological observations. In practice, this type of observation will
observations not be done routinely by meteorologists, but it is useful to know the kind of growth
·stages which are observed. For this purpose, wheat and maize have been chosen.
Additionally, a short example is given of modelling the reproductive devel-
opment as a ~nction of actual weather for operational purposes, such as timing
of crop protection, or irrigation activity. Degree-days (see section 8.2) are used to
calculate the total amount of heat required to reach a given phenological stage.
Generally, such a simple model will be site-specific - but it can be practically
quite useful.

9.4.3.1 Wheat The following eight main phases of wheat development may be considered repre-
sentative of the life cycle of the wheat crop - sowing date, sprouting, tillering,
stem extensions, earing; flowering, ripening and harvesting:
(i) Sowing date - note the first and last date of sowing;
(ii) Sprouting phase - the appearance of seedlings is the first observed stage
of the cycle because germination takes place under ground;
Start: When the appearance of seedlings is judged to be taking place
without interruption and at an increasing rate;
End: The occurrence of the last seedlings in uninterrupted succession;
(iii) Tillering phase - this is marked by the appearance of lateral buds, or by
that of the fourth leaf which takes place almost simultaneously and is
easier to see;

166
CHAPTER 9 - AGROMETEOROLOGICAL INSTRUMENfS AND OBSERVATION

Start: When it is judged that the number of plants having their fourth leaf
is sufficient to imply that the process is uninterrupted.
(iv) Stem extension phase - this is a concealed phase which consists in the
emergence above ground of the flower bud at the upper end of the future ear-
bearing stem. This is the first manifestation of the flowering process which
can be observed microscopically directly in the field. The observational tech-
nique is to cut the seedling at ground level and make a longitudinal slit with
a finger nail or a pen-knife. If all that is found are folded leaves, the plant is
still in the vegetative stage. If, on the contrary, it is possible to detect an
internode and the floral bud is visible above the closely juxtaposed nodes,
the plant has begun to form its stem. At one centimetre above ground level,
the flower bud is between one and two centimetres long and is visible to the
naked eye. Such observation of the stem extension requires some practice
but, when one is accustomed to it, it is no longer necessary to cut and split
the shoot but simply to feel its base; if the shoot feels hollow when it is
lightly squeezed at ground level, the plant has begun to form its stem.
Start: Since observation is difficult, one simply records the date when the
flower shoot can be observed on the majority of the plants
inspected.
(v) Earing phase - observations noted when the ears free themselves from
the last of the leaves enfolding them and protrude above the plant.
Start
Climax } According to the criterion followed for
End recording phase dates in dense crops;
(vi) Flowering phase - flowering, which follows immediately on heading,
means the appearance of the anthers beyond the spikelets after fertilization
has taken place inside them (wheat being a cleistogamous plant).
Start } According to the criterion followed for
End recording phase dates in dense crops;
(vii) Ripening - the consequence of fertilization is the formation of the grain,
which increases in size with successive changes in its water content and
chemical composition, between green ripeness and hard ripeness. Three
stages of ripeness can be usefully distinguished:
• Start of milky ripeness: when the grains in the ears are squeezed
between fingers they release a whitish fluid. This sub-phase takes place
at the time when the wheat plant first begins to turn yellow;
• Start of waxy ripeness: when squeezed between the fingers the grains
should be malleable like wax;
• Start of full ripeness: the grains have lost their water content when they
can be split by fingernail.
(viii) Harvest
Where necessary, observations may be confined to merely sowing, the start of sprouting,
the start of stem extension, the start of heading and the end of full ripeness. Although
the suggested record for the phenological observation of wheat is based on six
phenological stages, Large (see Peterson, 1965) suggested four phenological stages:
tillering, stem extension, heading and ripening. However, each of the first three stages
was sub-divided into a number of sub-stages. A sketch of 'growth stages in cereals' as
suggested by Large, is presented in Figure 9.5.

9.4.3.2 Maize The following eight main phases of maize development may be considered repre-
sentative of the life cycle of the maize crop - sowing date; sprouting, tillering,
incipient cobbing, pollination, earing, ripeness and harvesting:
(i) Sowing date - the first and last date of sowing to be noted.
(ii) Sprouting phase:
Start } with sparse crops only a subjective estimate of
End the intensity of sprouting can be formulated.
(iii) Tillering phase - this requires observation of lateral shoot formation,
which is an important point for agronomical purposes. Since not all plants
of a tillering stand actually form lateral shoots, the possession of lateral

167
WMO LECTIJRE NOTES - AGRICULTIJRAL METEOROLOGY

Figure 9.5 Stem extension - - - - Heading Ripening

stage
Development of cereals,
STAGE
(Feekes-Large scale). STAGE STAGE
1~·1 11
10.5
fir st flowering
spikelet wheat
of head
just
visible
. . - - - - - - Tillering ------1
STAGE ligule of
last leaf
8
STAGE last leaf vi~~~e
7
v/~~~e
STAGE
6
STAGE
STAGE 5

.~:,tl
STAGE 4 leaf-
STAGE 3 leaf- ·sheaths
STAGE 2 tillers sheaths strongly
1 tillering formed lengthen erected

~ - '

- --~~~~
~~...:::....,,:--=-'-

shoots by 20 per cent of the plants can be taken as a limit, and it may be
indicated whether the variety has more or less than this percentage.
(iv) Incipient cobbing - the cob, which is still in the stem, is just visible
peeping out of the leaves surrounding it without any attempt being
required to separate them.
Start: when the cob begins to push its way out of the leaf surrounding
it in 20 per cent of the plants. It is not necessary for the cob to
protrude above the plant at this time - it can be counted as an
incipient cob if it can be seen, even when looking at the plant
from. above.
Climax: when 50 per cent of the plants have incipient cobs.
(v) Pollination - this commences when the air begins to contain pollen
from the plants under observation and when it is possible for it to adhere
to neighbouring stigmas. This phase occurs some 7 to 12 days after incipi-
ent cobbing depending on the variety, the sowing time and the weather
during the year in question.
Start: when 20 per cent } of the cobs of the plants are releasing pollen,
End: when 80 per cent as can be seen by shaking them.
(vi) Earing phase - the emergence of stigmas usually takes place between 2
and 5 days after the start of pollination,
Start: when 20 per cent } of the plants show stigmas protruding from
End: when 80 per cent the future cobs.
(vii) Ripeness
Start of milky ripeness: when, in 20 per cent of the plants, the grains release
a whitish fluid when squeezed. There is some correspondence between this
sub-phase of maturity and the bracts turning yellow. It is therefore possible
to avoid breaking grains by counting 20 per cent of the plants whose ears
are beginning to turn yellow;
End of full ripeness: when, in 80 per cent of the plants, the grains in the
ears cannot be split by a finger nail;
(viii) Harvest
For less detailed phenological observations of maize one can record only the dates
of the following phases : sowing, start of sprouting, 20 per cent of plants with
incipient cobs, 20 per cent of plants releasing pollen, 20 per cent of plants with
visible stigmas, and 80 per cent of plants with full ripeness.

9.4.3.3 Avocado Davenport (1982) has broken the life ·cycle of the avocado crop growing season
into eight stages (see Table 9 .3). Climatic fluctuations during this period play a
dominant role in the achievement of final yield, and microclimatic manipulations

168
CHAPTER 9 - AGROMETEOROLOGICAL INSTRUMENTS AND OBSERVATION

during short critical periods can help in improving yield. Growth stages were then
compared with degree-day sums, using 12°C as base temperature, i.e. adding from
each day:
[(Tmax + Tmin)/2] - 12°C
and calculating regression of log-normals of this sum and Davenport stage
numbers (Table 9.3). This was then validated against independent new observa-
tions in a glasshouse (at much higher temperatures) and in a field plantation, with
the following result (Lomas, 1997).
Table 9.3
Davenport Development Temperature sum value
stage (Fuerte avocado tree) Model output Glasshouse Field
1 Buds unaffected
2 Swollen flower buds
3 Inflorescence extension 25-35 35 49
4 Beginning of open flowering 40-50 90
5 Full open flowering 65-85 100
6 End of open flowering 100-150 150 194
7 Beginning of fruit set 220-300 280
8 Fruit set 350-400 370 350

If we make a graph of the results, it is obvious that the model performs well
for the last stages (7,8), but not so well for the middle stages (4,5,6). This means
that one of two things has occurred:
• The validation data were insufficiently compatible with the model range (e.g.
different site climate or a year with anomalous weather); or,
• The real relation between phenological avocado development and the
temperature sum defined here is insufficiently linear (see section 8.2).
In conclusion, our predictive knowledge of avocado phenology has
increased with this experiment, but additional work of this type is required to
remove uncertainties.
This investigation is typical of much weather-related research on develop-
ment. The usual purpose is that by tracking accessible weather parameters it
becomes possible to estimate in advance when the phenological situation will
require some operational action. A frequent error in validating a modelled predic-
tion by independent observations (such as those tabulated above) is to calculate
only a correlation coefficient and be satisfied if it is significant at some level of
accuracy. We must use any relation we have developed to our advantage, but we
must also, in the meantime, think of improving it.

9.5 REMARKS ON Special agrometeorological problems often require special field investigations. The
EXPERIMENTAL economics of experimentation and effort imply that the type of equipment,
PROCEDURES layout, data collecting retrieval and processing facilities should not be more elab-
orate than is needed for obtaining the answer to a particular question. Frequently,
much information can be obtained by simple 'probe' observations. However,
savings in 'hardware' can easily be overtaken by the cost of staff time required to
supplement information which was not measured. If for useful analysis of data
from simple 'probes' it is necessary to try 'reconstructing' weather data at some
site from available sources (such as synoptic charts), then experience shows that
the man-hours which are needed to achieve reconstruction in an acceptable way,
may often cost considerably more than the initial outlay which would have been
needed to measure the reconstructed data. This leads to an important principle:
do the field job with minimal equipment, but not with less. In a very real sense
there can be an inverse relationship between the advance cost of increasing on-
site instrumentation facilities and the subsequent cost of processing the data.
In extended agrometeorological field work, prior to actual trials, some time
should be devoted to testing equipment and layout. One aim of such preliminary
work is to gain experience in the operation, reading, data collecting and

169
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

processing procedures. Part of this experience will be on the suitability of

available experimental means, because often financial constraints and staff
shortages necessitate the use of sub-standard equipment. However, even
seemingly excellent equipment may fail miserably in field conditions. Humidity
and heat, insects and dirt, or simply the requirement to work continously have
defeated many instruments which performed well in the laboratory. It is also
important to work out a representative distribution of available equipment, and
to dispose of it in such a way that some guidance is obtained on the degree of
inevitable variability which exists in the investigated environment.
During the experiment, constant scrutiny and analysis of the observations
should be undertaken to ensure that the initially adopted observation procedures
are indeed the best that can be achieved. It is equally, if not more,. important to
document systematically the procedures followed, what the surroundings looked
like, and all the unforeseen occurrences during an investigation. While most of us
are aware that no experiment is finished until its analysis has been written up,
experience also shows that there is often little time for analysis at the most oppor-
tune moment, during and· immediately after the experiment. Much later, when
the opportunity finally arises to work on the material obtained, success is only
possible if there is enough descriptive material, such as a well-used logbook and
pictures of the observation sites.
Nowadays there are (budget permitting) good modern technical observation
facilities, like field-usable data loggers driven by solar and wind power, which
should be exploited fully. However, machines are stupidly predictable. There is no
real substitute for the dedicated field observer, because only he or she can deal
properly with the unexpected which always occurs during experiments.

Review questions on Chapter 9

A. Questions on basic understanding:

9.1 What are the reasons for screening thermometers? Describe the main
features of the design of a standard Stevenson temperature screen and
explain the function of these features.
9.2 At night (especially in clear weather) the reading of an exposed 'grass
minimum' thermometer is below that of a thermometer in the screen.
Explain why this is the case. Discuss the expected differences in reading
'grass minimum' thermometers which are exposed respectively over
grass, over bare ground and over stone.
9.3 Discuss the difficulties of measurement of precipitation by a simple rain-
gauge, particularly:
• Why is it necessary to adopt a standard height of the rim above
ground level?
• What are the advantages of surrounding a gauge with a turf wall?
• What distance should be kept to surrounding obstacles?
• What is the size of the area for which its measurements are repre-
sentative?
9 .4 Describe the Piche evaporimeter and discuss some of the problems
related to its use.
9.5 Discuss and compare the advantages and disadvantages of five methods
of soil moisture measurement: (i) Gravimetric; (ii) Tensiometers; (iii)
Conductivity blocks; (iv) Neutron scattering; and (v) TDR.
· 9 .6 Describe the eight main phases of wheat development and discuss
briefly the method of determining the stem extension phase.
9. 7 Comment on the following statements and advice and state with
reasons whether each is right or wrong:
(i} When calibrating a thermograph, decisions on corrections/adjust-
ments should not be based upon a single comparison.
(ii) Checks based upon comparisons of daily maximum values are
likely to be less accurate than those based on daily minimum
values ..
CHAPTER 9 - AGROMETEOROLOGICAL INSTRUMENfS AND OBSERVATION

(iii) Checks using the daily maximum or minimum temperatures are

less satisfactory than comparisons based upon spot readings
distributed throughout the diurnal cycle.
(iv) The best method of checking the zero setting of a Hg-in-steel
thermograph against a standard Hg-in-glass thermometer is to use
a water bath,with the water constantly agitated.
(v) Because of its robustness and firm trace, the twin-bulbed Hg-in-
steel instrument is frequently used in remote areas as a wet and dry
bulb sensor. A continuous run of many hours with zero depression
can mean 100 per cent humidity, but it may be a symptom of
malfunctioning. Explain.
(vi) More weight should be given to comparisons obtained in over-
cast/windy weather than in clear/calm weather.
(vii) The stability of psychrometer calibration (separation of the traces)
should be checked by running both bulbs dry. Explain.
(viii) When using the twin-bulbed Hg-in-steel instrument to measure
soil temperature at shallow depths (say 10 cm and 20 cm), it is
prudent to separate the pens by a distance equivalent to a temper-
ature difference of more than l0°C. Explain why this is desirable.
(ix) If you want to obtain representative wind speed data, you need
simultaneous wind direction data as welL

B. Questions requiring extended application:

9 .8 Compare the presented development schemes of wheat, maize and
avocado, and devise a scheme of basic phases which would fit all three
of them, and possibly also other crops.
9 .9 Describe and discuss the basic differences between meteorological and
biological observations. Extending the argument, describe the main
differences between biological and meteorological observations for
precise research purposes and the corresponding observations for opera-
tional use.
9 .10 Which station measurements could be replaced by remote-sensed satel-
lite data without prejudicing the availability of essential information?
9 .11 Write a short essay on the evaporation of a Class A pan and a sunken
tank, as compared to the evaporation from grassland and from a free
water surface.
9.12 List the processes (except precipitation) which result in water being
present on foliage. Also discuss the features which lead to the differen-
tial deposit of dew on different bodies. Use these facts to consider the
unrepresentative nature of many methods of dew measurements (espe-
cially those in which isolated bodies or surfaces are freely exposed) in
relation to the wetness of crop foliage.
9 .13 Assuming that the temperature in a Stevenson screen outdoors was 0°C,
and that we wanted to check a thermograph reading by comparing it
with a good mercury thermometer which had been stored indoors at a
temperature of 20°C, and had a lagtime)... = 100 sec at the 1 m/s flow
speed in the Stevenson screen - how many minutes would it take for
the thermometer to adjust to the screen-temperature within 0.1 °C?
9 .14 What are the operational advantages and disadvantages of using quick-
response against slow-response temperature measuring systems in
agrometeorological field work? Consider specific cases - e.g. frost
surveys; monitoring of conditions for disease warning systems; studies
of the effects of soil temperatures on plants growth. Consider also the
measurement of other meteorological elements such as wind and
humidity in relation to response time.
9.15 List the measurements discussed in Chapter 9 which are not likely to be
made at a common synoptic weather station. Range them in order of
importance for agrometeorology and state briefly why each is more or
less important.

171
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

C. Questions involving practical work or local data:

9 .16 The difference between the reading from a grass minimum thermometer
and a simultaneous temperature reading in a nearby screen is not
uniform as temperature falls from the late afternoon, through the night
until the following morning. Undertake trials in which the temperatures
given by a screen instrument and an exposed grass minimum ther-
mometer are taken at hourly or more frequent intervals from
mid-afternoon through to the following morning. Plot the results,
noting when temperatures reach their lowest point, and in particular the
relationship between depression of the temperature given by the grass
minimum instrument below that in the screen and the standard param-
eter (screen minimum - grass minimum).

172
APPENDIX I
NOTES ON THE STATISTICAL ANALYSIS OF METEOROLOGICAL
DATA

Excellent accounts of the use of statistical methods in meteorological analysis are

presented in CL-G and in the WMO publications (WMO-Nos. 327 and 335 - both
by Lowry). More material is given in WMO-No. 100 (Chapter V), and WMO-TN
Nos. 71 and 81. Other basic references are Conrad and Pollak (1950), Brooks and
Carruthers (1953), Munn (1970), Spiegel (1961), and Chapter 8 of Stull (1988).
Analytical procedures range from the orderly tabulation of data and their
display in tables or in diagrams (and by mapping, a complex study in its own
right) to advanced methods of mathematical statistics - probability and sampling
theory, time series examination by harmonic analysis or power spectra, analysis of
variation and of contingencies. However, in attempts to obtain precision using
sophisticated methods, the value and power of simple statistical procedures must
not be overlooked. Simple two-way contingency tables which, for example, set
maximum or minimum or mean temperatures against the associated wind speed,
are often already quite valuable for practical advice. Climograms, displaying the
joint average variation of two parameters in the course of a year, are also very
instructive (see question 8.1).
The underlying objective is obviously to achieve the highest degree of quan-
tification possible. We need, however, to recognize that the accuracy and precision
of the basic data varies widely - from information on the presence or absence of
some characters or property, to continuous chart or digital output with a very high
degree of resolution.
An almost inescapable requirement in all field work is graphing data as they
become available. One particularly useful case is the plotting of calibration checks
- zero drifts, alterations in range and sensitivity are most easily spotted from
graphic displays and remedial measures can be put in hand with the minimum of
delay. This reduces the risk of very time-consuming retrospective corrections of a
series of data due to e.g. instrumental malfunction or the use of bad data.
One of the very negative influences of the increasing availability of data
loggers is that they tempt the field experimentator to use the logged data imme-
diately for analysis, without first checking their quality graphically. Visual display
of the rough data in scatter diagrams or other graphs, if not done already in the
course of obtaining the data, should be a first step in any reliable analysis.
When numerous data or analysis results are presented in a graph as mean
values of one parameter per class of another parameter (e.g. average relative
humidity for temperature classes of 0°-l 1°-2°, 2°-3°, etc.), the graph should
0
,

always contain variability bars, preferably of the standard deviation of the data in
the class around its plotted mean value. A graph which shows an empirical rela-
tion without such bars cannot show the reliability of the relation and is mainly
useful as nomogram.
Two characteristics of meteorological (and in general geophysical) data
deserve further emphasis:
(i) Many encountered frequency distributions are certainly not normal (i.e.
Gaussian) - many lack even approximate symmetry about the central mode, and
indeed, some do not show a single 'hump' or central mode at all.
An interesting example are observations of hourly cloud amounts and their
complement, hours of sunshine, which typically have a U-shaped distribution, so
that 'average cloudiness' becomes a rather useless parameter to compute. For
hourly precipitation, the frequencies have a reverse J-shaped distribution (see
WMO-No. 327, p. 112), since the most frequent hourly precipitation amount is nil.
In these emphatically non-Gaussian cases the arithmetic mean has little
significance, although the mode(s) correspond to some real properties of the
physical situation. The mean deviation and standard deviation are numerical
quantities which may have little physical meaning; the range between maximum

173
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

and minimum has a meaning, while quartiles, deciles and interquartile range can
convey useful information about these distributions.
Wind gusts provide another example of interest; the majority of them are
locally generated and have a quasi-Gaussian distribution about the average wind-
speed, while a small number are generated by thunderstorms and are very large.
In such cases the safest representative of average parameter properties is the
median value.
Often the most useful method of summarizing information from non-
normal distributions is by the use of cumulatives, the percentages of the total
number of items which have values 'less than' successively increasing limits (alter-
natively 'more than' a successive decreasing sequence of limiting values).
However, the distribution of statistical parameters obtained from samples of the
original (non-normal) population, e.g. the mean value of sample means, do tend
towards normality (see WMO-No. 327).

(ii) In general, observations of .a meteorological event made in sequence

either according to intervals of time or intervals of space, are rarely independent
in the statistical sense. For example, it is obvious that the temperature at any
given moment will be partly conditioned by the temperature one hour or one day
earlier. This is an expression of the persistence which arises from the fact that the
atmospheric engine possesses inertia - it has built-in lags, which ensure that if
some external input is suddenly applied there will be some delay before a
complete response is evident. Weather forecasting, incidentally, relies heavily on
this property of persistence.
The degree of coupling between successive measurements of the same
element is expressed by the coefficient of serial- or auto-correlation. WMO-No.335
discusses the autocorrelation between the daily maximum temperature on
particular days with that three days later, i.e. with an interval of three days. The
temperature on Day 1 is associated with that .on Day 4, Day 2 with Day 5 and so
on. If there is no systematic connection between maximum temperatures three
days apart, the autocorrelation coefficient tends to zero. If relatively high (or low)
maximum temperatures are associated with high or low temperature three days
apart, the coefficient approaches + 1. The greater the interval the less the correla-
tion - this is often, but by no means always, the situation in meteorology.
The effect of persistence upon the precision of a mean value based upon a
series of observations is discussed by Brooks and Carruthers (1953). A detailed
examination of this property of meteorological events is beyond the scope of this
text; however, the general implications of the topic can be appreciated from the
two examples below.
Example 1: Brooks and Carruthers (1953) in their discussion of upper-level
winds state that: "in applying tests of significance or assessing errors, n wind
observations cannot be regarded as n independent observations. In general winds
recorded at one or more times per day in runs of consecutive days may be assumed
roughly equivalent to D/2 independent observations, where D is the number of
days".
It should be stressed that this example deals with upper-level, i.e. synoptic-
scale, winds (see section 5.1). Any estimate of independent period should be
adjusted to scale. For boundary-layer winds, or in general boundary-layer param-
eters, this means that the length of an independent period is about 3 hours, i.e. in
D days we have approximately 8 D independent boundary-layer observations.
Example 2: Brooks and Carruthers (1953) quote specific examples of meteor-
ological events, e.g. days of snow, days of fog, and define a persistence factors, and
conclude that: " .... to a first approximation, a phenomenon having a persistence
factor s has a tendency to occur in groups of average size s. This causes length of
runs (sequences of observations of a particular event) to be s times as long, and the
variance (i.e. the square of the standard deviation) to be s times as great, as would
be expected if there were no persistence. It should, however, be stressed that when
computing a mean value from n observations, all observations are included and
the mean is "'£x/n".

174
APPENDIX I - NOTES ON THE STATISTICAL ANALYSIS OF METEOROLOGICAL DATA

Persistence becomes particularly important when considering the character of

a sequence of observations and measures of dispersion. Such effects are illustrated
by an example on temperature which can be found in Carruthers (1949). Suppose
we wish to estimate a mean temperature in a month of 31 days formed by taking
a single observation at a fixed hour on each day throughout the month. From the
sequence, 10 observations at 3-day intervals will produce a better estimate of the
31-day mean than 14 or 15 observations on consecutive days. The 'persistence
length' for temperature has been found to be about 3 days (see example in Lowry,
WMO-No. 335, p. 103). For scattered days, means of samples of 10 or a few obser-
vations will be distributed about the 31-day mean with a standard deviation of
cr/✓n (where cr = standard deviation of the single observations), but if the observa-
tions are consecutive the s.d. of the n-day means will be cr/✓(n/3), since
n consecutive observations are equivalent to only n/3 independent observations.
As mentioned earlier, the autocorrelation coefficient does not necessarily
decrease with increasing interval. To illustrate this point, consider an elementary
and rather artificial example. Supposing we have a long series of monthly mean
temperatures, then because of the strong seasonal variation the temperature of a
particular named month (e.g. January) will tend to be more closely associated with
that of the next January (a twelve-month interval) than with next June, a six-
month interval. Generally, if some meteorological element exhibits a periodicity,
for example if it tends to reach higher values every six months, or every twelve
months, then the correlation between values six (or twelve) months apart will
tend to be higher than that at shorter intervals. In other words, with increasing
interval the autocorrelation coefficient will first decrease and then increase. One
method of identifying regularities and periodicities in time series is that of using
autocorrelation coefficients.

A frequent problem in meteorological data analysis is the occurrence of so-called

outliers. If one or two values in a dataset are suspiciously different from the other
data points (e.g. the irradiation values measured for a given Sun elevation inter-
val in bright weather), two questions may arise. First, are these 'outlying' data due
to measurement error? or to some physical reason - in our irradiation example,
maybe the temporary presence of some unnoticed cloud? Second, should these
values be disregarded, or given a special status in further analysis? Before deciding
on this, however, one should first make sure that the difference is sufficiently
large that it really requires attention. Meteorological data often show a lot of
scatter, and it is bad science to eliminate outlying values arbitrarily from a dataset,
unless one can prove that their presence in the dataset is illogical. Chauvenet's
criterion provides an objective test to determine whether an outlying point is
statistically acceptable within a dataset.
This criterion states that the presence in a set of n data of a certain value with
an occurrence probability smaller than l/(2n) is incompatible with the assump-
tion that the data are normally distributed. Limiting standardized values given by
Parratt (1961) for this criterion at various values of n can (for n ~ 7) be accurately
represented by:
Ix-xi
- - = 1.42 + 0.301 In (n - 4)
crx
Typical values of this limiting ratio vary from around 2, for n around 10, to
more than 3 for n>200.
For application, we calculate the standard deviation crx and the average of the
entire dataset, including the presumed outliers. Data which differ from the average
more than the product of crx with the appropriate limiting ratio are then marked as
'not normal' in the objective statistical sense of that word. An investigation can then
take place to determine if there is a particular reason for their excessive deviation. If
they are eliminated from the dataset, the mean and crx must be recalculated, and
maybe the criterion should be applied a second time. In practice, the use of this
objective outlier criterion often indicates that certain data points, which subjectively
appear very excessive, are objectively still within a statistically acceptable range.

175
APPENDIX II
UNITS, CONVERSION FACTORS, USEFUL CONSTANTS

Conversion to S.I. (Standard International) units:

Dimensions (S.I.-unit: m)
1 inch = 0.0254 m 1 pouce = 0.0271 m
1 foot= 0.3048 m 1 pied = 0.3248 m
1 yard= 0.9144 m 1 fathom = 1.829 m
1 mile= 1.6093 km 1 nautical mile = 1.8532 km
1 Angstrom= 10-10 m 1 American pint = 0.4731 x I0-3 m3
1 acre= 4047 m2 1 British pint= 0.5683 x I0-3 m3
1 hectare (ha) = 104 m2 1 British gallon= 4.546 x I0-3 m3
1 American gallon= 3.785 x I0-3 m3
Velocity (S.I.-unit: ms-I)
1 ft s-I = 0.3048 m s-I 1 km hrI = 0.2778 m s-I
1 mile hrI (mph) = 0.4470 ms-I 1 knot =1 nautical mile hrI = 0.5148 ms-I
Mass (S.I.-unit: kg) and density (S.1.-unit: kg m-3)
1 ounce (oz)= 0.02835 kg 1 stone= 6.350 kg
1 pound (lb)= 0.4536 kg I lb ft-3 = 16.02 kg m-3
Pressure (S.I.-unit: 1 pascal (Pa) =1 newton m-2 =1 kg m-I s-2)
1 millibar (mb) = 100 Pa =1 hPa 1 mm mercury= 133.32 Pa
1 lb ft- 2 = 4.882 Pa 1 m water = 9 .81 x 103 Pa
1 lb inch-2 = 6.895 kPa 1 atm(osphere) = 1013.2 hPa
Energy (S.I.-unit: 1 joule CT) =1 newton m =1 kg m2 s-2)
1 (gram-) calorie (cal)= 4.187 J 1 erg (=l dyne-cm)= I0-7 J
1 British Thermal Unit (Btu) = 1055 J 1 kilowatt-hour (kWh) = 3.60 MJ
Power (S.I.-unit: 1 watt (W) =1 J s-I)
1 cal s-I = 4.187 W 1 Btu hrI = 0.293 W
Energy flux density (S.1.-unit: W m-2)
1 langley (Ly) =1 cal cm-2 min-I = 697.8 W m-2
1 cal cm-2 day-I= 0.4846 W m-2 1 mW cm-2 = 10 W m-2
Thermal conductivity: 1 cal cm-I K-I s-I = 4.187 x 102 J m-I K-I s-I
1 Btu ft-I °F-I = 6.23 x 103 J m-I K-I
Specific heat: 1 cal K-I g-I = 4.187 x 103 J kg-I K-I
Thermal capacity: 1 cal cm-3 K-I = 4.187 x 106 J m-3 K-I
Thermal diffusivity: 1 cm 2 s-I = IQ-4 m2 s-I
Useful constants and quantities:
Solar constant 1 380 W m-2
Earth's radius (of sphere of equal volume) 6 371.2 km
Acceleration of gravity at 45° and g = 9.807 m s-2
Mean Sea Level (MSL)
Stefan-Boltzmann constant 5.67 x I0-8 W m-2 K-4
Avogadro number (molecules/mole in ideal gas) 6.022 x 1Q23 moI-I
Specific gas constant for dry air R = 287 J kg-I K-I
Specific heat capacity at constant pressure for dry air CP = 1 005 J kg-I K-I
Standard temperature and pressure (STP) p = 1 013.2 hPa, T = 0°C
Specific latent heat of vaporization of water at STP 2.50 MJ kg-I
Specific latent heat of fusion of ice at STP 0.334 MJ kg-I
Density of air at STP 1.292 kg m-3
Density of ice at 0° C 917 kg m-3

177
WMO LECTIJRE NOTES - AGRICULTURAL METEOROLOGY

1 radian = 5 7.3 degree = 5 7° 18'

n degrees Fahrenheit (°F) = (n - 32). 5/9 degrees Celsius ( C)
0

n degrees Celsius (centigrade)= n + 273.16 Kelvin (K)

1 part per million (1 ppm, see section 6.7) = 12.2 (M/T) mg m-3 if M = molecu-
lar weight
Prefixes: h = hecto- = x 102
k = kilo- =x 1Q3
M = mega- =x 1Q6
c = Centi- =x 10-2
m = milli- =x 10-3
µ=micro- =x lD-6

178
APPEND IX III
ESTIMATION OF SOLAR ELEVATION AND DAY LENGTH
Approximation with procedure by H. R. A. Wessels (see Holtslag and Van Ulden, 1993)

The solar altitude above the horizon, h, at a given time and location is needed for
all basic calculations of incoming radiation. For estimation of h, Sun path
diagrams have been constructed (see e.g. Lowry, WMO-No. 335, section 8.11), but
that must be done anew for each latitude. The following approach provides a
rough estimate of h, accurate to± 2°, needing only date, time and location coor-
dinates. For much radiation work this is quite sufficient. A useful additional result
is the number of hours of daylight, S.
Take a location on earth at northern latitude <I> and at western longitude B.
At Firenze, 44 ° North and 11 ° East, we have B = -11. For the entire calculation we
must convert all angles to radians (degrees/57.3), reconverting final results to
degrees.
First estimate the day-number d. If M is the number of the actual
month (1 to 12) and D is the number of the day in the month (1 to 31), then
d == 30 (M - 1) + D. Then we can calculate the solar longitude SL (in radians) from:
SL= 4.871 + 0.0175 d + 0.033 sin (0.0175 d)
The solar declination 8 is given by: 8 = arcsin (0.398 sin(SL)).
The day length S (from sunrise to sunset) follows then from
7.5 S = arccos (- tan <I> tan 8), in hours of duration when the arccos-calculation
result is expressed in degrees.
If t is the actual universal time (UT= time at B = 0°, 'Greenwich time') in
hours, we can calculate the local hour angle a, through which the earth must turn
to bring the meridian of the given location directly under the Sun, as follows (in
radians):
a = - B + 0.043 sin(2 SL) - 0.033 sin(0.0175 d) + 0.262 t - 1t

Then the solar elevation h follows from sin h = sin o sin <I> + cos 8 cos <I> cos a.

As an example we calculate for 15 June: d = 165, SL= 7.767 radians, o = + 23.36°.

Then for Firenze at 1200 local time= 1100 UT: a= -3.9°➔ h == 69°, S = 15.3 hr.
On the same day of the year at Buenos Aires (35°S, 58°W) at 0930 local time
(= 1350 UT) we obtain: a= -35.4°➔ h == 23°, S = 9.7 hr.
More elaborate and precise formulas for calculating these parameters and
other solar information, e.g. the Sun's azimuth, are given in CL-G (Annex AZ) and
by Oke (1987).

179
APPENDIX IV
RELEVANT WMO PUBLICATIONS, AND AGROMETEOROLOGICAL
CORE LIBRARY

Three compendia are referred to on a regular basis throughout this publication

because they are considered to be essential references (see Preface).
• Compendium of Lecture notes for training Class IV meteorological
personnel, Volume II - Meteorology, B. J. Retallack (WMO-No. 266,
2nd ed., 1984) GM-II
• Compendium of Lecture notes in Climatology for Class III and Class IV
meteorological personnel, B. J. Garnier (WMO-No. 726, 1992) CL-G
• Compendium of Lecture notes for training Class IV agrometeorological
personnel, A. V. Todorov (WMO-No. 593, 1980) AG-T

Two older WMO compendia, now out of print, deserve particular mention
because of their quality:
• Compendium of Lecture Notes in Climatology for Class IV Meteorological
Personnel (W. P. Lowry, 1972), WMO-No. 327.
• Compendium of Lecture Notes in Climatology for Class III Meteorological
Personnel (W. P. Lowry, 1972), WMO-No. 335.
Two current WMO compendia out of the series WMO-No. 364 for Class I and
Class II meteorological personnel, edited by A.C. Wiin-Nielsen, contain actual
material of large relevance to certain sections of this agrometeorological
compendium:
• Compendium Vol. I, Part 2: Physical meteorology, B. J. Retallack (1973);
• Compendium Vol. II, Part 6: Air chemistry _and air pollution meteorology,
R. E. Munn and H. Rodhe (1985).
Three current and regularly updated WMO publications contain fundamen-
tal application information:
• Guide to Climatological Practices, WMO-No. 100 (2nd ed. 1983).
• Guide to Agricultural Meteorological Practices, WMO-No. 134 (2nd ed. 1981)
• Guide to Meteorological Instruments and Methods of Observation, WMO-No. 8
(5th ed. 1983).
It will have become obvious that, out of the various publications of the
WMO, the Technical Notes series is particularly useful. The following list gives
titles of other Notes relevant to agrometeorology. The double asterisk symbol (**)
at the end of some of the Technical Notes refers to presence of a short abstract in
the appendix to Smith, L. P., 1975: Methods in Agricultural Meteorology.
For convenient access, WMO has issued all technical notes on agricultural
meteorology on the following set of CD-ROMs.
• Publications of the Commission for Agricultural Meteorology, 1954-1999,
Volume I - Technical Notes, Volume II - Technical Notes, CAgM Reports and
Brochures, World Meteorological Organization, 2001.

WMO TECHNICAL NOTES:

The following publications are in print:
No. 101 Meteorology and grain storage. (WMO-No. 243, 1969 and 1990) (**)
No. 157 Techniques of frost prediction and methods of frost and cold protection.
(WMO-No. 487, 1978) (See TN No. 51).
No. 173 Weather and airborne organisms. (WMO-No. 562, 1980) (See TN No. 54).
No. 180 Weather-based mathematical models for estimating development and ripen-
ing of crops. (WMO-No. 620, 1980) (See TN Nos.151, 17 4).
No. 182 Analysis of data from international experiments on lucerne.
(WMO-No. 629, 1986).
No. 184 Land use and agrosystem management under severe climatic conditions.
(WMO-No. 633, 1986).

181
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

No. 186 Land management in arid and semi-arid areas. (WMO-No. 662, 1989).
No. 188 Applications of meteorology to atmospheric pollution problems.
(WMO-No. 672, 1987) (See TN Nos.114, 121, 139, 177).
No. 190 Weather, climate and .a nimal performance. (WMO-No. 684, 1988) (See
TN No. 122).
No.192 Agrometeorological aspects of operational crop protection. (WMO-No. 687,
1988).
No. 193 Agroclimatology of the sugar-cane crop. (WMO-No. 703, 1988).
No. 196 Climate variabili~ agriculture and forestry. (WMO-No. 802, 1994).

The out-of-print technical notes with an asterisk (*) marked beside them are now
available in electronic format on the CD-ROM of the Publications of the
Commission for Agricultural Meteorology, 1954-1999 (see reference above).
The following publications are no longer available in printed form:
No. 4 (*) Energy from the wind - assessment of suitable winds and sites.
(WMO-No. 32, 1954) (See TN Nos. 63, 175).
No. 10 (*) The forecasting from weather data of potato blight and other plant
diseases and pests. (WMO-No. 42, 1955) (**) (See TN No. 41).
No. 20 (*) The climatological investigation ofsoil temperature. (WMO-No. 72, 1958) (**).
No. 32 (*) Meteorological service for aircraft employed in agriculture and forestry.
(WMO-No. 96, 1960) (**).
No. 41 (*) Climatic aspects of the possible establishment of the Japanese beetle in
Europe. (WMO-No. 110, 1961) (**) (See TN No. 10).
No. 42 (*) Forecasting for forest fire service. (WMO-No. 110, 1961) (**).
No. 43 (*) Meteorological factors influencing the transport and removal of radio-
active debris. (WMO-No. 111, 1961) (**).
No. 51 (*) Protection against frost damage. (WMO-No. 133, 1963) (**) (See TN No. 15 7).
No. 53 (*) The effect of weather and climate upon the keeping quality of fruit.
(WMO-No. 137, 1963) (**).
No. 54 (*) Meteorology and the migration of the desert locust. (WMO-No. 138,
1963). (**) See conference proceedings TN No. 69 (1965, WMO-
No.171), and TN No. 173.
No. 55 (*) The influence of weather conditions on the occurrence of apple scab.
(WMO-No. 140, 1963) (**) (+ then-existing instrumentation for
measuring wetness of leaf).
No. 56 (*) C. P. de Brichambaut, C. C. Wallen: A study of agroclimatology in semi-
arid and arid zones of the Near East. (WMO-No. 141, 1963) (See
TN No.179).
No. 59 (*) J. van Eimem, et al.: Windbreaks and shelterbelts, 188 pp. (WMO-No. 147,
1964) (**).
No. 63 (*) Sites for wind power installations. (WMO-No. 156, 1964) (See TN
No. 175).
No. 65 (*) A survey of human biometeorology. (WMO-No. 160, 1964) (**).
No. 71 (*) Statistical analysis and prognosis in meteorology. (WMO-No. 178, 1966).
No. 76 (*) Instruments and measurements in hydrometeorology. (WMO-No. 191, 1966).
No. 83 (*) Measurement and estimation of evaporation and evapotranspiration.
(WMO-No. 201, 1966). (**) (Revision of TN No. 11, WMO-No. 42, 1955).
No. 85 (*) Precision des mesures pyrheliometriques (Accuracy of measurements by
pyrheliometers). (WMO-No. 209, 1967) (**).
No. 86 (*) J. Cocheme and P. Franquin: An agroclimatology survey ofa semi-arid area
in Africa south ofthe S(lhara. (WMO-No. 210, 1967) (See TN No. 125).
No. 96 (*) Air pollutants, meteorology and plant injury. (WMO-No. 234, 1969) (**)
(See TN No. 147).
No. 97 (*) Practical soil moisture problems in agriculture. (WMO-No. 235, 1969)
(**).
No. 99 (*) Meteorological factors affecting the epidemiology of wheat rust.
(WMO-No. 238, 1969) (**).
No. 105 (*) Artificial modification of clouds and precipitation. (WMO-No. 249,
1969). (See TN Nos. 1, 13, 154).

182
APPENDIX IV - RELEVANT WMO PUBLICATIONS, AND AGROMETEOROLOGICAl CORE LIBRARY

No. 114 Meteorological practice in air pollution. (WMO-No. 274, 1970) (See TN
Nos. 121, 139, 177, 188).
No. 118 Protection of plants against adverse weather. (WMO-No. 281, 1971) (**).
No. 119 The application of micrometeorology to agricultural problems.
(WMO-No. 298, 1972) (**).
No. 121 Dispersion and forecasting of air pollution. (WMO-No. 319, 1972). (See TN
Nos. 114, 139, 177, 188).
No. 122 Some environmental problems of livestock housing. (WMO-No. 325, 1972)
(**) (See TN Nos. 113, 159, 190, 191).
No. 123 The assessment of human bioclimate - a limited review of physical param-
eters. (WMO-No. 331, 1972) (See TN No. 65).
No. 125 L. H. Brown and J. Cocheme: A study of the agroclimatology of the high-
lands of eastern Africa. (WMO-No. 339, 1973) (See TN No. 86).
No. 131 J. Seemann: Climate under glass. (WMO-No. 373, 1973).
No. 132 Applications of meteorology to economic and social development.
(WMO-No. 375, 1974) (See TN Nos. 164, 168).
No. 133 L. P. MacHattie, F. Schnelle: An introduction to agrotopoclimatology.
(WMO-No. 378, 1974).
No. 136 Mulching effects on plant climate and yield. (WMO-No. 388, 1975).
No. 137 Meteorology and the Colorado potato beetle. (WMO-No. 391, 1975).
No. 138 Drought and agriculture. (WMO-No. 392, 1975) (See TN No. 186).
No. 139 Climatological aspects of the composition and pollution of the atmosphere.
(WMO-No. 393, 1975). (See TN Nos. 114, 121, 177, 188).
No. 144 Rice and weather. (WMO-No. 423, 1975).
No. 147 Review of present knowledge of plant injury by air pollution.
(WMO-No. 431, 1976) (See TN No. 96).
No. 151 Crop-weather models and their use in yield assessments. (WMO-No. 458,
1977) (See TN Nos. 174, 180).
No. 152 Radiation regime of inclined surfaces. (WMO-No. 467, 197"7).
No. 154 The scientific planning and organization ofprecipitation enhancement exper-
iments, with particular attention to agricultural needs. (WMO-No. 478,
1977) (See TN No. 105).
No. 159 Weather and parasitic animal diseases. (WMO-No. 497, 1978).
No. 160 Soya bean and weather. (WMO-No. 498, 1978).
No. 161 Estudio agroclimatol6gico de la zona andina. (WMO-No. 506, 1978).
No. 164 The economic value of agrometeorological information and advice.
(WMO-No. 526, 1980) (See TN Nos. 132, 168).
No. 167 Meteorological factors affecting the epidemiology of the cotton leaf worm and
the pink bollworm. (WMO-No. 532, 1980).
No. 168 The role of agrometeorology in agricultural development and investment proj-
ects. (WMO-No. 536, 1980) (See TN Nos. 132, 164).
No. 172 Meteorological aspects of the utilization of solar radiation as an energy
source. (WMO-No. 557, 1981).
No. 17 4 The effect of meteorological factors on crop yields and methods of forecasting
the yield. (WMO-No. 566, 1982) (See TN Nos. 151, 180).
No. 175 Meteorological aspects of the utilization of wind as an energy source.
(WMO-No. 575, 1981) (See TN Nos. 4, 63).
No. 177 Review of atmospheric diffusion models for regulatory applications.
(WMO-No. 581, 1982). (See TN Nos.121, 139, 188).
No. 178 Meteorological aspects of certain processes affecting soil degradation -
especially erosion. (WMO-No. 591, 1982) (See TN Nos. 20, 97).
No. 179 Agroclimatology of the humid tropics of South-East Asia. (WMO-No. 597,
1982) (See TN No. 56).
No. 191 Animal health and production at extremes of weather. (WMO-No. 685,
1989) (SeeTNNos.113, 122,159).

183
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

Orders for missing WMO Publications (give WMO-number!) should be

addressed to:
Sales and Distribution of Publications
World Meteorological Organization
7 bis, Avenue de la Paix
P.O. Box 2300,
CH-1211 Geneva 2
Switzerland
Tel: (+44-22) 730 81 11
Fax: (+41-22) 730 80 22
E-mail: [email protected]

AGROMETEOROLOGICAL CORE LIBRARY

This compendium is a pedagogical introduction, not a handbook or an encyclo-

pedia on agrometeorological issues. Whoever wishes to have more background
information will need access to other books. The list below is a proposal for a 13-
book library for people, who either want to specialize in agrometeorology, or want
to support their pres.e nt knowledge of agrometeorology with study of closely
related other subjects. Three points need to be noted about this list: (a) All the
titles are in English; (b) They have not been published by WMO (those have
already been listed above); and (c) They have been published (or revised) after
1980 and were still available in 1997.

Basic reviews:
Baldy, C., Stigter, C. J., 1997: Agrometeorology of multiple cropping in warm climates.
Oxford and IBM Publ. Comp., New Delhi, India, 237 pp.; ISBN 81-204-1191-
9. Also: INRA (Paris) English ed.; ISBN 2-7380-0772-4.
Geiger, R., Aron, R. H., Todhunter, P., 1995: The climate near the ground (5th
edition). Vieweg, Germany, 528 pp.; ISBN 3-528-08948-2.
Griffiths, J. F. (ed.), 1994: Handbook of agricultural meteorology. Oxford University
Press, United Kingdom, 320 pp.; ISBN 0-19-506240-X. (expensive)
Jackson, I. J., 1989: Climate, water and agriculture in the tropics (2nd edition).
Longman, United Kingdom, 377 pp.; ISBN 0-582-02159-6.
Lenschow, D. H. (ed), 1986: Probing the atmospheric boundary layer. American
Meteororological Society, Boston, 269 pp.; ISBN 0-933876-63-7.
Linacre, E., 1992: Climate data and resources. Routledge, New York, 366 pp.; ISBN 0-
415-05 703-5.
Lowry, W. P., Lowry, P. P., 1989: Fundamentals of biometeorology: the physical envi-
ronment. Peavine (Oregon, United States), 2 vols., 650 pp.,
ISBN 0-882002-05-9.
Oke, T. R., 198 7: Boundary layer climates (2nd edition). Methuen/Routledge, United
Kingdom, 435 pp.; ISBN 0-415-04319-0.
Rosenberg, N. J., Blad, B. L., Verma, S. B., 1983: Microclimate: the biological environ-
ment (2nd edition). Wiley, United States, 495 pp.; ISBN 0-471-06066-6.

Advanced handbooks:
Brutsaert, W. (1982): Evaporation into the atmosphere. Reidel/Kluwer, Dordrecht,
299 pp.; ISBN 90-277-12476. (expensive)
Jones, H. G. (1992): Plants and microclimate. Cambridge University Press, 428 pp.;
ISBN 0-521-41502-0/0-521-42524-7.
Kaimal, J.C., Finnigan, J. J. (1994): Atmospheric boundary layer flows, their structure
and measurement. Oxford University Press, 289 pp.; ISBN 0-19-506239-6.
(expensive)
Monteith, J. L., Unsworth, M. (1990): Principles of environmental physics (2nd
edition). Arnold, United Kingdom, 291 pp.; ISBN O 7131 2931 X.

184
REFERENCES

Abels, H., 1892: Beobachtungen der taglichen Periode der Temperatur im Schnee
und Bestimmung des Warmeleitungsvermogens des Schnees als Funktion
seiner Dichtigkeit. Rep. f Meteor., 16, Nr. 1.
Akesson, N. B., 1965: Pesticide drift residue. Aerial Applicator, 2 (3). 5.
Allen, R. G., Smith, M., Pereira, LS., Perrier, A., 1994: An update for the calcula-
tion of reference evapotranspiration. ICID Bulletin, 43, pp. 35-92.
Amble, H. R., 1959: Atmospheric salinity at various places in United Kingdom.
C.l. Report No. 104, Min. of Supply, London, 23 pp.
Alharthi, A., Lange, J., 1987: Soil water saturation: dielectric determination. Water
Resources Research, 23, pp. 591-595.
Anda, A., Stephens, W. (1996): Sugar beet production as influenced by row orien-
tation. Agronomy Journal, 88, pp. 991-996.
Anonymous, 1950: Railw. Engin. Maint., Chicago, 46, p. 1132.
Arnfield, A. J., 1979: Evaluation of empirical expressions for the estimation of
hourly and daily totals of atmospheric long wave emissions under all sky condi-
tions. Quarterly Journal of the Royal Meteorological Sodety, 105, pp. 1041-1052.
Atzema, A. J., 1993: Moisture content of cereals at harvesting time by comparing
microclimate values and standard weather data. Netherlands Journal of
Agricultural Science, 41, pp. 167-178.
Atzema, A. J., 1995: Determination of the quality of forecast moisture content of
cut grass using forecast weather elements. Journal of Agricultural Engineering
Research, 61, pp. 1-10.
Avissar, R., Mahrer, Y., 1988: Mapping frost-sensitive areas with a three-
dimensional local-scale numerical model. Journal of Applied Meteorology, 27,
pp. 400-426.
Baier, W., Robertson, G. W., 1966: A new versatile soil moisture budget. Canadian
Journal of Plant Science, 46, pp. 299-315.
Baier, W., Robertson, G. W., 1968: The performance of soil moisture estimates as
compared with the direct use of climatological data for estimating crop yields.
Agricultural Meteorology, S, pp. 17-31.
Baier, W., 1969: Concepts of soil moisture availability and their effects on soil
moisture estimates from a meteorological budget. Agricultural Meteorology, 6,
pp. 165-178.
Baier, W., 1973: Crop-weather analysis model: review and model development.
Journal of Applied Meteororology, 12, 937-947.
Baker, C. R. B., 1980: Some problems in using meteorological data to forecast the
timing of insect life cycles. Bulletin EPPO, 10, 83-91.
Baldy, C., Stigter C. J., 1997: Agrometeorology of multiple cropping in warm climates.
Oxford & IBM Puhl.Comp., 237 pp. (Original: en Francais, 1993, INRA, Paris)
Battan, L. J., 1969: Weather modification in the U.S.S.R., Bulletin of the American
Meteorological Society, SO, pp. 924-945.
Baver, L.D., 1948: Soil Physics (2nd Ed.). John Wiley, New York, 398 pp.
Beadle, M., 1966: Inexpensive modification of the Jordan sunshine recorder.
Weather, 21, pp. 98-101.
Bean, A., Alperi, R. W., Federer, C. A., 1975: A method for categorizing shelterbelt
porosity. Agricultural Meteorology, 14, pp. 417-429.
Bedson, G., 1992: Leaf wetness information paper. Austr.Bur.Meteor., unpubl.
Beer, T., 1991: The interaction of wind and fire. Boundary-Layer Meteorology, 54, pp.
287-308
Beljaars, A. C. M., 1982: The derivation of fluxes from profiles in perturbed areas.
Boundary-Layer Meteorology, 24, pp. 35-55.
Beljaars, A. C. M., 1987a: On the memory of wind standard deviation for upstream
roughness. Boundary-Layer Meteorology, 38, pp. 95-101.

185
'WMO ~~cmRE NOil:S- 7\'GRI'Ol!TLT:UML ME'fE0R'OI;0GY

Beljaars, A. C. M., 1987b: The measurement of gusts at routine wind stations - a

review. WMO/CIMO Instruments and Methods of Observation Report 31.
Beljaars, A. C. M., Holtslag, A. A. M., 1991: Flux_parameterization over land surfaces for
atmospheric models. Journal ofApplied Meteororology, 30, pp. 327-341.
Benichou, P., Le Breton, 0., 1987: Prise en compte de la topographie ·p our la carto-
graphie des champs pluviometriques statistiques. La Meteorologie, 7, No. 19.
-Benincasa, F., Maracchi, G,, Rossi, ,P., 1991: Agrometeorologia. -P atron (Bolqgna),
426 ·pp.
:Black, CA. {ed.), 19'65: Meth0ds -0Lsoil -ana1ysis ..Am. Soc. Agron., Monogr.aph .9.
Blanc, M.1., 1960: Soil temperature: introduction .and -importaace .in ag-rometeo-
rology. .Final Report - lnter-r.egional seminar on tropical agrometeorology, Maracay,
'"Venezuela, 45-Si.
Bohren, -:C. -E., 1987: Clouds in ,a glass of-beer. Wiley, New York, 195 pp.
Bordreuil, C, 1979: 'Bulletin il nformation. Meteo-France, No.. 49.
Bowen, E.G., 1967.:-cloud seeding. Sdence Joumdl (United Kingdom) 3(8), pp. 69-73.
Bowen, A. J., Lindley, D., 1977: A windtunnel investigation of the wind speed and
turbulence chaucteristics close to the .g round over various escarpment ·shapes .
.Boundary-Layer Meteorology, 12, :PP-· 259-2'71.
Bowers, S. A., Hanks, R. D., 1965: Reflection of radiant ·e ner,g y from soils. Soil
Science, 100, ,p. BO.
·B raham R. (ed.), 198·6: Pred;pitation enhancement- a scientific challenge. American
Meteorological Society, Meteor. Monograph ;.211.,, No. 43, 171 pp.
Brooks, C. E. P., Carruthers, N., 1953: Handbook of statistical methods in meteorology.
H.M. Stationery Office, London, 412 pp.
Brocks, K., 1939: Nachtliche Temperaturminima in Flirchen mit verschiedenem
Boschung-swinkel. Meteor. Zeitschr., 56, pp. 378-383.
Brunt, D., 1932: Notes on radiation in the atmosphere. Quarterly Journal of the
Royal Meteorological Society, 58, pp. 389-420.
Brutsaert, W., 1975: Comments on surface roughness parameters and the height
of dense vegetation. Journal of the Meteorological Society oflapan, 53, pp. 96-97.
Brutsaert, W., 1982: Evaporation into the atmosphere. Reidel, Dordrecht, 299 pp.
Businger, J. A., 1963: The glasshouse (greenhouse) climate. In: Van Wijk, W. R.
(ed.), Physics of plant environment, North Holland Publ. Co., pp. 277-318.
Callander, B. A., 1990: Agrometeorological information for use in agricultural
production, including weather forecasting. In: Price-·B udgen A. .(ed.): "Using
meteorological information and products", Ellis Horwood, pp. 437-454.
Carruthers, N., 1949: The accuracy of a mean of n temperature observations as an
estimate of the mean -temperature for a particular month. Meteorological
Magazine, 78, ,pp. 65.
·Changnon, S. A., 1971: Hail fall characteristics related to crop damage. Journal of
Applied Meteorology, ,10, pp. 270-27 4.
:Changnon, S. A., 1975.: The ,paradox of p1armed weather modification. Bulletin of
the American Meteorelqgical Society., 56, ip.p. 27-37.
Changnon, S. A., Ivens, J. L., 1981: History repeated: the forgotten hail cannons
of Europe. Bulletin o('the American Meteorological Society, 62, pp. 368....,375.
Changnon, S. A., 1992: Contents of climate _predictions des1red by agricultural
decision makers. Journal of Applied Meteorology, 31, p.p. 1488-1491.
Christoffer, J., Jurksch, G., 1985.: Untersuchung von Anemometerstandorten
mittels sektorabhangiger Boenfaktoren. Meteor. Rundschau, 38, pp. 43-48.
Coile, T. S., 1936: Soil samples. Sot/ science, -42, pp. 139-142.
Collier, C. G., 1986: Accuracy of n1infall estimates by radar. Journal of Hydrology,
-83, :P.P· .207-23.5.
Collier, RH., Finch, S., Phelps, K.; 1991: Forecastin,g the liming of attacks of Delia
radicum on cruciferous crops. Bulletin EPPO, 21, pp. 419-424.
Conrad, V. A. , Pollak, L. W., 1950: Methods in climatology. Harvard Univ. Press,
Mass., 2nd ed., 459 pp.
·Cotton, W. R., Pielke, R. A., 1995: Human impacts on weather .and climate.
Cambridge Univ. Fress, 288 pp.
Coulson, K. L., 1-975: Solar and terrestrial radiation. Acad.Press, London, 322 pp.

186
REFERENCES

Counihan, J., Hunt, J. C. R., Jackson, P. S., 1974: Wakes behind 2-dim. surface
obstacles in turbulent boundary layers. /.Fluid Mech., 64, pp. 529-563.
Crawford, T. V., 1964: Computing the heat requirements for frost protection.
Journal of Applied Meteorology, 3, pp. 750-760.
Dalton, J., 1802: Experimental essays on the constitution of mixed gases,
on the force of steam or vapor from water and other liquids in different
temperatures, both in a Torricellian vacuum and in air, on evaporation
and on the expansion of gases by heat. Mem. Manchester Lit. Phil. Soc., S,
pp. 535-602.
Dalton, G. E., 1974: The effect of weather on the choice and operation of harvest-
ing machinery in the United Kingdom. Weather, 29, pp. 252-260.
Davenport, A. G., 1960: Rationale for determining design.wind velocities. Journal
of the Structural Division, American Society of Civil Engineers, 86, pp. 39-68.
Davenport, T. L., 1982: Avocado growth and development. Proc. State Hortic. Soc.,
95, pp. 92-96.
Davis, J. R., 1955: Frost protection with sprinklers. Mich.State Coll., Extension
Bulletin, 327.
De Bruin, H. A. R., 1973: Gegevens betreffende neerslag en verdamping verzameld in
Salland gedurende 1970-1972. Royal Netherlands Meteorological Institute,
Scientific Report WR-73-4.
De Bruin, H. A. R., Keijman, J. Q., 1979: The Priestley-Taylor evaporation model
applied to a large shallow lake in the Netherlands. / ournal of Applied
Meteorology, 18, pp. 898-903.
De Bruin, H. A. R., Kohsiek, W., Van den Hurk, B. J. J.M., 1993: A verification of
some methods to determine the fluxes of momentum, sensible heat and water
vapour, using standard deviations and structure parameters of scalar meteoro-
logical quantities. Boundary-Layer Meteorology, 63, pp. 231-257.
De Bruin, H. A. R., Van den Hurk, B. J. J. M., Welgraven, A. D., 1995: A series of
global radiation at Wageningen for 1928-1992. International Journal of
Climatology, IS, pp. 1253-1272.
De Bruin H.A.R., Nieveen J.P., De Wekker S.F.J., Heusinkveld B.G., 1996: Large
aperture scintillometry over a 4.8 km path for measuring areally-average sensi-
ble heat flux - a case study. Prepr. 22nd AMS Con(. Agric. Forest Meteor. (Atlanta),
pp. 153-156.
De Bruin, H. A. R., Lablans, W. N., 1998: Reference crop evapotranspiration deter-
mined with a modified Makkink equation. Hydrological Processes, 12,
pp. 1053-1062.
Dehne, K., 1989: Preliminary results of the WMO automatic sunshine duration
measurement comparison 1988/89 in Hamburg. WMO Instrument of
Observation and Methods Report 35, (WMO/TD-No. 303) pp. 27-32.
Delaunay D., 1984: Cartographie regionale de l'energie eolienne - application a la
Bretagne. Centre Sc.Techn.Batiment, Nantes, France, Rep.EN-CLI-84.20.
Delecolle R., 1987: Cereal crop development models. Proc. Workshop on Models in
Agriculture and Forest Research (San Miniato, ed. F.Miglietta) pp. 107-111.
Deosthali V., Lomas, J ., 1996: An agroclimatic analysis of rainfed sorghum produc-
tion under semi-arid conditions (Maharashtra, India). Indian Journal of
Agricultural Research, 2, pp. 24-38.
De Vries, D. A., 1958: The thermal behaviour of soils. In: Climatology and
Microclimatology (UNESCO, Paris), pp. 109-113.
De Weille, G. A., 1964: Forecasting crop infection by the potato blight fungus. Meded.
Verb. Roy. Neth. Meteor. Inst., 82, 144 pp.
Donald, C. M., 1963: Competition among crop and pasture plants. /.Sci.exp.Biol.,
15, pp. 282-313.
Doorenbos, J., Pruitt, W. 0., 1975: Crop water requirements. FAO Irrigation and
Drainage Pap.24, FAO, Roma, 179 pp.
Doorenbos, J., Kassam, A. H., 1979: Yield response to water. FAO Irrigation and
Drainage Pap.33, FAO, Roma, 193 pp.
Duckham, A. N., 1974: Climate, weather and human food systems - a world
view. Weather, 29, pp. 242-257.

187
WMO LECTURE NOTES - AGRICULTIJRAL METEOROLOGY

Edwards, R. S., Claxton, S. M., 1964: The distribution of air-borne salt of marine
origin in the Aberystwyth area. Journal of Applied Ecology, 1, pp. 253-263.
Esteves A., De Rosa, C., 1989: A simple method for correcting the solar radiation
readings of a Robitzsch-type pyranometer. Solar energy, 42, pp. 9-13.
Fajer, E. D., Bazzaz, F. A., 1992: Is carbon dioxide a 'good' greenhouse gas? Global
Environmental Change, pp. 301-310.
Federer, B., et al., 1986: Main results of Grossversuch IV. Journal of Climate and
Applied Meteorology, ZSr pp. 917-957.
Fisher, R. A.,. 1924: The influence of rainfall distribution on the yield of wheat at
Rothamsted. Phil. Trans Roy. Soc. London B 213, pp. 89'--142.
Friehe,. C. A., 1986: Fine-scale measurements of velocity, temperature and humid-
ity in the atmospheric boundary layer. In: 'Probing the atmospheric boundary
layer (ed. D. H. Lenschow, publ. American Meteorological Society, Boston),
pp. 29;...38..
Fritschen, L. J., Gay, L. W., 1979: Environmental instrumentation. Springer Verlag,
N.Y., pp. 216.
Frost, W~, 1973: Review of data and prediction techniques for wind profiles around
man-made surface obstructions. AGARD/NATO Conf. Proc. No. 140 on 'Ftight
in turbulence', Paper 4, pp. 1-18.
Frost, J. R., Stevens, R. J., Laughlin,. R. J., 1990: Effect of separation and acidifica-
tion of cattle slurry on ammonia volatilization and on the efficiency of slurry
nitrogen for herbage production. Journal· of Agricultural Sdence, 115, pp. 46-56.
Gandemer, J., Guyot, A., 1981~ La protection contre le vent. Centre Scient.Techn.
Batiment, Nantes/Paris, pp .. 132.
Gandin, L. S., 1970: The planning of meteorological station networks. WMO TN 111
(WMO-No. 265).
Gardner, H. R., 1968: Availability and measurement of soil water. In: Water deficits
and plant growth~.Ed.TT. Kozlowski, pp. 107-135.
Garnier B.J., Ohmura A., 1968: A method of calculating the direct short wave radi-
ation income of slopes. Journal of Applied Meteorology, 7, pp. 796-800.
Garratt J.R., 1978: Flux profile relations above tall vegetation. Quarterly Journal of
the Royal Meteorological Society, 104, pp. 199-211.
Garratt J .R., 1990: The Internal • Boundary Layer - a review. Boundary-Layer
Meteorology, SO, pp. 171-203.
Gates D.M., 1968: Energy exchange in the atmosphere. Proc. Unesco Symp., Copenhagen
- Functioning of the terrestrial ecosystems at the primary production level.
Geiger R., 1965/1995: The climate near the ground. 4th ed.: Harvard Univ. Press, 611
pp.; 5th ed. (revised by R. H. Aron and P. Todhunter) Vieweg, 528 pp.
Georg, J. G., 1970: A numerical model for prediction of the nocturnal temperature
in the atmospheric surface layer. Journal of Applied Meteorology, 9, pp. 711-713.
Geslin, H., 1944: Etude des lois de croissance d'une plante en fonction des facteurs du
climat. Paris, Imprimerie Nationale, 116., Centre National de Recherches
Agronomiques, Service de Documentation, Versailles.
Geslin, H., Bouchet, R. J., 1966: Climat et production vegetale. La Bioclimatologie.
regards sur la France. Instit.Nation. Rech. Agronomique, Versailles, pp. 35-38.
Getz, R.R., 1992: Report on the measurement of leaf wetness. CAgM Report No. 38,
WMO/TD-No. 478, 10 pp.
Gillespie, T. J., Barr A., 1984: Adaption of a dew estimation scheme to a new crop
and site. Agricultural and forest meteorology, 31, pp. 289-295.
Gloyne, R. W., 1952: Daily maximum temperature of the surface of the ground.
Meteorological Magazine, 81, pp. 203-206.
Gloyne, R. W., 1953: Radiative minimum temperature over a grass surface and
over a bare soil surface. Meteorological Magazine, 82, pp. 263-267.
Gloyne, R. W., 1954: Temperature under glass. Sdent. Hort., XI (1952-54) pp. 158-162.
Gloyne, R. W., 1964: Some characteristics of the natural wind and their modifica-
tion by natural and artificial obstructions. Scient. Hortic., 17, pp. 7-19.
Gloyne, R. W., 1972: The diurnal variation of global radiation on a horizontal
surface - with special reference to Aberdeen (U.K.). Meteorological Magazine,
101, pp. 44-51.

188
REFERENCFS

Golder, D., 1972: Relations among stability parameters in the surface layer.
Boundary-Layer Meteorology, 3, pp. 47-58.
Grace, J., 1977: Plant response to wind. Academic Press, London, 204 pp.
Graser, E. A., 1985: Micrometeorology of sorghum at two row spacings. Ph.D.Thesis
Univ.Nebraska, 389 pp.
Gray, D. M., Male, D. H. (eds.), 1981: Handbook of snow - principles, processes,
management and use. Pergamon, 776 pp.
Greacen, E. L. (ed.), 1981: Soil water assessment by the neutron method. CSIRO
Div. of Soils, Adelaide, Australia, 140 pp.
Greeley, R., Iversen, J. D., 1985: Wind as a geological process. Cambridge Univ.Press,
U .K., 333 pp.
Griffiths, J. F. (ed.), 1994: Handbook of agricultural meteorology. Oxford University
Press, U.K., 320 pp.
Gunn, D. L., Yeo, D., 1951: Bellani spherical pyranometer or radiation integrator:
the calibration of an improved model and some suggestions for further
improvements. Quarterly Journal of the Royal Meteorological Society, 77,
pp. 293-301.
Hahn, L., McQuigg, J. D., 1970: Evaluating climate records for rational planning
of livestock shelters. Agricultural Meteorology, 7, pp. 131-141.
Hanson, K. J., 1963: The radiation effectiveness of plastic films for greenhouses.
Journal of Applied Meteorology, 2, pp. 793-797.
Harrington, R., Woiwod, I. P., 1995: Insect crop pests and the changing climate.
Weather, SO, pp. 200-208.
Haude, W., 1940: Ergebnisse der allgemeinen met. Beobachtungen ... Rep. Sc.
Exped. to N. W. Prov. China (Sven Hedin), IX Meteorology, l, Stockholm 1940.
Haverkort, A. J., 1987: Climate and potato crop growth interactions in Africa's
continental divide region. Acta Hortic., 214, pp. 137-147.
Heusinkveld B.G., 1998: New optical method to measure leaf wetness. Prepr.lOth
American Meteorological Society Symposium Meteorological Observations and
Instrumentation, Phoenix (USA), p. 300.
Hewson, W. H., Longley, R. W., 1944: Meteorology: theoretical and applied. Wiley,
N.Y., 468 pp.
Hickman, M. J., 1970: Measurement of humidity. Nation.Phys.Lab.Note Appl.Sci., 4,
HMSO (H.M. Stationery Office), London, 32 pp.
Hiemstra, G., Bor N., Harbers, J., Mauritz, R., Peters, G., Wartena, L., 1994: Weerkunde
voor de agrarische sector. Educ.Partn.Ned., Houten, Netherlands, 245 pp.
HMSO (London), 1982: Observers Handbook (4th ed.), 230 pp., ISBN O 11 400329 7
Hogg, W. H., 1964: Frost prevention in Dutch Light frames. Agricultural
Meteorology, l, pp. 121-129.
Hogg, W. H., 1968: Climate and surveys of agricultural land use. Proc. Unesco Symp.
Methods in Agroclimatology, Reading'66, pp. 281-289.
Holmes, J. W., 1958: Aspects of soil moisture measurement with reference to arid
soils. Proc. Symp. on Arid Zone Climatology, Unesco, (1), pp. 295-300.
Holmes, J.W., Turner, K. G., 1958: The measurement of water content of soils by
neutron scattering. A portable instrument for field use. Journal of Agricultural
Engineering Research, 3, pp. 199-204.
Holtslag, A. A. M., Van Ulden, A. P., 1983: A simple scheme for daytime estimates
of the surface fluxes from routine weather data. Journal of Climate and Applied
Meteorology, 22, pp. 517-529.
Hooghart, J.C. (ed.), 1987: Evaporation and weather. TNO Comm.Hydro/. Res.Proc.,
39, 98 pp.
Howell, T. A., Phene, C.J., Meek, D. W. and Miller, R.J., 1983: Evaporation from screened
Class A pans in a semi-arid climate. Agricultural Meteorology, 29, pp. 111-124.
Hudson, N., 1971: Soil conservation. Batsford Ltd., London, 320 pp.
Htitte, P., 1968: Experiments in wind-throw and wind-damage in Germany: site
and susceptibility of spruce forests to storm damage. In: 'Forestry' Supplement
- Wind Effects in the Forest. Ix. Univ. Press, pp. 20-26.
Idso, S. B., Jackson, R. D., 1969: Thermal radiation from the atmosphere. Journal
of Geophysical Research, 74, pp. 5397-5403.

189
WMO LEC:TIJRE' NOTES - AGRfCULTtJRAL METEOROLOGY

Idso, S. B., Allen, S. G., Choudhury, B. J., 1988: Problems with porometry: meas-
uring stomata! conductance of potentially transpiring plants. Agricultural and
Forest Meteorology, 43, pp. 49-58.
Iqbal, M., 1983: An introduction to solar radiation. Acad.Press Canada, 390 pp.
Jacobs, A. F. G., 1985: Turbulence around a thin solid fence. Agricultural and Forest
Meteorology, 34, pp. 315-321
Jacobs, A. F. G., Nieveen, J. P., 1995: Formation of dew and the drying process
within crop canopies. Meteorological Applications, 2, pp. 249-256.
Jensen, M., 1954: Shelter effect. Danish Tech.Press, Copenhagen, 264 pp.
Johnson, L., 1965: Ann. Rep.for 1964~ Internat. Rice Res. Inst., pp. 185-194.
Kaempfert, W., 1951: Ein Phasendiagramm der Besonnung. Meteor. Rundschau.,. 4,
pp. 141-144
Kaimal, J. C., Finnigan, J. J., 199'4: Atmospheric boundary layer flows, their structure
and measurement. Oxford' Univ.Press, 289 pp.
Kaiser, H., 1959: Die Stromung an Windschutzstreifen. Ber. D. Wetterd., 7, Nr. 53.
Kasten, F., Czeplak, G., 1980: Solar and terrestrial radiation dependent on the
amount and type of cloud. Solar energy, 24, pp. 177-189.
Keane, T., 1986: Climate, weather and Irish agriculture. AGMET, Fae. Agriculture,
Univ: Dublin, Ireland; 3-29 pp.
Kerr, J.P., ThurteU, G. W., Tanner, C. B., 1967: Integrating pyranometer for clima-
tological observer stati?Ons and me~soscaTe networks. Journal of Applied
Meteorology, 6, pp. 688-694.
Kessler, E., Wilk, K. E., 1968: Radar measurement of precipitation for hydrological
purposes. WMO/lHD Report No. 5.
Kessler, E., 1993: Wind chill errors. Bulletin of the American Meteorological Society,
74, pp. 1743-1744.
Kind, R. J., 1990: Mechanics of aeolian transport of snow and sand. f. Wind Engin.
Ind. Aerodyn., 36, pp. 855-866.
King, E., 1970: Beitrage zum Gewii.chshausklima. In: Der Erwerbsgartner 24 and 43
(1969) and 17.
Knighting, E., 1950: A note on nocturnaJ cooling. Quarterly Journal of the Royal
Meteorological Society, 76, pp. 173-181.
Kondratyev, K. J., Manolova, M. P., 1960: The radiation balance of slopes. Solar
Energy, 4, 14 ..
Kondratyev, K. J ., 1969: Radiation in the atmosphere. Academic Press.
Kramer, P. J., Boyer, J. S., 1995: Stomata and gas exchange. In: 'Water relations of
plants and soils'. Academic Press, pp. 257-272.
Kramer, M. L., Porch, W. M. ( ed.), 1990: Meteorological aspects· of emergency response.
American Meteorological Society, 119 pp.
'Kroon, L. J.M., De Bruin, H. A. R., 1993: Atmosphere-vegetation interaction in
local advection conditions: effect of lower boundary conditions. Agricultural
and Forest Meteorology, 64, pp. 1-28.
Kruizinga, S., Yperlaan, G. J., 1978: Spatial interpolation of daily totals of rainfall.
Journal of Hydrology, 36, pp. 65-73.
Landsberg, H., 1968: Physical climatology (2nd ed.). Gray Printing, Penn., USA.
Lawrence, E. N., 1955: Effects of a windbreak on the speed and direction of the
wind. Meteorological Magazine, 84, pp. 244-251.
Leopold, A. C., 1964: Plant growth and development. McGrawHill, USA, 466 pp.
Lewis, T., 1965: Effect of an artificial wind-break on the areal distribution of flying
insects. Annals Appl. Biol., 55, pp. 503-512.
Linacre, E., (1992): Climate data and resources. Routledge, USA, ISBN 0-415-05703-5,
366 pp.
Lomas, J., Gat, Z., 1967: The effect of wind-borne salt on citrus production near
the sea in Israel. Agricultural Meteorology, 4 (6), pp. 415-425.
Lomas-J., Shashoua Y., Cohen A., 1969: Mobile surveys in agrotopoclimatology.
Meteor. Rundschau, 22, 96--101.
Lomas, J., Shashoua, Y., 1973: The effects of rainfall on wheat yields in an arid
region. In: Plant response to climatic factors, Proc. Uppsala Symposium, Unesco,
pp. 531-HS.

190
REFERENCES

Lomas, J., Mandel, M. M., 1973: The quantitative effects of two methods of sprin-
kler irrigation on the microclimate of a mature ·avocado plantation. Agricultural
Meteorology, 12, pp. 35-48.
Lomas, J., 1977: Topoclimatology as an aid in the siting of the avocado crop. Israel
Meteorol. Res. Papers, 1, pp. 94-100.
Lomas J., Levin J., 1979: Irrigation. In: Seemann, J., et al., 'Agrometeorology',
Springer, pp. 217-264.
Lomas, J., 1984: Efficient utilization of climate resources for rainfed agriculture in
Israel. Bull. Soc. Botanique, Fr.131, pp. 501-515.
Lomas, J ., 1988: An agrometeorological model for assessing the effect of heat
stress during the flowering and early fruit set on avocado yields. /.Am.Soc.
Hort.Sci., 113, pp. 172-176.
Lomas, J ., 1992: Analysis of effects of heat stress during flowering on the yield of
avocado under Mediterranean climatic conditions. Agricultural and Forest
Meteorology, 59, pp. 207-216.
Lomas, J., Zamet, D. A., 1994: Long-term analysis and modelling of agroclimatic
effects on national avocado yields in Israel. Agricultural and Forest Meteorology,
71, pp. 315-336.
Lomas, J., 1997: The phenology of the flowering process of the avocado pear and
its application. The Fruit Grower (Florida Grower Ass.), 4, pp. 74-89.
Lowry, W. P., 1967: Weather and life. Academic Press, N.Y., 306 pp.
Lowry, W. P., Lowry, P. P., 1989: Fundamentals of biometeorology. Vol.1: The physi-
cal environment. Peavine Publ., Oregon, USA), 310 pp.
Machardy, W. E., Gadourny, D. M., 1989: A revision of Mills's criteria for predict-
ing apple scab infection periods. Phytopathology, 79, pp. 304-310.
Mak.kink, G. F., 1957: Ekzameno de la formula de Penman. Netherlands Journal of
Agricultural Science, 5, pp. 290-305.
Marshall J. K., 1967: The effect of shelter on the productivity of grasslands and
field crops. Field Crop Abstracts, 20, pp. 1-14.
Marshall, J. K., 1972: Drought, land-use and soil erosion. In: Drought,
its Perception, Effects and Relief. Ed. J. V. Lovell, Angus and Robertson,
Melbourne.
Martinez-Lozano, J. A., Tena, F., Onrubia, J.E., De la Rubia, J., 1984: The historical
evolution of the Angstrom formula and its modifications - review and bibli-
ography. Agricultural and Forest Meteorology, 33, pp. 109-128.
Maselli, F., Petkov, L., Maracchi, G., 1997: Extension of climate parameters over
the land surface by the use of NOAA-AVHRR and ancillary data. photogramme-
try engineering and remote sensing (in press).
Mason, B. J., 1975: Clouds, rain and rainmaking (2nd ed.). Cambridge Univ. Press,
U.K., 200 pp.
Mazzarella, D. A., 1972: An inventory of specifications for wind measuring instru-
ments. Bulletin of the American Meteorological Society, 53, pp. 860-871.
McAdam, J. H., 1980: Tatter flags and climate in the Falkland Islands. Weather, 35,
pp. 321-327.
McCulloch, J. S. G., Wangati, F. J., 1967: Notes on the use of the Gunn Bellani
radiometer. Agricultural Meteorology, 4, pp. 63-70.
Mcllveen, R., 1992: Fundamentals of weather and climate (2nd ed.). Chapman &
Hall, U.K., 497 pp.
McIntosh, D. H., Thom, A. S., 1969: Essentials of meteorology. Wykeham Publ.,
U.K., pp. 238.
McNaughton, K. G., Jarvis, P. G., 1984: Using the Penman-Monteith equation
predictively. Agricultural Water Management, 8, pp. 263-278.
McNaughton, K. G., Jarvis, P. G., 1991: Effects of spatial scale on stomatal control
of transpiration. Agricultural and Forest Meteorology, 54, pp. 279-301.
McQuigg, J. D., 1968: A review of problems, progress and opportunities in the use
of weather information in agricultural management. In: Agroclimatological
Methods. Proc. Unesco Symp., Reading, England.
Meteorological Department, U.AR., 1970: Sand storms and dust storms in U.A.R.
Technical Note No. 1. Cairo, 35 pp.

191
WMO LECTURE NOTES - AGRICULTURAL METEOROLOGY

Meteorological Office, 1981: Handbook of meteorological instruments (2nd ed.,

8 volumes). Publ.Met.O.919, H.M. Stationery Office, London, U.K.
Mohammed, A. E., Stigter C.J., Adam H.S., 1996: On shelterbelt design for combat-
ing sand invasion. Agricultural Ecosystem Environment, 57, pp. 81-90.
Molga, M., 1962: Agricultural meteorology II: outline of agrometeorological problems.
Centr. Inst. Inf. Nauk.-Techn. Ekonom., Warszawa, 253 pp.
Monteith, J. L., 1957: Dew. Quarterly Journal of the Royal Meteorological Society, 83,
pp. 322-341.
Monteith, J. L., 1965: Evaporation and environment. In: 'State and movement of water
in living organisms', 19th Symp. Soc. Exp. Biol., Academic Press, pp. 205-234.
Monteith,J. L., 1966: Physical limitations to crop growth.Agric. Progress, 41, pp. 9-23.
Monteith, J. L., 1968: Analysis of the photosynthesis and respiration of field crops
· from vertical fluxes of carbon dioxide. Proc. Symp. Functioning of terrestrial
ecosystems at the primary production level, Copenhagen, pp. 349-355.
Monteith, J. L., 1975: Principles of environmental physics. Edward Arnold, U.K.,
pp. 241.
Mount, L. E., 1979: Adaptation to thermal environment: man and his productive
animals. Edward Arnold, U.K.
Munn, R. E., 1970: Biometeorological methods. Academic Press, N.Y., 336 pp.
Munn, R. E., Rodhe H., 1985: Air chemistry and air pollution meteorology
(Compendium of lecture notes for Class I and Class II). WMO-No. 364, Vol.II, 6.
Nadler, A., Dasberg, S., Lapid, I.., 1991: Time Domain Reflectometry measurements
of-water content and electrical conductivity of layered soil columns. Soil Science
Society of American Journal, 55, pp. 938-943.
Nagel, J. F., 1962: On the measurement of dew. Arch.Met.Geoph.Biokl.B, 11,
pp. 403-423.
Nageli, W.,. 1943: Untersuchungen tiber die Windverhaltnisse im Bereich von
Windschutzstreifen. Mitt.Schweiz. Anst. forstl. Versuchw., 23.
Nieuwstadt, F. T. M., Van Dop, H. (ed.), 1982: Atmospheric turbulence and air pollu-
tion modelling. Reidel, Dordrecht, 358 pp.
Oke, T. R., 1987: Boundary layer climates (2nd ed.). Methuen, London, 435 pp.
Onyewotu, L.O.Z., Ogigirigi M.A., Stigter CJ., 1994: A study of competitive effects
between a Eucalyptus shelterbelt and an adjacent millet crop. Agricultural
Ecosystem Environment, 51, pp. 281-286.
Orlanski, I., 1975: A rational subdivision of scales for atmospheric processes.
Bulletin of the American Meteorological Society, 56, pp. 527-530.
Page, J. K. (ed.), 1986: Prediction ofsolar radiation on inclined surfaces. Solar Energy
R&D European Community, Series F vol.3. Reidel, Dordrecht, 459 pp.
Paltridge, G. W., Platt, C. M. R., 1976: Radiative processes in meteorology and clima-
tology. Elsevier, 318 pp.
Parratt, L. G., 1961: Probability and experimental errors in science. Wiley, New York.
Pasquill, F., Smith, F. B., 1983: Atmospheric diffusion (3rd ed.). Ellis Horwood, U.K.
Pedgley, D., 1982: Windborne pests and diseases - meteorology of airborne organisms.
Ellis Horwood, Chichester, 250 pp.
Pedro, M. L Gillespie, T. J., 1982: Estimating dew duration, utilizing standard
weather station data. Agricultural Meteorology, 25, pp. 297-310.
Penman, H. L., 1943:. Daily and seasonal changes in the surface temperature of
fallow soil at Rothamsted. Quarterly Journal of the Royal Meteorological Society, 69,
pp. 1-16.
Penman, H. L., 1948: Natural evaporation from open water, bare soil and grass.
Proc. Roy. Soc., London, Ser. A 193, pp. 120-145.
Penman, H. L., Schofield, R. K., 1951: Some physical aspects of assimilation and
transpiration. Symp. Soc. Exper. Biol., 5, 115-129.
Penman, H. L., 1956: Evaporation, an introductory survey. Netherlands Journal of
Agricultural Science, 4, pp. 9-29.
Perarnaud, V., Raynal, N., 1991: Agrometeorologie. Meteor.Nation. Cours et
Manuels, 4, 183 pp.
Pereira, H.G., 1959: Practical field instruments of radiation and evaporation.
Quarterly Journal of the Royal Meteorological Society, 85, pp. 253-261.

192
REFERENCES

Perry, A. H., Schofield, W. H., Joubert, P. N., 1969: Rough-wall turbulent bound-
ary layers./. Fluid Mech., 37, pp. 383-413.
Peterson, R. F., 1965: Wheat. Interscience Puhl.Inc., New York.
Petkov, L., Pieri, M., Maselli, F., Maracchi, G ., 1996: Study of modelling of temper-
ature spatial variability by NOAA-AVHRR thermal imagery. Photogrammetry
Engineering and Remote Sensing, 51, pp. 127-136.
Pike, W. S., 1989: One hundred years of the Dines pressure tube anemometer.
Meteorological Magazine, 118, pp. 209-214.
Platt, R. B., Griffiths, J. F., 1964: Environmental measurements and interpretation.
Reinhold, New York, 235 pp.
Popov, A. M., 1975: Modelling of the planetary boundary layer of the atmosphere
in the roughness layer. Atm.Oc.Phys., 11, pp. 354-358.
Popov, G. F., 1990: Agrometeorological crop monitoring and yield forecasting. In:
Price-Budgen A., ed.; 1990: 'Using meteorological information and products'. Ellis
Horwood (N.Y.) and WMO, pp. 427-435.
Price-Budgen, A., ed.), 1990: Using meteorological information and products. Ellis
Horwood (N.Y.) and WMO, 487 pp.
Priestley, C. H. B., Taylor, R. J., 1972: On the assessment of surface heat
flux and evaporation using large-scale parameters. Monthly Weather Review, 100,
pp. 81-92.
Primault, B., 1972: Etude mesoclimatique du Canton de Vaud. Office Cantonal
Vaudois de l'Urbanisme, Lausanne 1972, 185 pp.
Rakovec, J., 1989: Thunderstorms and hail. Theor. Appl. Climatol., 40,
pp. 179-186.
Rakovec, J., Gregorcic B., Kranjc A., Mekinda T., Kajfez-Bogataj L., 1990: Some
evaluations of hail suppression system efficiency in Slovenia, Yugoslavia. Theor.
Appl. Climatol., 41, pp. 157-171.
Ramdas, L. A., Dravid, R. K., 1936: Soil temperature in relation to other factors
controlling the disposal of solar radiation at the earth's surface. Indian /. Agr.
Sci., 2, pp. 131-143.
Ramdas, L.A., 1961: Crops and weather in India. Ind. Council for Agr. Res., New
Delhi, 127 pp.
Raupach, M. R., Thom, A.S., 1981: Turbulence in and above plant canopies.
Annual Review Fluid Mech. 13, pp. 97-129.
Reiff, J., Forbes, G. S., Spieksma, F. T. M., Reynders, J. J., 1986: African dust reach-
ing northwestern Europe: a case study to verify trajectory calculations. Journal
of Climate and Applied Meteorology, 25, pp. 1533-1567.
Reifsnyder, W. E., Abers, B., 1994: Systems for evaluating and predicting the effects of
weather and climate on wild/and fires. WMO-No. 496, 2nd ed.
Rietveld, M. R., 1978: A new method for estimating the regression coefficients in
the formula relating solar radiation to sunshine. Agricultural Meteorology, 19,
pp. 243-252.
Rijtema, P. E., 1965: An analysis of actual evapotranspiration. Agric. Res. Rep., 659,
Pudoc, Wageningen.
Riley, J. R., Giles, W. L., 1965: Agricultural meteorology in relation to the use of
pesticides in the USA. Agricultural Meteorology, 2, pp. 225-245.
Robertson, W. K., Hammond, L. C., Johnson, J. T., Boote, K. J., 1980: Effects of
plant water stress on root distribution of corn, soybean and peanuts in sandy
soil. Agronomy Journal, 72, pp. 548-550.
Robinson, N. (ed.), 1966: Solar radiation. Elsevier, 347 pp.
Rogers, W.S., Modlibowska I., 1954: Low temperature injury to fruit blossom - IV.
Further experiments in water sprinkling as an anti-frost measure. /. Hort. Sci.,
24, pp. 126-141.
Rosenberg, N. J., Blad, B. L., Verma, S. B., 1983: Microclimate: The biological
environment (2nd edition). Wiley, U.S.A., 495 pp.
Rumney, R.P., 1986: Meteorological influences on the spread of foot-and-mouth
disease. /.Appl.Bacterial., Sympos.Suppl. 105S-l 14S.
Rutter, N., 1968: Geomorphic and tree shelter in relation to surface wind condi-
tions, weather, time of day and season - shelter effect of an old-established

193
WMO LECTURE NOTES - AGRICULTIJRAL METEOROLOGY

shelter block of European larch on a slope of mean gradient 1 in 4. Agricultural

Meteorology, S, pp. 319-349.
Sadek, M. F., Shahin, M. M., Stigter, C. J., 1997: Evaporation from the reservoir of
the High Aswan Dam, Egypt: a new comparison of relevant methods with
limited data. Theor.Appl.Climatol., 56, pp. 57-66.
Sarker, R. P., Biswas, B. C., 1986: Agroclimatic classification for assessment of
crop potential and its application to dry farming tract of India. Mausam, 3 7,
pp. 27-38.
Scaetta, H., 1935: Les avalanches d'air dans les Alpes et dans les hautes montagnes
de 1' Afrique centrale. CieL et Terre, SI, pp. 79-80.
Scharringa, M., 1976: Nachtvorst. Royal Netherlands Meteorological Institute,
Verspr.Opstellen, 5.
Schnelle, F., 1948:- Einfilhrung in die Probleme der Agrometeorologie. Ed. Ulmar,
Stuttgart, p. 154.
Schnelle, F., 1955: Pflanzen-Phiinologie. Akad. Verl. Ges., Leipzig, 299 pp.
Scorer, R. S., 1978: Environmental aerodynamics. Ellis Horwood, 488 pp.
Seemann, J., 1973: Climate under glass. WMO Technical Note No. 131, 40 pp.
Seghi, L., Orlandini, S., Gozzini, B., 1996: Survey of simulation models of plant
disease. CeSIA-IATA, Firenze, 167 pp.
Seginer, I., 1975a: Atmospheric-stability effect on windbreak shelter and drag.
Boundary-Layer MeteoroloKif, 8~ pp. 383-400.
Seginer, I., 1975b: Flow around a windbreak in oblique wind. Boundary-Layer
Meteorology, 9, pp. 133-141
Sellers, W. D., 1965: Physical climatology. Univ.of Chicago Press, 272 pp.
Silva, A. M., Rosa, R., 1985: Radiative environment inside a greenhouse.
Agricultural and Forest Meteorology, 33, pp. 339-346.
Slatyer, R. 0., Mcllroy, I. C., 1961: Practical microclimatology-with special reference
to the water factor in soil-plant-atmospheric relationships. CSIRO, Australia and
Unesco, Paris, pp. 175 + appendices.
Slatyer, R. 0 ., 1969: Physiological significance of internal water relations to crop
yields. In: Physical aspects of crop growth, Ed. R. C. Dinauer, pp. 53-58.
Smith, F.B., Carson D.J., Oliver H.R., 1972: Mean wind-direction shear through a
forest canopy. Boundary-Layer Meteorology, 3, pp. 178-190.
Smith, L. P., Gloyne, R. W., 1951: Shielded thermometer mounts. Meteorological
Magazine. 80, pp. 203-204.
Smith, L. P., 1975: Methods in agricultural meteorology. Elsevier, 210 pp.
Spiegel, M. R., 1961: Theory and problems of statistics. Schaum's Outline Series
(McGrawHill), 3-59 p.p.
Stanhill, G., 1965: The concept of potential evapotranspiration in arid-zone agriculture.
Proc. Montpellier Unesco Syrup., Meth. ofPlant Ecophysiology, pp. 109-117.
Stanhill, G., 1965: Observations on the reduction of soil temperature. Agricultural
Meteorology, 2, 197-204.
Stanhill, G., 1983: The distribution of global solar radiation over the land surfaces
of the earth. Solar Energy, 31, 95-10-4.
Sterk, G., Herrmann, L., Bationo, G., 1996: Wind-blown nutrient transport and
soil productivity changes in southwest Niger. Land Degradation Development, 7,
pp. 325-335.
Sterk, G., Spaan, W. P., 1997: Wind erosion control with crop residues in the Sahel.
Soil Science Society of America Journal, 61, 911-917
Stigter, C. J., Uiso, C. B. S., 1981: Understanding the Piche evaporimeter as a
simple integrating mass transfer meter. Appl. Sci. Res., 37, pp. 213-223.
Stigter, C. J., 1983: An explanation of uncertainties in point cloudiness/solar
energy relationships. Journal of Climate and Applied Meteorology, 22,
pp. 336-338.
Stfgter, C. J., 1984: Mulching as a traditional method of microclimate manage-
ment. Arch. Meteor. Geph. Biokl., B35, pp. 147-154.
Stigter, C. J., Kainkwa, R. M. R., Musabilha, V. M. M., 1987: A near-equatorial
comparison of two instruments measuring diffuse solar radiation. Zeitschr.
Meteor., 37, pp. 161-164.

194
REFERENCES

Stigter, C. J., Jiwaji, N. T., Kainkwa, R. M. R., Juma, F. S., Musabilha, V. M. M .,

Mabuba, K. I., Waryoba, J. M., 1989: The National Radiation Measurement
and Calibration Centre in Dar-es-Salaam, Tanzania: an appraisal of the
early work. Proc. TECIMO IV, WMO Instrument Observation Method Report 35
(WMO-T.P. No. 303), pp. 371-376.
Strangeways, I. C., 2000: Measuring the natural environment Cambridge Univ. Press,
ISBN 0-521-57310-6.
Stricker, H., Brutsaert W., 1978: Actual evapotranspiration over a summer period
in the Hupsel catchment. Journal of Hydrology, 39, pp. 139-157.
Stull, R. B., 1988: An introduction to boundary layer meteorology. Kluwer, 666 pp.
Suomi, V. E., Franssila, M. and Islitzer, N. F., 1954: An improved net radiation
instrument. Journal of Meteorology, 11, pp. 276-282.
Sutton, 0. G., 1953: Micrometeorology. McGrawHill, London, 334 pp.
Svabik, 0., 1989: Review of meteorological aspects of hail defense in Austria.
Theor. Appl. Climatol., 40, pp. 247-254.
Swinbank, W. C., 1963: Longwave radiation from clear skies. Quarterly Journal of
the Royal Meteorological Society, 89, pp. 339-348.
Tarpley, J. D., 1990: Use of weather satellite data in agricultural applications : solar
radiation, max/min temperature, and daily precipitation. In: Weather satellites:
systems, data and environmental applications, P. Krishna Rao, et al. (eds); publ.
American Meteorological Society) pp. 460-466.
Tennekes, H., 1973: The logarithmic wind profile. Journal of Atmospheric Sciences,
30, pp. 234-238.
Thomas, G. A. et 10 al., 1981: Evaluation of row spacing and population density
effects on grain sorghum over a range of northern Australian environments.
Austr. J. Exp. Agric. Anim. Husb., 21, pp. 210-217.
Thornthwaite, C. W., Holzman, B., 1939: The determination of evaporation from
land and water surfaces. Monthly Weather Review, 67, pp. 4-11.
Thornthwaite, C. W., 1948: An approach towards a rational classification of
climate. Geographical Review, pp. 55-94.
Thornthwaite, C. W., 1954: Topoclimatology. In: Proc. Toronto Meteor. Conference
1953, Royal Meteorological Society, London, pp. 227-232.
Thornthwaite, C. W., Mather, J. R., 1955: Instructions and tables for computing
potential evapotranspiration and the water balance. Publications in Climatology
X (3), Lab. of Climatology, N.Jersey, USA.
Tisdall, A.L., 1951: Comparison of methods of determining apparent density of
soils. Austr. f. Agric. Res., 2, pp. 349-354.
Topp, G. C., Davis, J. L., Annan, A. P., 1980: Electromagnetic determination of soil
water content: measurements in coaxial transmission lines. Water Resource
Research, 16, pp. 574-582.
Trewartha, G. T., 1954: An introduction to climate. McGraw Hill, N.Y., 402 pp.
Van Arsdel, E. P., 1965: Micrometeorology and plant disease epidemiology.
Phytopathology, 55, pp. 945-950.
Van Eimern, J., 1968: Soil climate relationships. The dependence of soil tempera-
ture on radiation, albedo, moisture and agricultural practices. Proc. WMO
Seminar, Melbourne, pp. 491-522.
Van Ulden, A. P., Holtslag, A. A. M., 1985: Estimation of boundary layer parame-
ters for diffusion applications. Journal of Climate and Applied Meteorology, 24,
pp. 1196-1207.
Van Wijk, W. R. (ed.), 1963: Physics of plant environment. North Holland Publ. Co.,
Amsterdam, 382 pp.
Ventskevich, G. Z., 1961 : Agrometeorology. Publ. Isr.Progr.for Sci.Transl., Jerusalem,
300 pp. (Russian orig.: Gidrometeoizdat 1958).
Verstegen, M. W. A., Henken, A. M., Van der Hel, W., Brandsma, H. A., 1987:
Thermal requirements of growing pigs from birth to slaughter. In: Energy
metabolism in farm animals (publ.Martinus Nijhoff), pp.133- 149.
Waggoner P.E., Pack A.B., Reifsn yder W.E., 1959: The climate of shade- a tobacco
tent and a forest stand compared to open fields. Connecticut Agr. Expr. Stn. (New
Haven) Bulletin, 626, 39 pp.

195
WMO LECTIJRE NOTES - AGRICULTIJRAL METEOROLOGY

Waggoner, P. E., Miller, PM., De Roo, H. C., 1960: Plastic mulching principles and
benefits. Conn. Agr. Expr. Stn. (New Haven) Bulletin, 634, 44 pp.
Wallace, J.M., Hobbs, P. V., 1977: Atmospheric science, an introductory survey. Academic
Press.
Wang, J. Y., 1967: Agricultural meteorology. Agriculture Weather Information Service,
Calif., USA, 693 pp.
Wang,J. Y., Bruton, M. D., Luchessa, C. E., Roper, T.J., 1982: The grower's weather guide
for farming practices. Milieu Information Service, San Jose, Cal., USA, 64 pp.
Wartena, L., 1996: T-som te veel autoinatisme. Boerderij, 81, No. 20, pp. 22.
Wieringa, J., Van Lindert, F. X. C. M., 1971: Application limits of double-fin and
coupled wind vanes. Journal of Applied Meteorology, 10, pp. 137-145.
Wieringa, J., 1976: An objective exposure correction method for average wind
speeds measured at a sheltered location. Quarterly Journal of the Royal
Meteorological Society, 102, pp. 241-253.
Wieringa, J., 1977: Wind Tepresentativity increase due to an exposure correction,
obtainable from past analog station wind records. Proc. TECIMO Con(,
WMO-No. 480, pp. 39-44.
Wieringa, J., 1980: Representativeness of wind observations at airports. Bulletin of
the American Meteorological Society, 61, pp. 962-971.
Wieringa, J., 1986; Rm~ghness-dependent geographical interpolation of surface
wind speed averages. Quarterly Journal of the Royal Meteorological Society, 112,
PP~867-889.
Wieringa, J., 1992: Updating the Davenport roughness classification. J. Wind
Engin. Industr. Aerodyn., 41, pp. 357-368.
Wieringa, J., 1993: Representative roughness parameters for homogeneous terrain.
Boundary-Layer Meteorology, 63, pp. 323-363.
Wieringa, J., 1996: Does representative wind information exist? f. Wind Engin.
Industr. Aerodyn., 65, pp. 1-12.
Wieringa, J., 1996b: Is agrometeorology used well in European farm operations?
EC-DG-XII Rep. No.711/DOC D, 24 pp . .
Wilson, J. W., Brandes, E. A., 1979: Radar measurement of rainfall - a summary.
Bulletin of the American Meteorological Society, 60, pp. 1048-1058.
Wollny, E., 1878: Untersuchungen iiber den Einfluss der Exposition auf die
Erwarmung des Bodens. Forsch a.d. Geb.d. Agrik. Physik., l, pp. 263-294.
Woodward, F. I., 1987: Climate and plant distribution. Cambridge Univ. Press,
174 pp.
Yakuwa, R., 1945: Uber die Bodentempera~ren in den verschiedenen Bodenarten
in Hokkaido. Geophysical Magazine, 14, pp. 1-12.
Yang, S., Logan, J., Coffey, D. L., 1995: Mathematical formulae for calculating the
base temperature for growing degree days. Agricultural and Forest Meteorology,
74, pp. 61-74.
Yoshino, M. M., 1975:·Climate in a small area. Univ.of Tokyo Press, 549 pp.
Zemel, Z., Lomas, J., 1976: An objective method for assessing representativeness
of a station network measuring minimum temperature near the ground.
Boundary-Layer Meteorolqgy, 10, pp. 3-14.

f96