Wmo 100

WORLD METEOROLOGICAL ORGANIZATION
INTRODUCTION TO
CLIMATE CHANGE:
LECTURE NOTES FOR
METEOROLOGISTS
Prepared by
David D. Houghton
10.0 360
354(1990)
7.5 320
5.0 280
2.5 240
0.0 200
-2.5 160
-5.0 120
-7.5 80
-10.0 40
160 120 80 40 0
WMO-No. 926
Secretariat of the World Meteorological Organization
Geneva – Switzerland
WMO TECHNICAL PUBLICATIONS
relating to education and training
WMO No.
114 — Guide to qualifications and training of meteorological personnel employed in the provision of meteorological services
for international air navigation. 2nd edition, 1974. (French–Spanish)
258 — Guidelines for the education and training of personnel in meteorology and operational hydrology. 4th edition, 2001, in
preparation. (English)
266 — Compendium of lecture notes for training Class IV meteorological personnel. Volume I—Earth science; 1970. (English);
Volume II—Meteorology; 1984. (English–French)
364 — Compendium of meteorology for use by Class I and Class II meteorological personnel.
Volume I, Part 1—Dynamic meteorology. (French–Spanish), Part 2—Physical meteorology. (French–Spanish),
Part 3—Synoptic meteorology. (English–French), Volume II, Part 1—General hydrology. (English), Part 2—Aeronautical
meteorology. (English–French–Spanish), Part 3—Marine meteorology. (English–French–Spanish), Part 4—Tropical
meteorology. (English), Part 5—Hydrometeorology. (English), Part 6—Air chemistry and air pollution meteorology.
(English–French–Spanish)
182 — International meteorological vocabulary. Second edition, 1992. (E/F/R/S)
385 — International glossary of hydrology. Published jointly by WMO and UNESCO; 2nd edition, 1992
407 — International cloud atlas. Volume I—Manual on the observation of clouds and other meteors. Reprinted in 1995.
Volume II (plates), 1987.
551 — Lecture notes for training Class II and Class III agricultural meteorological personnel. 1980 edition. (Spanish)
593 — Lecture notes for training Class IV agricultural meteorological personnel. 1982 edition. (English–French–Spanish)
622 — Compendium of lecture notes on meteorological instruments for training Class III and Class IV meteorological personnel.
1986 edition. Volume I., Part 1—Meteorological instruments, Part 2—Meteorological instruments maintenance
workshops, calibration laboratories and routines. Volume II, Part 3—Basic electronics for the meteorologist. (English)
649 — El Niño phenomenon and fluctuations of climate—Lectures presented at the thirty-sixth session of the WMO Executive
Council (1984), 1986. (English)
659 — Marine cloud album. 1987 edition. (English)
669 — Workbook on numerical weather production for the tropics for the training of Class I and Class II—meteorological
personnel. 1986 edition. (English–Spanish)
701 — Mesometeorology and short-range forecasting lecture notes and students’ workbook for training Class I and Class II—
meteorological personnel. Volumes I and II. (English, 1990; Russian, 1988)
712 — Mesoscale forecasting and its applications—Lectures presented at the fortieth session of the WMO Executive Council
(1988). 1989. (E/F/R)
726 — Compendium of lecture notes in climatology for Class III and Class IV personnel. Part I—Lecture notes; Part II—Student’s
workbook; Part III—Notes for instructors. 1992 edition.
738 — Meteorological and hydrological risk assessment and disaster reduction—Lectures presented at the forty-first session of
the WMO Executive Council (1989). 1991. (E/R)
770 — Methods of interpreting numerical weather prediction output for aeronautical meteorology TN-No. 195 (2nd edition).
1999.
771 — Special topics on climate—Lectures presented at the forty-second session of the WMO Executive Council (1990).
1993. (E/R)
795 — Scientific lectures presented at the Eleventh World Meteorological Congress (1991). 1993
798 — Climate change issues—Lectures presented at the forty-fourth session of the WMO Executive Council (1992). 1994.
(English)
805 — Lectures presented at the forty-fifth session of the WMO Executive Council (1993). 1994. (E/F)
822 — Lectures presented at the forty-sixth session of the WMO Executive Council (1994). 1995. (E/F)
845 — Lectures presented at the Twelfth World Meteorological Congress (1995). 1997. (English)
866 — Scientific lectures presented at the forty-eighth session of the WMO Executive Council (1996). 1997. (English)
910 — Lectures presented at the forty-ninth session of the WMO Executive Council (1997). 2000 (English)
911 — Lectures presented at the fiftieth session of the WMO Executive Council (1998), 2000. (English)
916 — Forecasting in the 21st Century. 2000. (English)
WORLD METEOROLOGICAL ORGANIZATION
INTRODUCTION TO
CLIMATE CHANGE:
LECTURE NOTES FOR
METEOROLOGISTS
Prepared by
David D. Houghton
WMO-No. 926
Secretariat of the World Meteorological Organization
Geneva – Switzerland
2002
© 2002, World Meteorological Organization
ISBN No. 92-63-10926-5
NOTE
The designations employed and the presentation of material in this publication

do not imply the expression of any opinion whatsoever on the part of the
Secretariat of the World Meteorological Organization concerning the legal status
of any country, territory, city or area, or of its authorities, or concerning the
delimitation of its frontiers or boundaries.
TABLE OF CONTENTS
FOREWORD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
ACKNOWLEDGEMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
INTRODUCTION .............................................. 1
CHAPTER 1 UNDERSTANDING THE CLIMATE SYSTEM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

1.1 Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.1.1 Climate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.1.2 Climate system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.1.3 Climate change . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 General overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.3 Radiation processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.3.1 Introductory comments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.3.2 Radiative energy budget . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3.2.1 Solar radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3.2.2 Terrestrial radiation . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.3.2.3 The ‘greenhouse effect’ . . . . . . . . . . . . . . . . . . . . . . 7
1.3.2.4 Role of radiation in the overall energy balance . . . . 8
1.4 Characteristics of climate system components . . . . . . . . . . . . . . . . . 10
1.4.1 Introductory comments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.4.2 Atmosphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.4.3 Ocean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.4.4 Land surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.4.5 Cryosphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.4.6 Biosphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
1.5 Feedbacks in the climate system . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
1.5.1 Radiation energy transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
1.5.1.1 Temperature feedback . . . . . . . . . . . . . . . . . . . . . . . 21
1.5.1.2 Albedo feedback . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
1.5.2 Heat energy transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
1.5.3 Biosphere interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
1.6 Global nature of the climate system . . . . . . . . . . . . . . . . . . . . . . . . . 22
1.6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
1.6.2 Ozone hole in the stratosphere . . . . . . . . . . . . . . . . . . . . . . . 22
1.6.3 El Niño — Southern Oscillation (ENSO) . . . . . . . . . . . . . . . . 22
1.6.4 Monsoon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
1.6.5 Volcanoes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
1.7 Regional nature of the climate system . . . . . . . . . . . . . . . . . . . . . . . . 24
1.7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
1.7.2 Geography of climate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
1.7.3 Local variations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
1.7.3.1 Rainfall . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
1.7.3.2 Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
CHAPTER 2 NATURAL TEMPORAL VARIABILITY IN THE CLIMATE SYSTEM . . . . . . . . . . . . . . . . . . 27

2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.2 Basic forcing mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.2.1 External forcing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.2.1.1 Astronomical effects . . . . . . . . . . . . . . . . . . . . . . . . 27
2.2.1.1.1 Variations in solar radiation emission . . 27
2.2.1.1.2 Diurnal and annual cycles of solar
radiation input . . . . . . . . . . . . . . . . . .. 28
2.2.1.1.3 Variations in orbital parameters of
the Earth . . . . . . . . . . . . . . . . . . . . . . .. 29
2.2.1.1.4 Meteors . . . . . . . . . . . . . . . . . . . . . . . .. 29
iii
INTRODUCTION TO CLIMATE CHANGE
2.2.1.2
Geological effects . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.2.1.2.1 Tectonics . . . . . . . . . . . . . . . . . . . . . . . . 30
2.2.1.2.2 Volcanoes . . . . . . . . . . . . . . . . . . . . . . . 30
2.2.2. Interaction of climate system components . . . . . . . . . . . . . . 30
2.2.2.1 Ocean effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.2.2.2 Cryosphere effects . . . . . . . . . . . . . . . . . . . . . . . . . . 33
2.2.2.3 Biosphere interactions . . . . . . . . . . . . . . . . . . . . . . . 34
2.2.2.4 Internal atmospheric processes . . . . . . . . . . . . . . . . 35
2.3 Observed climate variability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.3.1 Surface temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.3.2 Precipitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.3.3 Severe weather . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
2.3.4 Ocean conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
CHAPTER 3 HUMAN IMPACTS ON THE CLIMATE SYSTEM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.2 Atmospheric greenhouse gas enhancement . . . . . . . . . . . . . . . . . . . . 41
3.2.1 Natural greenhouse gas constituents . . . . . . . . . . . . . . . . . . . 42
3.2.2 New greenhouse gases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.3 Atmospheric aerosol enhancement . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3.3.1 Types of aerosols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3.3.2 Radiative impacts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3.3.2.1 Direct impacts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
3.3.2.2 Indirect impacts . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
3.4 Change of radiative effects of clouds . . . . . . . . . . . . . . . . . . . . . . . . . 48
3.5 Change of radiative properties of the land surface . . . . . . . . . . . . . . . 49
3.6 Summary of human impacts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
CHAPTER 4 MODELLING CLIMATE CHANGE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
4.2 Basics for modelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
4.2.1 Governing physical equations . . . . . . . . . . . . . . . . . . . . . . . . 52
4.2.2 Parameterization of physical processes . . . . . . . . . . . . . . . . . . 54
4.2.3 Mathematics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
4.2.4 Computers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
4.3 Current climate models and their performance . . . . . . . . . . . . . . . . . 57
4.3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
4.3.2 Overall climate model evaluation . . . . . . . . . . . . . . . . . . . . . 58
4.3.2.1 Current climate conditions . . . . . . . . . . . . . . . . . . . 58
4.3.2.2 Past climate conditions . . . . . . . . . . . . . . . . . . . . . 61
4.3.3 Evaluation of climate model components . . . . . . . . . . . . . . . 62
4.3.3.1 Atmospheric component . . . . . . . . . . . . . . . . . . . . 62
4.3.3.2 Ocean component . . . . . . . . . . . . . . . . . . . . . . . . . 64
4.3.3.3 Land-surface component . . . . . . . . . . . . . . . . . . . . . 64
4.3.3.4 Cryosphere component (sea-ice models) . . . . . . . . . 65
4.3.4 Sensitivity of climate models to model formulation, boundary
conditions and parameterization . . . . . . . . . . . . . . . . . . . . . . 66
4.3.5 Update from 2001 IPCC Report . . . . . . . . . . . . . . . . . . . . . . . 68
CHAPTER 5 CLIMATE PREDICTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . ............... 69

5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . ............... 69
5.2 Predictability . . . . . . . . . . . . . . . . . . . . . . . . . . . . ............... 69
5.3 Short-term climate forecasts . . . . . . . . . . . . . . . . ............... 71
5.4 Medium-range climate forecasts . . . . . . . . . . . . . ............... 73
5.5 Long-range climate prediction . . . . . . . . . . . . . . . ............... 74
5.6 Predictability for regional climate . . . . . . . . . . . . ............... 75
5.6.1 Global climate models . . . . . . . . . . . . . . . ............... 75
5.6.2 Statistical downscaling technique . . . . . . ............... 75
5.6.3 Regional climate models . . . . . . . . . . . . . ............... 76
5.7 Update highlights on climate modelling from the 2001 IPCC Report 76
CHAPTER 6 OBSERVATIONS FOR LONG-TERM CLIMATE MONITORING . . . . . . . . . . . . . . . . . . . . 78

6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
6.2 Key principles for long-term climate monitoring . . . . . . . . . . . . . . . . 78
iv
TABLE OF CONTENTS
6.3 Status of selected observations critical for climate change . . . . . . . . . 79

6.3.1 Observations for basic forcing factors . . . . . . . . . . . . . . . . . . 79
6.3.1.1 Solar radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
6.3.1.2 Greenhouse gases . . . . . . . . . . . . . . . . . . . . . . . . . . 80
6.3.1.2.1 Carbon dioxide . . . . . . . . . . . . . . . . . . . 80
6.3.1.2.2 Ozone . . . . . . . . . . . . . . . . . . . . . . . . . . 80
6.3.1.2.3 Water vapour . . . . . . . . . . . . . . . . . . . . . 80
6.3.1.3 Aerosols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
6.3.2 Observations for feedbacks from climate system components 81
6.3.2.1 Clouds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
6.3.2.2 Oceans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
6.3.2.3 Surface hydrology . . . . . . . . . . . . . . . . . . . . . . . . . . 82
6.3.2.4 Surface land cover . . . . . . . . . . . . . . . . . . . . . . . . . . 82
6.3.2.5 Cryosphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
6.3.3 Observations for climate responses . . . . . . . . . . . . . . . . . . . . 83
6.3.3.1 Surface temperature . . . . . . . . . . . . . . . . . . . . . . . . 83
6.3.3.2 Precipitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
6.4 Strategies for improving long-term climate monitoring . . . . . . . . . . . 84
6.4.1 Data recovery and recalibration (‘rehabilitation’) . . . . . . . . . . 84
6.4.2 Reanalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
6.4.3 Increasing the number of measurements . . . . . . . . . . . . . . . . 86
6.4.4 New measurement systems . . . . . . . . . . . . . . . . . . . . . . . . . . 86
CHAPTER 7 MODELLING, DETECTION, AND ATTRIBUTION OF RECENT AND FUTURE CLIMATE CHANGE 87
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
7.2 Model results for climate change . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
7.2.1 Recent climate change . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
7.2.2 Future climate change . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
7.2.2.1 Mean conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
7.2.2.2 Variability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
7.2.2.3 Changes in extreme events . . . . . . . . . . . . . . . . . . . 95
7.2.2.3.1 Wind . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
7.2.2.3.2 Temperature . . . . . . . . . . . . . . . . . . . . . . 96
7.2.2.3.3 Precipitation . . . . . . . . . . . . . . . . . . . . . 97
7.2.3 Reducing model uncertainties and improving climate change
estimates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
7.3 Detection and attribution for causes of recent climate change . . . . . 98
7.3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
7.3.2 Recent progress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
CHAPTER 8 POTENTIAL IMPACTS OF CLIMATE CHANGE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103

8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
8.2 Terrestrial ecosystems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
8.2.1 Agriculture (plant crops) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
8.2.2 Forests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
8.2.3 Deserts, land degradation and desertification . . . . . . . . . . . . 105
8.3 Freshwater resources management . . . . . . . . . . . . . . . . . . . . . . . . . . 106
8.4 Sea-level rise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
8.5 Storms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
8.6 Human health . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
8.6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
8.6.2 Potential direct effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
8.6.3 Potential indirect effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
8.6.3.1 Vector-borne diseases . . . . . . . . . . . . . . . . . . . . . . . 112
8.6.3.2 Water-borne and food-borne diseases . . . . . . . . . . . 112
8.6.3.3 Agricultural productivity and food supplies . . . . . . 112
8.6.3.4 Air pollution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
8.6.4 Stratospheric ozone depletion and increased Earth-surface
ultraviolet radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
CONCLUDING REMARKS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114

REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
SUBJECT INDEX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
v
FOREWORD
The atmosphere is the essential physical and chemical environment for life.
Changes, anthropogenic or otherwise, to the physical and chemical properties
of the atmosphere have the potential of affecting directly the quality of life
and even the very existence of some forms of life.
Human-induced climate change, in particular, as well as other global
environmental issues such as land degradation, loss of biological diversity and
stratospheric ozone depletion, threatens our ability to meet very basic human needs,
such as adequate food, water and energy, safe shelter and a healthy environment.
A great majority of the scientific community, whilst recognizing that scientific
uncertainties exist, believe that human-induced climate change is inevitable.
Further proof of the reality of climate change was made available through the work
of the IPCC, established by WMO and UNEP in 1988. The recently released IPCC's
Third Assessment Report forms the most comprehensive picture of the state of the
climate and the global environment yet published, confirming that earlier
judgements and projections of global mean temperature increases were
underestimated. The IPCC also concludes that human activity is having a
discernible effect on the environment, and that global temperatures are projected
to increase at a rate unprecedented in the last thousand years.
A majority of experts believe that important reductions in net greenhouse
gas emissions are technically feasible due to a wide range of technologies and
policy measures in the energy supply, energy demand and agricultural and
forestry sectors. Besides, the anticipated adverse effects of climate change on
socio-economic and ecological systems can, to some degree, be reduced
through proactive adaptation measures. Consequently discussions are taking
place under the UNFCCC and the Kyoto Protocol seeking how to best cope
with this issue, particularly in developing mitigation and adaptation strategies
to prevent future generations from excessively negative impacts and to reduce
the world vulnerability to these changes.
The meteorological community and, in particular, the Weather Services are
now being increasingly solicited by the media, the general public and national and
private institutions for information and guidance on climate issues. It is then
essential that the technical and professional staff of weather services have the
necessary background and knowledge of basic concepts and approaches of the issue
of climate change in order to provide authoritative responses.
Professor David D. Houghton, from the Department of Atmospheric and
Oceanic Sciences, University of Wisconsin – Madison, USA, kindly agreed, as part of
his sabbatical leave, to prepare these lecture notes. They aim at enabling
meteorologists, hydrologists and oceanographers to gain a more comprehensive
understanding of the topics related to the issue of climate change.
I wish to express my sincere appreciation and gratitude to Professor Houghton
for the high quality and comprehensiveness of his text in such a difficult and
complex subject. I feel fully confident that these notes will be highly appreciated by
meteorologists and related professionals interested in expanding their knowledge
and understanding of the science of climate change.
(G.O.P. Obasi)
Secretary General
vii
ACKNOWLEDGEMENTS
The cooperation and support of many persons and institutions were essential for
producing this book. I am deeply indebted to the University of Wisconsin–
Madison and its Department of Atmospheric and Oceanic Sciences for a
sabbatical leave which gave me the time needed for the project and to the
Education and Training Programme (ETRP) at the World Meteorological
Organization (WMO) in Geneva for providing work facilities. I thank Dr Gustavo
Necco, Director of WMO’s Education and Training Department, for his support
and guidance and Eliette Tarry, also from the Education and Training
Department, for her handling of logistical arrangements.
I am very grateful for the comprehensive scientific reviews by Dr Hartmut
Grassl, Director of WMO’s World Climate Programme (WCP), and by Professor
Igor Karol, Voeikov Main Geophysical Laboratory, St. Petersburg, Russian
Federation. They provided essential calibration, constructive comments and addi-
tional information for the book.
I appreciated the warm welcomes and constructive conversations with many
persons in the climate research groups at the Hadley Centre, Bracknell, United
Kingdom, and the Max Planck Institute for Meteorology, Hamburg, Germany. In
particular, I wish to thank Geoff Jenkins and Lennart Bengtsson for hosting me on
these visits. Thanks are also due to Lev Karlin, who hosted my visit to the WMO
Training Center at the Russian State Hydrometeorological Institute in
St. Petersburg, and his assistant, Edward Podgaisky, and to Vilma Castro for
hosting my visit to the WMO Training Center at the University of Costa Rica, San
Jose, Costa Rica. These visits provided me with useful insights on the environ-
ments in which this book may be used as well as constructive comments on the
science of climate change.
Many other persons contributed to this project. They included: Maurice
Blackmon, Francis Bretherton, Jay Fein, Richard Hallgren, Ben Santer, Kevin
Trenberth and Warren Washington, for their comments in the formative stages of
the project; David Parker, Thomas Spence, Paul Try and William Rossow for infor-
mation they provided on meteorological observations; Ulrich Cubasch for his
insights on climate change education; Walter Fernandez for calling to my atten-
tion a recent book about the role of the sun in climate change; Steve Hammond
for information on computers; and Linda Hedges and Jean Phillips for helping
with clarifying references; and Pete Pokrandt for helping me to scan figures into elec-
tronic files. I also want to thank Heather McCullough for her work on the index.
Finally, I am very appreciative for the comprehensive editorial proofing of
the entire book by Linda Keller, my son Eric Houghton, and my wife Barbara
Houghton and all others who assisted in editing. Barbara provided support and
constructive review throughout the project that were critical for successful
completion of the project.
David D. Houghton
Professor, Department of Atmospheric and Ocean Sciences
University of Wisconsin—Madison
Madison, Wisconsin, United States of America
10 July 2001
ix
INTRODUCTION
Climate change and the need for environmental protection are global problems
and call for a knowledgeable response from all countries in order to be effectively
addressed. In recognition of this, the ten-year plan (1996–2005) of the Education
and Training Programme (ETRP) of the World Meteorological Organization
(WMO) places a high priority on enhancing a global system approach in its
education and training work in member countries, particularly in developing
countries. The plan states that personnel will need to have available an integrated
training programme for understanding the ocean-land-atmosphere system,
whether for monitoring purposes under the Global Atmosphere Watch (GAW),
for learning about physical concepts and understanding patterns, or for predic-
tion of tomorrow’s weather and long-term climate changes.
Considerable attention has been given to climate change by the scientific
community, government bodies and the public media. However, many issues are
not fully understood. It is important that the professional operational community
of meteorologists, hydrologists, and oceanographers become more knowledgeable
on this subject in order to be able to respond to the needs to monitor climate
change, to incorporate climate change perspectives into their own work, to help
governing bodies to understand the scientific issues, and to provide information
to the general public.
These lecture notes are intended to enhance familiarity with the broad scope
of topics related to climate change. They provide material on the science of climate
change assuming that the users already have a basic understanding of geophysical
fluid dynamics, and relevant physical processes such as radiation transfer, diffu-
sion, the hydrological cycle, and cloud physics along with some understanding of
air chemistry, hydrology, and oceanography. Individuals needing advanced level
material for climate change studies or research should refer to the basic references
listed at the beginning of the reference section on page 115.
The term ‘climate change’ is used with different meanings and perspectives.
In some cases it may refer to all environmental change or include natural varia-
bility. It is most useful to think of climate change as one of several symptoms of
human-produced environmental change with both global and local perspectives.
A global perspective is appropriate in recognition of the global interactions
involving the component physical systems fundamental to climate change. The
local perspective is essential because it is the local impacts which are of signifi-
cance to individuals and communities, and because it is at the local level where
measurements must be obtained from all parts of the world in order to properly
describe climate and predict its changes.
With the current vigorous research programs and advances in observation
technology and analysis it is clear that the specific assessments of climate change
will be continually evolving. It is hoped that these lecture notes will give a back-
ground which remains relevant to understanding the advancing science of
climate change.
There is a review of the characteristics and physical processes of the climate
system in Chapter 1 followed by a discussion of its variability first from natural
causes in Chapter 2 and then from human activity in Chapter 3. The description
of and advances made with numerical climate models are presented in Chapter 4
followed by a focus on climate predictability in Chapter 5. Chapter 6 presents
important requirements for observations needed to identify and understand
climate change. Progress in the isolation and analysis of recent climate change is
discussed in Chapter 7 followed by examples of climate change impacts in
Chapter 8.
1
The primary seven references for these lecture notes are listed at the top of
the list of references on page 115. They are labeled with numbers running from
one to seven, and are cited in the text, particularly in tables and figures. Note that
the book on long-term climate monitoring, the three reports of the
Intergovernmental Panel on Climate Change (IPCC), and the climate system
modelling book, numbered 2, 3, 4, 6 and 7, respectively, contain material from a
very large number of contributors. A few other references are cited in the text.
Figures and tables adapted from the source material contain numerous reference
citations. A complete list of references is provided at the end of the text.
The third report of Working Group I of the IPCC which updates the scien-
tific analysis and conclusions for climate change appeared just at this book was
being finalized (July 2001). That report (cited as Reference no. 7 here) basically
confirmed and strengthened the scientific conclusions from the previous reports
used in this book. Some material was incorporated into this book; however, it was
not considered necessary to include all of the details.
2
CHAPTER 1
UNDERSTANDING THE CLIMATE SYSTEM
1.1 Climate is generally defined as the average state of the atmosphere for a given
DEFINITIONS time scale (hour, day, month, season, year, decade and so forth) and generally for
1.1.1 a specified geographical region. The average-state statistics for a given time scale
CLIMATE including all deviations from the mean are obtained from the ensemble of condi-
tions recorded for many occurrences for the specified period of time. Thus the
mean temperature for the month of May in Ankara, Turkey, is obtained from
measurements considered representative for Ankara averaged over the month of
May from a record of many years. Climate descriptors also include conditions at
the Earth’s surface such as ocean temperatures and snow cover.
The average-state description involves a wide range of variables depending
on what is of interest. Temperature and precipitation are the most commonly
used; however the list may include wind, cloudiness and sunshine, pressure, visi-
bility, humidity and elements with noteworthy human impacts such as severe
storms, excessively high and low temperatures, fog, snow and hail. The method
of description focuses on statistical parameters, the mean and measures of vari-
ability in time such as the range, standard deviation, and autocorrelations.
It is important to identify the difference between weather and climate.
Weather involves the description of the atmospheric condition at a single instant
of time for a single occurrence. In general, climate may be thought of as an
average of weather conditions over a period of time including the probability for
distributions from this average.
1.1.2 The climate system is defined as the five components in the geophysical system,
CLIMATE SYSTEM the atmosphere and four others which directly interact with the atmosphere and
which jointly determine the climate of the atmosphere. The five components are
listed below:
(a) Atmosphere;
(b) Ocean;
(c) Land surface;
(d) Ice and snow surfaces (both land and ocean areas); and,
(e) Biosphere (both terrestrial and marine).
Figure 1.1 shows the scope of the climate system. Note that the two-way
arrows in the diagram identify explicit interactions between the atmosphere and
other components.
At this point it is appropriate to recognize that there are other factors, also
variable in nature, which contribute to determining the climate. These are consid-
ered ‘external’ forcing factors and include the sun, Earth orbital parameters,
land-ocean distribution, Earth topography (land and ocean), and basic composi-
tion of the atmosphere and ocean. These are important determiners of the climate
which, except for the basic composition of the atmosphere and oceans, are not
affected in return by the climate conditions.
1.1.3 Climate change in these lecture notes is defined as the change in climate attrib-
CLIMATE CHANGE uted directly or indirectly to human activity which, in addition to natural climate
variability, is observed over comparable time periods. The definition adopted by
the United Nations Framework Convention on Climate Change (UNFCCC)
focuses only on the human activity that alters the composition of the global
atmosphere and excludes other human activity effects such as changes in the land
surface. Sometimes the term ‘climate change’ is used to include all climate vari-
ability, which can lead to considerable confusion. Climate has variability on all
time and space scales and will always be changing.
3
Changes in the Atmosphere: Changes in the

Composition, Circulation Hydrological Cycle
Changes in
Solar input
. Atmosphere Clouds
Aerosols Air-Biomass
H2O, N2, O2, CO2, O3, etc.
. Coupling
Air-Ice Precipitation-
Coupling Evaporation
Terrestrial
Radiation
Heat Wind
exchange Stress
Human influences
Biomass
Sea-ice Landmass
Coupling
Ocean
Ocean Rivers
Lakes
Ice-Ocean Coupling
text
Land
Changes in the Ocean: Changes in/on the Land Surface:

Circulation, Biogeochemistry Orography, Land Use, Vegetation, Ecosystems
Figure 1.1 — Schematic view of the components of the global climate system (bold), their processes and interactions (thin
arrows) and some aspects that may change (bold arrows). [from page 55, Reference no. 3].
1.2 The definition for the climate system makes it clear that one has to have an
GENERAL OVERVIEW understanding of all of that system’s components (atmosphere, ocean, land
surface processes, cryosphere, and biosphere) in order to understand it. In reality
one needs to know a limited amount, dependent on the time scales considered,
about the non-atmospheric components to understand the interactions of those
components with the atmosphere. In general, these interactions occur primarily
at physical interfaces so that, for example, for ocean interactions, it is necessary
to know only the conditions at the oceanic upper boundary and for cryosphere
interactions only at the surface of the ice. To know such conditions, of course, it
is necessary to understand how they vary in relationship to conditions within the
ocean and ice. Unlike the other interactive components, the ocean is an easily-
movable fluid, as is the atmosphere, so that understanding the ocean for climate
system applications requires dealing with geophysical fluid dynamic and ther-
modynamic relationships as complex as those for the atmosphere. Hence, it
becomes necessary to use numerical model representation for the ocean compa-
rable to that used for the atmosphere. Current climate system research depends
strongly on coupled atmosphere-ocean numerical models.
This chapter focuses on the forcing and interaction processes of particular
relevance to climate change. Basic material on topics such as large-scale geophys-
ical fluid dynamics, synoptic-scale weather systems, turbulence, or the
hydrological cycle is not covered here. Radiation processes play a key role in the
climate change scenario and are discussed first in some detail followed by discus-
sion of relevant aspects of the five climate system components. Examples
demonstrating the global connections of the climate change processes are then
presented, and finally, there is discussion of regional-scale aspects of climate vari-
ability and climate change.
4
CHAPTER 1 — UNDERSTANDING THE CLIMATE SYSTEM
1.3 Electromagnetic wave energy transfer (radiation) accounts for nearly all energy
RADIATION PROCESSES transfer from the sun, and is the primary source of energy for the atmosphere and
1.3.1 the entire climate system. Such transfer is also the only way in which significant
INTRODUCTORY COMMENTS amounts of energy can leave the climate system. The energy of the global climate
system is nearly in balance with incoming and outgoing radiation transfers. A
change in one component will produce a different balanced state. The primary
human impact on the energy balance is to alter the radiative properties of the
atmosphere with respect to these two energy streams. This effect far out shadows
other anthropogenic energy sources and sink effects such as the heating due to
combustion and nuclear processes. Understanding the impacts of human envi-
ronmental change on radiation transfer processes in the atmosphere and on the
Earth’s surface is crucial to understanding climate change. Because of this central
role of radiation in climate change, a brief review of relevant radiation principles
is given here even though it is assumed that the student already has a basic under-
standing of radiation.
Radiation principles cover the production (emission) of radiation energy
from the internal energy (heat) of material substance and the transformation of
radiation into the internal energy of material substance (absorption). These radi-
ation principles also govern a number of processes that change the nature of the
radiation, but do not convert its energy to internal energy of matter: reflection,
refraction, diffraction and scattering. The production (emission) of radiation
energy depends upon internal energy (temperature) as well as other properties of
the emitting material substance. The destruction (absorption) of radiation
depends on the amount of incident radiation energy and properties of the absorb-
ing material substance except for its temperature. The properties of radiation
depend on its wavelength. Radiation can exist for a wide and continuous range of
wavelengths referred to as the radiation spectrum.
There are two primary forms of radiation relevant to the energy balance
properties of the climate system. The first is the ‘solar’ or ‘short-wave’ form
predominant in the radiation from the sun. This is primarily in the wavelength
range from 0.2 to 4.0 microns (a micron is one-millionth of a meter) which
encompasses the visible part of the spectrum. This short-wave radiation provides
a source of energy for the climate system as it is absorbed in the atmosphere,
clouds, ocean, land surface, and by living matter. The second form is the ‘terres-
trial’ or ‘long-wave’ type predominant in the radiation emitted by matter in the
climate system. The primary wavelength range for this form is from 4 to 60
microns which is entirely in the invisible infrared part of the spectrum.
Sometimes the solar and terrestrial radiation forms are called ‘visible’ and ‘invisi-
ble,’ respectively.
The overall differences in predominant wavelengths are due to the
differences in temperature of the emission sources for the radiation, about 6000 K
for the sun and in the range of 190-330 K for the climate system components
emitting terrestrial radiation. The terms ‘short-wave’ and ‘long-wave’ refer to the
wavelengths of these radiation forms relative to each other and should not be
confused with the same terminology used to describe the wavelengths of
radiation used for communications (radio and television).
The relative roles of the two primary forms of radiation in the energy balance
are complicated by the fact that the components of the climate system absorb as,
well as emit, long-wave infrared radiation. This leads to a very complex descrip-
tion of the long-wave radiation energy transfer processes (Schwarzchild’s
equation) which is complicated to solve for the atmosphere situation.
Large-scale radiation effect variations in the climate system are most
pronounced with respect to height and latitude. Horizontal variations in radiation
transfer do exist at the small scale as demonstrated by the difference in solar
heating at the Earth’s surface on the two sides of a hill, one facing the sun and the
other facing away from the sun. For general applications to the climate system,
only the vertical component of radiation and is considered. Energy transfer
magnitudes are discussed with reference to energy crossing horizontal surfaces.
5
1.3.2 Solar radiation incident upon the Earth system coming into the atmosphere from
RADIATIVE ENERGY BUDGET above leads to heating as it is absorbed by gases, aerosols and clouds in the atmos-
1.3.2.1 phere, and by the ocean, land, ice and biosphere elements at the Earth’s surface.
Solar radiation The absorption is proportional to the intensity of the incident solar radiation and
depends on the properties of these substances. As discussed above, the relevant
intensity is the component in the vertical direction. As the solar radiation is
absorbed, there is less radiation available to be absorbed at lower levels. Although
the radiation is initially in a narrow beam traveling in one direction from the sun,
reflection at surfaces and scattering within the atmosphere sends the solar radia-
tion all directions. For this reason, when one looks outdoors in the daytime one
sees light coming from all directions. The complete description of the absorption
effects (heating) must include the cumulative effects from radiation propagating
in from all directions.
The overall heating effects due to solar radiation absorption on a horizontal
surface in the climate system relate to the intensity of the solar beam coming into
the atmosphere from space and the angle of the solar beam to the local vertical.
The intensity depends on the temperature of the sun and the distance from the
sun to the Earth. The angle of the solar beam to the local vertical varies according
to a number of astronomical factors: latitude on Earth (distance from the
equator), longitude on Earth (time of day), and the orientation of the Earth’s axis
with respect to the sun (the solar declination angle) which varies according to the
time of year. Variations in these factors lead to large variations in heating from
day to night, from equatorial to polar regions, and from summer to winter. Figure
1.2 shows the variations in total daily solar radiation energy received at the top of
the atmosphere as a function of latitude and time of year. Variations in the Earth’s
orbit and solar conditions result in additional longer-term variations to be
discussed later.
The details of solar radiation absorption (heating) within the atmosphere
and at the Earth’s surface depend strongly on the properties of the absorbing
substance. The albedo (reflectivity) of sunlight from the Earth’s surface is
indicative of (inversely related to) the absorption of radiation by that surface. A
surface with a high albedo (high visible brightness) is heated much less than one
with a low albedo (low visible brightness). At the Earth’s surface, the albedo
ranges from about five per cent for ocean surfaces (with the sun high in the sky)
and the top surface of dark thick coniferous forests to 90 per cent for fresh snow.
Thick clouds in the atmosphere can also have an albedo nearly as high as fresh
snow. Since much of the reflected and back-scattered solar radiation travels back
out to space, it is never converted to heat in the climate system. The atmosphere
(gases, aerosols, and clouds) absorbs less of the incident radiation than the Earth
surface, so that solar heating effects are greater at the Earth surface than in the
atmosphere.
Figure 1.2 — Daily total of the

solar radiation incident on a unit
horizontal surface at the top of
the atmosphere as a function of
latitude and date in 106 J m-2 for
one day ( 11.6 W m-2). Shaded
areas represent the areas that are
not illuminated by the sun (from
Wallace and Hobbs, 1977; after
List, 1951). [from page 100,
Reference no. 5, with
permission of Springer-Verlag].
6
Terrestrial (long-wave) radiation is both emitted and absorbed by material

substances in the climate system. The absorption depends on the incident radia-
tion intensity and physical properties of the substances (except for temperature)
whereas the emission depends on the temperature and other physical properties
of the substances. The Earth’s surface and clouds have radiative properties that tend
to produce the maximum amount of terrestrial radiation given by ‘black body’ values
and to absorb incident terrestrial radiation completely. On the other hand, the radi-
ation emission and absorption characteristics of atmospheric gases have a large
variability depending on wavelength, as shown in Figure 1.3. The strongest effects
are exhibited by minor constituents in the atmosphere: water vapour, carbon
dioxide, ozone, nitrous oxide, and methane. These gases occur naturally and are
known as ‘greenhouse gases.’
1.3.2.3 The ‘greenhouse gas’ radiative properties just noted are much more pronounced
The ‘greenhouse effect’ for terrestrial radiation than for solar radiation. Panel b in Figure 1.3 shows the
magnitude of absorptivity for the entire depth of the atmosphere. Note that the
absorptivity effects are much larger for terrestrial radiation (wavelength range 4 to
60 microns) than for the solar radiation (wavelength range 0.15 to 4 microns). The
large absorptivity of the atmospheric gases for the terrestrial radiation together
with atmospheric temperatures in the 210-310K range [optimal for terrestrial radi-
ation emission as shown in Figure 1.3, Panel (a)] leads to emission of significant
Figure 1.3 — (a) Black body

curves for the solar radiation
(assumed to have a temperature
of 6000 K) and the terrestrial
radiation (assumed to have a
temperature of 255 K);
(b) low resolution absorption
spectra for the entire vertical
extent of the atmosphere; and
(c) for the portion of the
atmosphere above 11 km after
Goody (1964); and,
(d) the absorption spectrum for
the various atmospheric gases
between the top of the
atmosphere and the earth’s
surface
after Howard et al. [updated with
data from Fels and Schwarzkopf
(1988, personal communication)
between 10 and 100 µm].
[adapted from page 93,
Reference no. 3].
7
amounts of terrestrial radiation in all directions even where clouds are not present.
[Students should recall the Kirchhoff and the Wien displacement radiation laws.]
It is the downward component which retains energy in the climate system and
keeps equilibrium temperatures at the Earth’s surface and in the lower atmosphere
higher than would otherwise be the case. This enhanced temperature is said to
result from the ‘greenhouse effect.’
On a globally-averaged basis the observed surface temperature is about 33K
above the 255K expected with no atmosphere at all. This enhancement value
would be even greater (over 80K) if radiative effects for the clear atmosphere were
the only modifying factors [Manabe and Strickler, 1964]. Sensible and latent heat
fluxes from the Earth’s surface to the atmosphere along with atmospheric convec-
tion partially offset the surface temperature enhancement due to radiation
transfer back from the atmosphere.
1.3.2.4 When considering only the vertical transfers of energy, radiation energy transfers
Role of radiation in the overall have a dominant role in the overall energy balance of the globally-averaged
energy balance atmosphere and the Earth’s surface. Figure 1.4 summarizes these energy
transfers. The top of the diagram represents the top of the atmosphere, the
bottom the Earth’s surface, and the atmosphere is in the middle. The solar
radiation components are on the left side. The terrestrial radiation components
are on the right side, and the sensible and latent heat transfers are shown in the
center.
Note that about 30 per cent of the incoming solar radiation is returned to
space without being converted to heat (an albedo of about 30 per cent for the
Earth–atmosphere system); about half is absorbed at the Earth surface, and only
about 20 per cent is absorbed in the atmosphere. For the terrestrial radiation
Reflected Solar Incoming 235 Outgoing

107 Radiation Solar Longwave
342
107 W m-2 Radiation Radiation
342 W m-2 235 W m-2
Reflected by Clouds,
Aerosol and 77
40
Atmosphere Emitted by Atmospheric
77 Atmosphere 165 30 Window
Absorbed by Greenhouse
67 Atmosphere Gases
Latent
24 78 Heat
40 324
Reflected by 350 Back
Surface Radiation
30
390
168 24 78 Surface
Absorbed by Surface Thermals Evapo- Radiation 324
transpiration Absorbed by Surface
Figure 1.4 — The Earth’s radiation and energy balance. The net average incoming solar radiation of 342 Wm-2 is partially
reflected by clouds and the atmosphere, or at the surface, but 49 per cent is absorbed by the surface. Some of that heat is
returned to the atmosphere as sensible heating and most as evapotranspiration that is realized as latent heat in precipitation.
The rest is radiated as thermal infrared radiation and most of that is absorbed by the atmosphere which in turn emits
radiation both up and down, producing a greenhouse effect, as most of the thermal radiation lost to space comes from cloud
tops and parts of the atmosphere much colder than the surface. The partitioning of the annual global mean energy budget
and the accuracy of the values are given in Kiehl and Trenberth (1997). [from page 58, Reference no. 3].
8
emitted from the Earth only about 10 per cent is transmitted directly to space; the
remaining part is absorbed in the atmosphere.
The energy component labeled ‘back radiation’ is a key indicator of the
greenhouse effect. Note also that the magnitude of terrestrial radiation emitted
downward from the atmosphere to Earth and absorbed at the Earth’s surface is
nearly equal to the total solar radiation incident at the top of the atmosphere and
is about double the amount of solar radiation absorbed at the Earth’s surface. In
general, the magnitudes of radiation energy transfer are considerably larger than
those associated with the sensible and latent heat transfers (see Figure 1.4).
A basic aspect of radiation forcing is the systematic variation with latitude.
As had been shown in Figure 1.2, there is generally an overall reduction with
distance from the equator in the daily total solar radiation coming into the
Earth–atmosphere system, being more extreme in the winter season and nearly
absent at the time of the summer solstice. Seasonal and annual means for solar
radiation absorbed in the Earth-atmosphere system show poleward decreases in
both the summer and winter hemispheres (see Figure 1.5).
The net radiative forcing, of course, depends on both the input from solar
radiation and losses due to terrestrial radiation emission to space. Latitude varia-
tions of terrestrial radiation emission are much less than for solar radiation (Figure
1.5). This terrestrial emission depends on the temperature (on the absolute Kelvin
scale) both at the Earth’s surface, and in the atmosphere which has a smaller
percentage variation than that for the zenith angle factor change with latitude
which affects solar radiation absorption amounts.
The resulting net radiation forcing for the Earth–atmosphere system (see the
Figure 1.5 — Meridional profiles last panel in Figure 1.5) has a net excess in the tropical latitudes and a deficit in
at the top of the atmosphere, for the polar latitudes. If radiation transfer were the only process occuring, the equa-
annual, DJF and JJA mean torial regions would be hotter than observed and the polar regions colder than
conditions, of: observed. However, the transport of heat from the equatorial to polar regions by
(a) the zonal-mean albedo; atmospheric and oceanic circulations offsets this radiation imbalance and
(b) absorbed solar radiation; provides an overall energy balance at each latitude.
(c) emitted terrestrial radiation; In conclusion and as stated before, the primary connection between human
and (d) net radiation; activity and climate change is the alteration of the radiation transfer characteris-
(based on data from Campbell tics of the atmosphere. The change in greenhouse gas concentration and the
and Vonder Haar, 1980). No addition of other gases with similar characteristics will change the terrestrial radi-
corrections were made for global ation transfer. In addition, a change in aerosol concentration and perhaps related
radiation balance. [from page change in cloud cover will change the solar radiation transfer. Except on a very
128, Reference no. 5. With local scale, the energy transferred by radiation is far greater than any production
permission of Springer-Verlag]. rate of energy due to human activity.
80 400
ABSORBED (Wm-2)
ALBEDO (%) a SOLAR
b
(From Campbell and Vonder Haar, 1980)
60 300
DJF
JJA YEAR
40 200
20 100
80S 60 40 20 0 20 40 60 80N
150 0
80S 60 40 20 0 20 40 60 80N
100
-2
NET (Wm ) d
50
0 300
EMTTED (Wm-2) c
-50 250 TERRESTIAL
-100 200
-150 150
-200 100
80S 60 40 20 0 20 40 60 80N 80S 60 40 20 0 20 40 60 80N
9
1.4 A brief description of each of the five components of the climate system is
CHARACTERISTICS OF presented. The information on the atmosphere is more extensive as climate is
CLIMATE SYSTEM largely defined by conditions in the atmosphere. It is expected that students will
COMPONENTS already have background knowledge on the atmosphere, its circulations and phys-
1.4.1 ical processes. It is understood that students may have little or no background for
INTRODUCTORY COMMENTS the other four components. Therefore the material presented here deals only with
aspects relevant to interactions with the atmosphere.
1.4.2 A number of factors make the atmosphere a very complex fluid system. As a gas,
ATMOSPHERE it has compressibility and great mobility. It extends well above topographic barri-
ers and can sustain global-scale circulations. A number of forcing factors
including radiative heating and cooling, latent heat sources and sinks due to
phase change of water, and variations in Earth’s surface temperature give rise to
significant temperature variations in all three space dimensions and in time.
These temperature variations give rise to horizontal pressure gradient forces,
approximately consistent with the hydrostatic relationship, which are the basis
for horizontal atmospheric motions. Temperature variations also affect vertical
pressure gradient forces which, in instances of ‘static instability,’ can lead to large
local-scale vertical motions. The large-scale horizontal motions are greatly modi-
fied by Coriolis effects arising from the rotation of the Earth and attain sufficient
magnitudes to force smaller-scale transient motions. Fluid transport and nonlin-
ear processes together with interactions among the physical factors listed above
lead to a complex global general circulation system and embedded smaller-scale
systems with space and time scales all the way down to atmospheric turbulence
which we see in the gustiness of winds and the dispersion of smoke plumes.
The hydrological cycle is an important part of the atmospheric system.
Evaporation and condensation of water can transfer considerable amounts of
energy by both vertically and horizontally. The cloud component of the
hydrological cycle strongly affects transfers of both solar and terrestrial radiation.
The precipitation is the source of fresh water needed for life on land surfaces.
In the climate perspective, where statistics of atmospheric conditions are
considered, the general circulation and associated temperature, cloudiness and
precipitation patterns provide the primary bases for the mean climate conditions.
Some of the transient features have systematic variations in time which are asso-
ciated with the diurnal and annual cycles which are described directly in climate
descriptions. Examples include average high and low temperatures for the day,
monthly mean temperatures for each month of the year, and the annual range in
monthly mean temperature. The random transient features such as extratropical
cyclones, moist convection in the tropics and middle latitudes, and turbulence
contribute to the climate descriptions for extremes and also for the mean states if
appropriate correlations exist among the variables of the transient systems.
This last point deserves to be developed and illustrated. Let us, for example,
take the relatively random transient atmospheric feature, the cumulus cloud. If
the vertical circulations associated with cumulus convection have upward
motions with systematically higher temperature than the downward motions, it
would be expected that an ensemble of these weather systems would give a net
upward transport of sensible heat. In the same way, if in association with extrat-
ropical cyclones, the winds from the south were typically warmer than winds
from the north, an ensemble of these storm systems would give a net northward
transport of sensible heat. In many situations randomly-occurring transient
systems contribute significantly to the larger-scale conditions which determine
the longer-term climate mean states.
It is assumed that the student is aware of the general climate conditions over
the Earth. Many texts and atlases present climate maps, including the books refer-
enced in the introduction. Nevertheless, a few climate charts are shown here to
illustrate differences over the world as well as to present seasonal differences. It is
important to remember that climate conditions vary considerably over the world,
climate change would therefore affect people in different places of the world quite
differently. Warming of summers in Canada may be welcomed as an enhancement
10
of the growing season whereas warming of summers in the Sahara, where it is

already hot, may be an entirely negative development.
Figures are shown for the climatology of surface temperature in January and July
(Figure 1.6), precipitation in December-February and June-August (Figure 1.7), surface
airflow (and pressure field) in December-February and June-August (Figure 1.8), and
upper tropospheric airflow in December-February and June-August (Figure 1.9). The
surface temperature and precipitation are key parameters for surface living conditions.
The wind fields highlight atmospheric transport conditions at the surface where we live
and in the upper troposphere where the primary maxima of kinetic energy exist in the
circulation (the subtropical and polar jet streams). Both topography and ocean–land
temperature differences result in wavy patterns in the horizontal mean flow particularly
in the Northern Hemisphere.
The circulation in the atmosphere is sufficiently vigorous that material
injected into one part of the atmosphere can be spread quickly over broad regions.
It may take just days for a volcanic smoke plume or radioactive products to circle
the Earth. Constituents with sufficiently long lifetimes, such as carbon dioxide,
would be expected to have relatively uniform concentration throughout the
atmosphere. The large-scale north-south circulations are less vigorous than those
in the the east-west direction (see Figure 1.9). Vertical motions generally are much
smaller than horizontal motions so the vertical transport of atmospheric
constituents may be quite limited. This accentuates the build-up of atmospheric
pollutants in the lower layers of the atmosphere especially in local areas with large
sources of the pollutants.
The internal instabilities, feedbacks and nonlinear nature of the atmospheric
system can result in circulation features that appear quite unrelated to the basic
forcing due to energy fluxes at the Earth’s surface and to radiation energy transfer.
Examples of these include tropical cyclones and extratropical weather fronts.
Furthermore, circulations may have more than one equilibrium for a given exter-
nal forcing. Chaos theory deals with the variability characteristics in systems with
such multiple equilibria. This characteristic of the atmosphere adds additional
challenges and uncertainty to understanding and determining the outcomes for
climate change scenarios.
Figure 1.6 — Global map of (a)

the January surface temperature
January and (b) July surface
temperature. (From Shea
(1986), reproduced with
permission from the National
Center for Atmospheric
Research). [from page 7,
permission of Academic Press].
11
Figure 1.7 — Global distribution

of average precipitation rate for
December-January-February
(D,J,F; left panel) and June-July-
August (J,J,A; right panel) for
1988-1996
(from the Global Precipitation
Climatology Project [GPCP] of
the Global Energy and Water
Cycle Experiment [GEWEX].
Figure 1.8 — Global

distributions of the height
anomalies of the 1000-mb
pressure field in gpm from the
standard atmosphere height, 113
gpm, and vector plots of the
surface winds for northern winter
(left panel) and northern summer
(right panel) mean conditions.
Each barb on the tail of an arrow
represents a wind speed of 2 m s-1.
The isoheight lines can be
interpreted as isobars for surface
pressure reduced to sea level.
Figure 1.9 — Global distributions

of the height difference of the
200-mb pressure field in gpm
from 11,784 gpm and vector
plots of the 200-mb winds for
northern winter (left panel) and
northern summer (right panel)
mean conditions. Each barb on
the tail of an arrow represents a
wind speed of 5 m s-1.
[adapted from pages 151 and
152, Reference no. 5, with
12
13
1.4.3 The ocean has a major impact on the climate of the atmosphere. It covers approx-
OCEAN imately 71 per cent of the Earth’s surface and thus has a dominant role for
transfers of energy and other properties between the atmosphere and the Earth’s
surface. Its large heat capacity, made accessible for surface energy transfers by
circulations within the ocean, provides a moderating effect on temperature vari-
ability in the atmosphere. Oceanic currents transfer large amounts of heat energy
away from equatorial regions. Finally, the ocean is an important source for atmos-
pheric water vapour, as well as a source and sink for other greenhouse gases.
The ocean’s heat capacity exceeds that of the atmosphere by a factor of the
order of 1000. This is due to differences both in heat capacity per unit mass (the
specific heat of liquid water is about four times that of air), and in total mass
between the ocean and atmosphere. The ocean’s heat capacity bears upon atmos-
pheric temperature through oceanic transports (both horizontal and vertical) that
produce and maintain surface water temperatures warmer or colder than the
atmosphere resulting in large heat transfers. The depth to which the oceans inter-
act with the atmosphere depends on the time scale under consideration. For
diurnal variations the depth is small, of the order of five to 10 meters. For seasonal
variations the depth is 20-200 meters (the depth of the well-mixed oceanic surface
layer). The ocean is a major component in determining the climate and its varia-
tions for annual, annually-averaged and longer period conditions.
The strong influence of the ocean on surface air temperature is clearly
evident. Figure 1.10 shows climatological sea-surface temperature conditions for
January and July. Note that these temperatures are similar to those in the surface
air over the oceans shown in Figure 1.6. Ocean temperatures are maximum in
equatorial regions and decrease poleward. However, the poleward decrease in the
Figure 1.10 — Monthly mean

Sea-Surface Temperatures (SSTs)
for January and July. Dark areas
indicate sea ice; stippling
indicates land areas. The contour
interval is 2°C except for dashed
contours of 27 and 29°C. (from
Shea et al., 1990).
14
winter hemisphere is not as rapid as that experienced for land surface

temperatures. The large ocean heat capacity results in summer-to-winter
temperature differences over the ocean are generally much less than the summer-
to-winter differences over land areas at comparable latitudes. Adjacent land areas
downwind of the ocean, in fact, experience a moderation of winter cold
temperatures and summer hot temperatures compared to other land areas at the
same latitude. Note, for example, that in January at 50°N the mean surface air
temperatures are well below freezing in eastern north America, well above
freezing in the central Atlantic (as is the sea-surface temperature), and still above
freezing in western Europe (see Figure 1.6).
Ocean currents have a major influence on sea-surface temperature. It is impor-
tant to understand these currents to fully appreciate the ocean’s interactions in the
climate system. The laws of basic geophysical fluid dynamics apply as they do for
atmospheric motions. As for the atmosphere, spatial variations in heating, primarily
in the ocean surface layer, lead to horizontal pressure gradients which cause motion.
However, the oceanic condition is different from the atmosphere in two
fundamental ways. First, the primary forcing of the ocean is at the upper boundary,
whereas the primary forcing for the atmosphere is at its lower boundary.
Atmospheric winds above the ocean are a major factor in causing ocean surface
currents through surface friction processes. In contrast, for the atmosphere fric-
tional conditions at its lower boundary tend to reduce atmospheric motion.
Second, the density of the ocean water is determined primarily by its salinity and
temperature instead of pressure, temperature and water vapour content as in the
atmosphere. The water vapour factor is relatively unimportant for atmospheric
density except in hot and humid conditions. On the other hand, salinity can play
a major role for ocean density especially when temperatures are near freezing in
which case density changes very little with temperature. Salinity conditions in
polar ocean regions are important for determining whether or not significant
vertical motions occur in local areas. The variations in ocean water density
according to temperature and salinity are shown in Figure 1.11. The density value
is shown as the difference from 1000 kg m-3. Thus, for example, for a temperature
of 10°C and a salinity of 16 parts per thousand, the seawater density is approxi-
mately 1016 kg m-3.
The major surface-layer ocean currents are shown in Figure 1.12. Ocean
currents result in significant equator-to-pole transport of thermal energy. The
strong northward-moving currents off the east coasts of Asia and north America
(the Kuroshio and Gulf Stream currents) take warm water away from the tropics.
Currents towards the equator on the west side of continents (the California, Peru,
and Benguela currents) transport cold water towards the equator.
Currents at deep ocean layers form primarily due to pressure gradients from
density variations (thermohaline currents). The density variations are strongly
influenced by fresh water sources from land surface runoff and sea-ice melt and
Figure 1.11 — Contours of

seawater density anomalies
(difference from a reference
density of 1000 kg m-3) in
kg m-3 plotted against salinity
and temperature. [from page
175, Reference no. 1 with
15
Figure 1.12 — A map of the

major surface currents in the
world ocean during the northern
winter (from Tolmazin, 1985).
[from page 177, Reference no. 5,
with permission of Springer-
Verlag].
fresh water evaporation from the sea surface. These currents provide a coupling
between deep and surface ocean waters that involves a large portion of the ocean
and provides ocean impacts on climate system variability on time scales of
centuries, millennia and even longer.
The oceans play a significant role as a source and sink for atmospheric gases,
including greenhouse gases. Changes in ocean temperature can change the
holding capacity for gases and can result in a net outflow or intake from the
atmosphere. Particularly noteworthy is the case for carbon dioxide. It is estimated
Figure 1.13 — The global carbon that the carbon dioxide dissolved in the upper layers of the ocean is nearly 50 per
cycle, showing the reservoirs (in cent more than the total amount in the atmosphere (1020 versus 750 gigatons of
GtC) and fluxes (GtC/yr) relevant carbon content). See Figure 1.13. Thus, there is much potential for effects on the
to the anthropogenic perturbation atmospheric carbon dioxide concentrations and the resulting radiation impacts
as annual averages over the due to changes in the ocean.
period 1980 to 1989 (Eswaran et In conclusion, the ocean is a very important and interactive component in
al., 1993; Potter et al., 1993, the climate system. The atmosphere forces oceanic motions through surface
Siegenthaler and Sarmiento, friction and affects oceanic temperature through surface sensible, latent, and
1993). The component cycles are
simplified and subject to 60 Atmosphere
considerable uncertainty. In n 750
tio
duc
addition, this figure represents ro
yp n 1.6 e
average values. The riverine flux, ar tio
ir m pira us 90 5.5
p
t e s d-
particularly the anthropogenic ne nd r lan
l
a a g Fossil fuels and
portion, is currently very poorly ob .3 gin
Gl 61 an cement production
Ch
quantified and so is not shown Vegetation 610
Soils and detritus 1580 0.5
here. Evidence is accumulating 2190 92
that many of the key fluxes can
fluctuate significantly from year Surface ocean
to year (terrestrial sinks and 1020
sources: INPE, 1992; Ciais et al., 50 10
0
40
1995; export from the marine
Marine biota
biota; Wong et al., 1993). In 3
contrast to the static view
conveyed by figures such as this 91
.6
one, the carbon system is clearly 4
6
dynamic and coupled to the Dissolved Intermediate and

6
climate system on seasonal, organic carbon deep ocean
<700 38,100
interannual and decadal time-
0.2
scales (e.g. Schimel and Sulzman,
1995). [from page 77, Surface sediment
150
Reference no. 3].
16
radiative energy transfers at the ocean–atmosphere interface. The ocean affects

atmospheric temperature by virtue of its large heat capacity which is enhanced
by circulations that distribute its heat energy internally. As will be seen
elsewhere in this chapter, it is also is a source and sink for atmospheric water
vapour and other greenhouse gases. The ocean also has an important biosphere
component.
1.4.4 Land surface is an important interactive component of the climate system. It

LAND SURFACE covers 29 per cent of the Earth’s surface. Significant exchanges of heat, moisture,
and momentum occur between the atmosphere and the land surface, including
its biosphere. It is also the surface on which people live. The heat storage factor
of land surface with respect to atmospheric temperature variations is much less
than that for the oceans. Land has a lower specific heat than the ocean, and its
rigidity restricts heat transport to deeper levels. As a result, the depth of the soil
layer which is important for energy exchange interactions with the atmosphere is
only several meters for the annual-cycle time scale. A cave 20 meters underground
will remain at the same temperature all year round. Because of the small heat
capacity of the land surface, variations in atmospheric temperature just above the
surface are much larger over the land than over the ocean.
The energy and momentum exchanges between land surfaces and the atmos-
phere are similar to those for an ocean surface. Heat and latent heat (water
vapour) exchanges depend on temperature and water vapour pressure differences
between the land surface and the lower atmosphere, roughness of the land
surface, and surface atmospheric wind speed. The latter may be characterized by
wind conditions in the lowest ten meters of the atmosphere (the atmospheric
‘mixed layer’).
Radiation transfer is the other important energy exchange. The amount of
solar radiation absorbed by a land surface depends on both the amount of solar
radiation coming through the atmosphere (a highly variable quantity as discussed
before) and the albedo (reflectivity) of the land surface which is also highly vari-
able. The albedo ranges from five to 90 per cent and depends on the type of cover
for the land surface as shown in Table 1.1. The infrared radiation transfer is the
net of the infrared radiation emitted by the land surface (which is close to the
maximum ‘black body’ value and thus dependent only on temperature) and the
total downward infrared radiation produced by the atmosphere. Because of the
small heat capacity of the land surface, the radiative, sensible, and latent energy
transfers come close to balancing most of the time.
Topography of the land surface has a pronounced effect on large-scale atmos-
pheric circulations. particularly in the Northern Hemisphere. The Rocky
Mountains, which are oriented north-south transect the Northern Hemisphere
westerlies, and the Tibetan Plateau with its extreme height and aerial extent
affects flow over a large area. Topography is a factor in the wave patterns in the
upper tropospheric horizontal wind flow (shown in Figure 1.9), and also has
major effects on surface temperature and rainfall.
Alteration of land surface by human activity is an important factor in climate
change that adds to the effects of human-produced changes in the radiative char-
acteristics of the atmosphere.
Urbanization, cultivation for agriculture, irrigation, and deforestation
change the albedo of land surfaces and the surface sensible and latent heat trans-
fers. These factors can also greatly influence the local aspects of climate change.
1.4.5 The cryosphere — the ice component — has significant impacts on the climate
CRYOSPHERE system in several ways. It affects radiative and sensible heat transfers at the
Earth’s surface. It influences temperatures in the ocean and at the Earth’s surface
due to transfers between latent and sensible energy during melting and freezing.
Finally, its melting and freezing influences water runoff from land and ocean
salinity. Ice and snow exist primarily in the latitudes poleward of 30 degrees
latitude and are thus are unfamiliar to the majority of the world’s human
population. Although only about two per cent of all the water on Earth is frozen,
17
Table 1.1 — Albedos for various

surfaces. [from page 88 in Surface type Range Typical value
Reference no. 1, with (in per cent)
Water
Deep water; low wind, low altitude 5–10 7
Deep water; high wind, high altitude 10–20 12
Bare surfaces
Moist dark soil, high humus 5–15 10
Moist gray soil 10–20 15
Dry soil, desert 20–35 30
Wet sand 20–30 25
Dry light sand 30–40 35
Asphalt pavement 5–10 7
Concrete pavement 15–35 20
Vegetation
Short green vegetation 10–20 17
Dry vegetation 20–30 25
Coniferous forest 10–15 12
Deciduous forest 15–25 17
Snow and ice

Forest with surface snowcover 20–35 25
Sea ice, no snowcover 25–40 30
Old, melting snow 35–65 50
Dry, cold snow 60–75 70
Fresh, dry snow 70–90 80
it covers an average of 11 per cent of the world’s land surface and seven per cent
of its oceans.
There are many constituents to the cryosphere: land ice in polar ice sheets,
glaciers, permafrost, frozen ground, seasonal snow cover and sea ice. Table 1.2
summarizes the amounts of ice in the various categories. Note that although
Antarctica and Greenland between them account for 98 per cent of the world’s
land ice, the total area covered by ice and snow can be much larger in the
Northern Hemisphere winter. Figure 1.14 shows maps of ice coverage. The
Northern Hemisphere has a much larger seasonal range than the Southern
Hemisphere because of the larger amount of land area.
The albedos of ice and snow are higher than the albedo of the land or ocean
surface that they cover (see Table 1.1). Thus, their seasonal variations in coverage
will cause important seasonal variations in the Earth’s surface albedo. The impact
is less than might be first thought because in the winter season when coverage is
maximum, the solar radiation is at a minimum and solar energy is less important
in the atmospheric energy balance. Nevertheless, ice and snow introduce a
process of positive feedback as the expansion of ice and snow coverage increases
the albedo which, in turn, decreases solar heating. The resultant cooling acts to
further enhance the ice and snow cover.
Snow and ice cover have strong insulation effects for sensible heat transfer
which reduces heat transfer from the Earth, oceans, and lakes to the atmosphere.
Over the ocean (and lakes), snow and ice cover effectively cut off the moderating
effects of the water so that air temperature over the ocean ice cover can fall well
below freezing point.
Cryosphere conditions may not be directly evident to most people of the
world. However, because of the global nature of the climate system, the cryos-
phere component also influences lower latitudes.
18
Table 1.2 — Estimated global

Area Volume Per cent of
inventory of land and sea ice*. (km2) (km3) total ice
[After Untersteiner, 1984 from mass
page 15 in Reference no. 1, with
permission of Academic Press]. Land ice Antarctic ice sheet 13.9 × 106 30.1 × 106 89.3
Greenland ice sheet 1.7 × 106 2.6 × 106 8.6
* Not included in this table is the Mountain glaciers 0.5 × 106 0.3 × 106 0.76
volume of water in the ground Continous 8 × 106 (ice content)
that annually freezes and thaws at Permafrost 0.2–0.5 × 106 0.95
the surface of permafrost (‘active
layer’), and in regions without Discontinous 17 × 106
permafrost but with subfreezing Eurasia 30 × 106
winter temperatures.
Sea snow 2-3 × 103
(average maximum) America 17 × 106
Sea ice Max. 18 × 106 2 × 104

Southern Ocean
Min. 3 × 106 6 × 104
Max. 15 × 106 4 × 104
Arctic Ocean
Min. 8 × 106 2 × 104
Figure 1.14 — Maximum extent

of snow and sea ice during winter
(a-panels) and maximum extent
of snow and sea ice during
summer (b-panels) in the
Northern Hemisphere (upper
panels) and Southern Hemisphere
(lower panels) after Untersteiner
(1984). Permafrost regions are
also shown in the Northern
Hemisphere summer. The ice limit
was taken to be at concentrations
15 per cent in the Southern
Hemisphere. [from pages 208
and 209 in Reference no. 5,
with permission of Springer-
Verlag].
1.4.6 The biosphere is a component of the climate system that has a distinct role in the
BIOSPHERE interactions of both the oceans and land surface with the atmosphere. Vegetation
on the land surface and both plant and animal life in the oceans are all relevant
elements of the biosphere component that interact with the atmosphere.
Important exchanges between the terrestrial vegetation and the atmosphere are
summarized in Figure 1.15.
19
Figure 1.15 — Important Atmosphere

exchanges between the
atmosphere and terrestrial
Momentum Sensible Hheat
ecosystem. [from page 174, Radiation Latent Heat
Precipitation Trace Gases
Reference no. 6, with CO2
Deposition
permission of Cambridge NH4 CH4
NO3 N2 O
University Press]. SO4 .
O3 .
. .
.
.
Key Characteristics
Leaf Area
Height (roughness)
Albedo
Soil Moisture
Nutrient Status
Terrestrial Ecosystems
Climate conditions of the atmosphere have a direct effect on the type of

terrestrial plant growth at the Earth’s surface, as summarized in Figure 1.16. The
nature of the plant cover in turn feeds back on the atmospheric condition by
influencing the sensible and latent energy transfers from a land surface, as well
as surface layer turbulence in the atmosphere (through its roughness properties).
Furthermore, land vegetation is a significant reservoir for carbon with a total
carbon content nearly equal to that in the atmosphere (Figure 1.13). Changes in
the amount of land vegetation due, for instance, to forest cutting and burning or
simply seasonal changes have a direct impact on the carbon dioxide
concentration in the atmosphere. Along with dissolved inorganic carbon and
calcium carbonate solids, plant and animal life have key roles in the ocean, in
Figure 1.16 — An example of a

simple classification of
vegetation types of the world
based on annual precipitation
and mean annual temperature
(Whittaker, 1975). [from page
176, Reference no. 6, with
permission of Cambridge
University Press].
20
the carbon cycle which influences the concentration of the greenhouse gas, and
carbon dioxide in the atmosphere and results in a loss of carbon due to
sedimentation of carbonates at the ocean bottom.
The interactions between the biosphere and atmospheric climate have
produced a record of past climate conditions. Tree rings, fossil patterns, pollen
counts in ocean and lake bottom sediments, and coal and oil deposits are records
which give us information on past climates.
Humans, themselves, are members of the biosphere. Humans alter the bios-
phere directly by agricultural and forestry activities and indirectly by altering the
climate system in which the biosphere exists. It is important to understand these
various impacts of human activity in order to understand climate change.
The biosphere must be included in the climate system analysis in order to
understand climate change. It is a component that interacts with other climate
system components, and is the component where the effects of climate change
will be clearly evident to people.
1.5 There are numerous significant physical interactions among the components of
FEEDBACKS IN THE the climate system that are relevant to climate change. A brief overview of these
CLIMATE SYSTEM components is presented below, they include: radiation energy transfer, heat
energy transfer, and biosphere interactions.
1.5.1 Radiation is a primary mechanism for energy transfer in the climate system. At
RADIATION ENERGY TRANSFER the same time characteristics of the climate system itself have a great impact on
the magnitudes of radiative energy transfer. There are two key feedbacks as
described below.
The interrelationship between temperature and radiation provides a negative

1.5.1.1 feedback whereby radiation transfer tends to reduce variations in temperature
Temperature feedback and to stabilize temperature conditions. This situation arises for two reasons.
First, the magnitude of radiation emission from substance depends on the
(absolute) temperature of the substance raised to the fourth power. Second,
radiation emission represents a loss of energy from the substance causing its
temperature to decrease. In the climate system the radiation involved is of the
infrared (terrestrial) type. Thus increasing temperature will lead to increased
radiative cooling.
1.5.1.2 A primary energy source for the climate system is the absorption of solar (visible)
Albedo feedback radiation. The amount absorbed is dependent on the reflectivity (albedo) proper-
ties of the substance. Since there is a large variation of albedo for substances in
the climate system, significant feedbacks exist based on variations in amounts of
specific substances. Important albedo feedbacks exist for ice cover, cloud cover,
and land-surface characteristics. The feedback is positive for the ice cover because
an increase in ice cover raises the overall albedo at the Earth’s surface which tends
to reduce surface temperature thereby making it possible for the ice cover to
increase even more. An increase in cloud cover would also tend to cool the Earth’s
surface temperatures based on albedo effects since cloud albedo tends to be higher
than that of the Earth’s surface. However, clouds also affect infrared radiation
transfer so that the net effect on the Earth’s surface temperatures may be a
warming or a cooling. Overall land surface albedo varies according to land use,
the type of plants and ice cover. This provides an important albedo feedback
related to human activity, climatic conditions and biosphere cycles.
1.5.2 Sensible and latent heat energy transfers provide for important energy-related
HEAT ENERGY TRANSFER transfers between the components of the climate system and involve important
feedbacks. Vertical energy transfers between the ocean and atmosphere have
already been mentioned. To this must be added the energy transfers between land
and atmosphere and between ice and ocean water. The latent heat component
arises from the phase change of water between its vapour, liquid, and solid forms.
In many cases the transfer of latent heat energy can be as significant as that of
21
sensible heat. Recall that in the global mean, the latent energy transfer from the
Earth to the atmosphere was much larger than the sensible heat transfer (Figure
1.4). The dependence of the equilibrium saturation vapour pressure of water on
temperature introduces a significant role of temperature into the latent heat
energy feedback.
Because of their fluid nature, both the atmosphere and ocean transfer signif-
icant amounts of heat energy by horizontal motions. An important feedback
between the atmosphere and ocean exists with regard to the latitudinal (pole-
ward) energy transfer. If, for instance, the ocean poleward heat transport were to
change due to internal conditions, there would be a change in the oceanic latitu-
dinal temperature variations which would affect the latitudinal temperature
variations in the atmosphere. This, in turn, would alter atmospheric circulations
and poleward heat transports.
1.5.3 Important two-way feedbacks between the atmosphere and biosphere were
BIOSPHERE INTERACTIONS discussed earlier in Section 1.4.6. On one hand, the biosphere, as a central compo-
nent in the carbon cycle, is a key determiner of greenhouse gas concentration in
the atmosphere. On the other hand, the atmosphere, in particular its temperature
and precipitation, has a major influence on the biosphere. The biosphere in both
the atmosphere and oceans plays a significant role in the carbon cycle which
includes the atmospheric carbon dioxide.
1.6 The circulations in the atmosphere and ocean transmit changes in one region of
GLOBAL NATURE OF THE the climate system to broad sectors of the world. This means that many aspects
CLIMATE SYSTEM of both natural climate variability and climate change are global in nature. This
1.6.1 makes the climate change issue a global one which will require the understand-
INTRODUCTION ing of people everywhere and the participation of all countries in dealing with its
impacts. A few examples of this global scope are presented here.
1.6.2 The depletion of the stratospheric ozone layer in recent years has been shown to
OZONE HOLE IN THE be due to chemical effects arising from the introduction of chlorofluorocarbons
STRATOSPHERE (CFCs) into the atmosphere. The CFCs were primarily used in the manufacture of
refrigeration systems, in plastics blowing agents and in aerosol spray-can propel-
lants. Sources of these gases may have been originally in the industrialized
countries (primarily in the Northern Hemisphere); however, the primary effect
has been seen in the reduction of the stratospheric ozone concentrations at polar
latitudes, particularly in the Southern Hemisphere, with associated impacts on
human life in Australia and southern parts of South America. There has also been
significant stratospheric ozone reduction in the high latitudes of the Northern
Hemisphere. The long lifetime of CFCs means that the impacts are felt for many
years after the gases were released into the atmosphere. Thus, overall, the impacts
of CFC emissions by human activity extend far, both in time and space, from
their source points.
1.6.3 The El Niño phenomenon in its original definition referred to warmer-than-

EL NIÑO — SOUTHERN normal temperature conditions on the ocean surface off the coast of Peru. In
OSCILLATION (ENSO) recent times the definition has been expanded to refer to warmer-than-normal
conditions on and near the equator in the eastern half of the Pacific Ocean. Figure
1.17 shows the typical pattern for the sea-surface temperature in the eastern trop-
ical Pacific during El Niño. The specific example is for the very strong El Niño in
1997-98. This condition ties into an atmospheric oscillation called the Southern
Oscillation to cause anomalous conditions in both the ocean and atmosphere in
the tropical Pacific and Indian Ocean area. Figure 1.18 shows the patterns of
anomalies in surface atmospheric pressure of the Southern Oscillation which
accompanies El Niño. Notice that pressure in the eastern tropical Pacific area is
lower than the mean where the sea-surface temperatures are higher than the
mean. Figure 1.18 actually shows correlations of surface pressure to that observed
in Darwin, Australia, where the pressure is higher than the local mean value
during El Niño.
22
Figure 1.17 — Anomalous sea-

surface temperature for December
1997. Contour interval is 1°C.
Dashed contours indicate
negative anomalies. Anomalies
are departures from the adjusted
optimum interpolation
climatology (Reynolds and Smith,
1995). [from page 24, Climate
Prediction Center, 1997].
Observational data reveals ‘teleconnections’ (significant correlations)

between ENSO conditions in the Pacific Ocean areas and variability in many
other parts of the world. Correlations have been found not only in atmospheric
conditions but also in ocean conditions such as for Indian Ocean sea-surface
temperature anomalies as shown for the 1997-98 El Niño in Figure 1.17.
Correlations are quite evident for weather conditions in the western coastal
areas and southern region of the United States, the northeastern part of Brazil,
and the eastern part of Asia as will be discussed in Chapter 2. Clearly the fluid-
dynamic processes in the atmosphere and ocean do much to give global-scale
perspective to what may appear as a regional phenomenon.
1.6.4 The monsoon refers to quasi-stationary circulation patterns and associated

MONSOON weather that exist on a seasonal basis due to surface temperature contrasts
between continents and surrounding oceans. Monsoons cause large rainfall
amounts over tropical and subtropical regions of Asia and Africa in summertime.
Note, for example, the difference in precipitation amounts over south-east Asia
between July (summer) and January (winter) as shown earlier in Figure 1.7. This
regional rainfall pattern is a well-recognized aspect of a monsoon. However,
further examination shows that the Asian monsoon, in particular, has impacts
which extend over most of Asia, the Pacific Ocean and down into the Australian
and Indian Ocean areas; truly a global scale. A schematic example of the low-level
aspects of Asian monsoon circulation into these areas is shown in Figure 1.19.
Variation in conditions in one part of the monsoon region may relate to condi-
tions at quite distant locations. For instance, correlations have been found
between monsoon rain intensities over south-east Asia and sea-surface tempera-
tures in the eastern Pacific Ocean south of the equator.
1.6.5 The relatively recent major eruptions of El Chichon in Mexico in 1982 and Mount
VOLCANOES Pinatubo in the Philippines in 1991 serve as good examples of the global impacts
of volcanic eruptions. In both cases the aerosols and gases from the eruptions
spread around the world in the latitude bands of the volcanoes within a few
Figure 1.18 — Horizontal
distribution of the correlation
coefficient between annual-mean
sea-level pressure anomalies over
the globe and the corresponding
pressure anomalies in Darwin,
Australia (12°S, 131°E) as a
measure of the Southern
Oscillation. The map shows that
global shifts of atmospheric mass
take place during ENSO episodes
(adapted from Trenberth and
Shea, 1987). Areas with
anomalies greater than 0.4 are
stippled. [from page 422,
23
Figure 1.19 — Asia’s monsoon

circulation occurs in conjunction
with the seasonal shift in the
Inter-Tropical Convergence Zone
(ITCZ) [Shown as ITC in the
diagram]. (a) In January, a
strong high pressure develops over
Asia and cool, dry continental air
generates the dry winter
monsoon. (b) With the onset of
summer, the ITCZ migrates
northward and draws warm
moist air onto the continent.
[from page 179, Lutgens and weeks and, eventually over several months, to most other latitudes. In many parts
Tarbuck, 1995, with of the world these aerosols resulted in unusual colorations in the morning and
permission of Pearson evening skies. The dust from Mount Pinatubo persisting at stratospheric levels
Education Publications]. resulted in a decrease of solar radiation at the Earth surface and a measurable
decrease in global-mean surface temperature of several tenths of a degree Celsius
for the following year in the Northern Hemisphere.
1.7 Climate has a global nature, but it also has local-scale variability which is very
REGIONAL NATURE OF important for its impacts on life. Local variations are caused by topography,
THE CLIMATE SYSTEM differences in ground cover at the Earth’s surface and organization of weather
1.7.1 systems resulting from the atmospheric general circulation in tropical regions,
INTRODUCTION such as the Inter-Tropical Convergence Zone (ITCZ). Mean temperature and rain-
fall conditions change markedly with elevation of the land. Variations also exist
upstream and downstream of topography with effects such as ‘rain forests’ or ‘rain
shadows’ where the mean rainfall is greater or less, respectively, than in surround-
ing areas. At the very small scale, farmers know that the slope of the surface and
small valleys in fields can noticeably alter growing conditions. Even the areas of
shade and sun around one’s home provide for very small-scale climate variations
(microclimates) which affect plant growth. These local variations are significant
for defining one’s personal environment.
1.7.2 There are a wide range of climate conditions on Earth. It is useful to characterize
GEOGRAPHY OF CLIMATE them in terms of surface temperature and precipitation because these aspects
relate directly to the biosphere and human activities. A classification system that
has evolved from the original work of the Russian-born German climatologist,
Wladimir Köppen, in the early 20th century is commonly used (Köppen, 1931).
The original classification has five basic categories denoted by the first (upper
case) letter in the classification code. Subcategories denoted by letters after the
basic category letter provide details on seasonal variations in temperature and
rainfall. Note that temperature criteria refer to monthly mean values. An update
by Trewartha and Horn (1980) to this climatic classification is summarized below.
Note that it has two additional basic categories.
A (Tropical humid): Warm enough and sufficient moisture (on an annual basis) for
plant growth year round; essentially no winter with no frost in continental areas
and mean temperature of coldest month 18°C or greater in maritime areas.
B (Dry): Plant growth limited by moisture supply alone (Steppe and Desert climates)
C (Humid middle-latitude with long growing season): Sufficient moisture (on an
annual basis) for plant growth; monthly-mean temperature equal to or greater
than 10°C for at least eight months.
D (Humid middle-latitude with short growing season): Sufficient moisture (on an
annual basis) for plant growth; monthly mean temperature equal to or greater
than 10°C for at least four months, but less than eight months.
E (Boreal subarctic): Very short summer with monthly mean temperature equal to
or greater than 10°C for only one to three months.
24
F (Polar): Essentially no growing season; for the warmest month the mean-monthly
temperature is less than 10°C in the Tundra climate and less than 0°C in the Ice
Cap climate.
H (Highland): Not a type of climate; high elevations are an important factor in
climatic conditions
Figure 1.20 shows the overall climate characteristics for the land areas of the
world in terms of the classification summarized above together with subcate-
gories. The key in the diagram describes the subcategories.
1.7.3 Mean rainfall distributions in tropical areas can have large variability over small
LOCAL VARIATIONS distances. The rainfall in Africa is one example. Mean annual and seasonal rain-
1.7.3.1 fall in the Sahel area of Africa can change rapidly over small distances in the
Rainfall north-south direction due to the quasi-stationary and small-scale structure of the
ITCZ. Changes in annual rainfall amounts are as large as from roughly 1000 mm
at 10°N to 50 mm at 17°N (a distance of roughly 700 km). Similar variations occur
in the tropical region of South America. The largest variations are between the
western-most portion of the Amazon River basin and the west coast of South
America due to the Andes Mountains.
As an example for subtropical areas the mean annual rainfall in the India-
Pakistan area has large variations due to the topographic influences of the Tibetan
mountains and lower mountain ranges and from the structure and persistence of
the monsoon circulation. Rainfall values are as low as 200 mm in Pakistan and
reach 1000 mm in the western part of India only 400 km to the south-east. Three
hundred kilometers further down the west coast of India, rainfall values reach
3000 mm.
Ocean islands, especially those in the trade-wind areas, can have large varia-
tions in mean annual rainfall from one side to another depending on topography.
The Hawaiian Islands are a good example.
1.7.3.2 Temperature variations due to altitude or proximity to oceans can be large over
Temperature short distances. The nature of agricultural crops in tropical countries can change
rapidly with elevation in mountainous areas. This is illustrated in countries such
as Honduras and regions such as eastern Africa. Mean wintertime temperatures in
Juneau, Alaska are 20°C warmer than places 100 km inland.
25
26
Figure 1.20 — Climate of the earth [from G.T. Trewartha and L.H. Horn, 1980, inside cover, with permission from McGraw-Hill).
CHAPTER 2
NATURAL TEMPORAL VARIABILITY IN THE CLIMATE SYSTEM
2.1 The Earth’s climate exhibits natural variability on all time scales. Some of this, for
INTRODUCTION instance, that has occurred in surface temperature is much larger than anything
envisioned due to human impacts. It is a continuing challenge to demonstrate
variations that are due to anthropogenic causes rather than natural causes.
Variability magnitudes differ widely over the range of time scales. There are
a number of time scales for which the variability is markedly larger than it is on
slightly larger or smaller time scales. Many of these local peaks in magnitude can
be ascribed to identifiable forcing processes. An idealized variance spectrum for
the Earth’s surface temperature is presented in Figure 2.1 with identification of the
time scales of the maxima, where time scale is measured by the period of oscilla-
tion. The daily and annual cycles can be clearly seen. Many of the narrow peaks
relate to other astronomical and geological effects such as variations in the Earth’s
orbit, continental drift and mountain formation (forcing mechanisms ‘external’
to the climate system). Some of these are discussed in the next section. The
broader peaks relate to variability enhancements that include the effects of inter-
actions within the climate system (internal forcing mechanisms). The
temperature spectrum, as for other variables in the climate system, is ‘red’
meaning that amplitudes of the variations are larger for the longer time-scales.
The natural variabilities in the range of seasonal to millennial (three months to
thousands of years) may be considered as the most relevant in our discussion about
climate change. However, even the diurnal temperature variability is an important
factor in climate change as it is affected by changes in greenhouse gas concentration.
In this chapter basic forcing mechanisms for climate variability are discussed,
covering both external and internal types. Specific examples of variability are then
presented.
2.2 The sunspot cycle is a well-defined variation in the solar condition with a period
BASIC FORCING of about 11 years. The three-hundred year record shown in Figure 2.2 shows the
MECHANISMS regularity in the periodicity of the variation. The amplitude of the peaks varies
2.2.1 by a factor of two over the record and the periodicity itself ranges between 10
EXTERNAL FORCING
2.2.1.1
Astronomical effects
2.2.1.1.1
Variations in solar radiation
emission
Figure 2.1 — Idealized,

schematic spectrum of
atmospheric temperature between
10-4 and 1010 yr adapted from
Mitchell (1976). [from page 25,
27
Figure 2.2 — Annual mean

sunspot numbers from 1700 to
1991. [from page 288,
and 12 years. Although the presence of a sunspot itself (a relatively darkened

area on the sun’s surface) causes a reduction in solar radiation output, there are
extra bright regions called faculae, that are found in conjunction with sunspots,
that lead to an overall increase of solar radiation.
There is a direct relationship between the number of sunspots and the
increase in solar radiation. At the top of the atmosphere the difference in the solar
constant between the sunspot minimum and maximum is of the order of 1.5 Wm-2.
This difference results in an average change of solar radiation absorbed at the
Earth surface of about 0.2 Wm-2. This number is small, but not negligible, in
comparison with the 2.45 Wm-2 estimated as the overall change in greenhouse
gas forcing due to human-produced increases since the pre-industrial era.
The variation in the sunspot periodicity itself has also been shown to relate
to Earth surface mean temperatures, with shorter periods corresponding to
warmer temperatures. An overall review of solar effects on atmospheric climate is
presented by Hoyt and Schatten (1997).
2.2.1.1.2 The diurnal and annual cycles are both large amplitude variations. The rotation
Diurnal and annual cycles of solar of the Earth around its axis leads to the well-pronounced diurnal cycle which is
radiation input strongest in equatorial regions and does not exist at all at the poles where the sun
remains low in the sky (or just below the horizon) for the entire day.
The annual cycle of solar energy input is primarily caused by the tilt of the
Earth’s axis with respect to the plane of the Earth’s revolution around the sun.
This tilt is at present about 23.5° to the normal of the plane and leads to seasonal
variations in solar sunbeam zenith angle for all parts of the globe as well as vari-
ations in the length of the daylight period. Since the Earth’s orbit around the sun
is elliptical rather than circular, the variation in distance from the sun causes
28
CHAPTER 2 — NATURAL TEMPORAL VARIABILITY IN THE CLIMATE SYSTEM
Figure 2.3 — Schematic diagram

of the Earth’s elliptical orbit
about the sun showing the
critical parameters of eccentricity
(e), obliquity (Φ), and longitude
of perihelion (Λ) defined relative
to the vernal equinox. The size of
the orbit is defined by the
greatest distance between the
ellipse and its center point, which
is called the semi-major axis
length, ao. The Earth–sun
distance at any time (d), the
angle between the position of
Earth and perihelion that we call
the true anomaly (v), and the
angle between the position of
Earth and the vernal equinox (ω)
are also shown. [from page 303, additional fluctuations in the amount of solar radiation received on Earth. This
Reference no. 1, with distance factor produces a variation of 6 per cent in solar radiation intensity
permission of Academic Press]. between the July 4 minimum and the January 3 maximum, an annual variation
that is largely masked by the axis-tilt effects for daily total solar radiation input.
The Earth is currently closest to the sun in the Northern Hemisphere in
winter and furthest in summer. Thus, in the Northern Hemisphere this distance
variation tends to offset the seasonal variability in solar radiation input. In about
11 000 years the sun will be the closest in the Northern Hemisphere summer and
furthest away in the winter tending to make the Northern Hemisphere summers
hotter and winters colder than they are at present.
The annual cycle of daily total solar energy input varies considerably with
latitude as shown in Figure 1.2 earlier. The range is most extreme at the poles
where there are six months with no energy input. At the equator the range of the
annual cycle is very small and a small semi-annual variability is observed. The
seasonal shift of the latitude where the sun shines straight down at noon (shown
by the dashed line in Figure 1.2) results in significant shifts in the Intertropical
Convergence Zone (ITCZ) and its associated weather. As a result, land areas very
near to the equator (in parts of tropical Africa, for example) where the ITCZ
crosses over twice in the year may have two rainy seasons each year.
2.2.1.1.3 Three of the Earth’s orbital parameters have long-term variations that can cause large
Variations in orbital parameters of variations in the range of solar energy input over the annual cycle. These are the
the Earth eccentricity of the orbit, tilt of the Earth’s axis (obliquity) and the positioning of the
Earth’s axis (see Figure 2.3). The periods for oscillations in these three parameters are
approximately 100 000, 41 000, and 22 000 years respectively. Milutin Milankovitch
was the first person to theorize that these orbital variations could be responsible for
climate variability related to ice ages. The positioning involves a precession of the Earth’s
axis that leads to changes in the season of the year when the Earth is closest to the sun.
This precession effect may be described by changes in the longitude of perihelion
relative to the vernal equinox.
Variations in these three orbital parameters cause significant effects on the
amount of solar radiation received on the Earth as a function of season and lati-
tude. It is possible for all three to be in phase and cause variations in seasonal
insolation as large as 30 per cent in high latitudes (p. 307, Reference no. 1).
2.2.1.1.4 A far less predictable astronomical forcing effect is due to meteors. The impact of
Meteors large meteors with the Earth can cause a major short-term variation in the Earth’s
climate due to the production of large amounts of dust and smoke. Adding large
amounts of dust and smoke can greatly reduce the amount of solar radiation that
gets to the Earth surface causing surface temperatures to decrease. A current
theory for the extinction of dinosaurs is that a giant asteroid hit the Earth about
29
65 million years ago. This introduced so much dust that surface conditions
became much darker and much colder; the significant effect lasting for about
three years.
2.2.1.2 Tectonic effects such as continental drift and changes in mountains produce very
Geological effects significant changes in climate. Their evaluation over time is important for deter-
2.2.1.2.1 mining past climates. The time scales associated with these changes are the order
Tectonics of millions of years or greater. Since they are far greater than those relevant for
anthropogenic climate change effects, they are not discussed here.
2.2.1.2.2 Volcanoes can inject vast amounts of dust and gases into the atmosphere. An
Volcanoes individual eruption may affect climate conditions for up to three years. On
average newsworthy eruptions occur every 20 years; major eruptions with a
significant impact on the global climate will perhaps take place every 100 years
or so. (The statistics can be better restated as: in a given year the probabilities of
newsworthy and major eruptions are five and one per cent, respectively.)
Volcanoes provide intermittent forcing impacts and are not generally associated
with longer-term climate variations.
Volcanic products which get into the stratosphere can have a significant
impact. Products that remain in the troposphere, such as the dust, are subject to
rather rapid removal processes by gravitational settling and washout by precipita-
tion. On the other hand sulphates formed from the sulphur dioxide injected into
the stratosphere by more severe eruptions can have lasting effects. They are small
particles with very small settling speeds. The stratosphere has little up-and-down
motion and little precipitation to remove the sulphates. The primary effect of
sulphate aerosols on radiation transfer is to reduce the short-wave solar radiation
reaching the Earth’s surface. This causes the cooling associated with volcanic
events.
2.2.2 For variability time-scales of less than a month, the ocean generally has a
INTERACTION OF CLIMATE damping effect on amplitudes of climate system oscillations. Its heat capacity is
SYSTEM COMPONENTS large compared with that of other interactive components in the climate system,
2.2.2.1 and horizontal transport effects due to currents are small on these time-scales.
Ocean effects Thus, ocean temperature temporal variations are small, and variability of energy
exchange processes that depend on temperature differences will be primarily a
function of atmospheric variations.
On a seasonal time-scale, the ocean can increase the synoptic-scale transient
fluctuations in the extratropical atmosphere by virtue of its slow temperature
variations. By remaining nearly constant in temperature as the continents cool off
in winter, surface temperature contrasts at the coasts of continents in the
Northern Hemisphere increase. This fosters baroclinic processes in atmospheric
disturbances and enhances synoptic-scale variability. Synoptic-scale activity can
have a net effect on the general circulation due to correlations among synoptic-
scale flow and temperature components. For longer time scales of variability,
processes within the ocean begin to provide mechanisms that allow for two-way
interactions with atmospheric processes. These interactions foster significant
amplitudes of variability at the interannual and longer time scales. A striking
example of this is the ENSO phenomenon which introduces into the climate
system a pronounced global-scale variability on an interannual time-scale with
periods ranging mostly from two to seven years.
ENSO has both ocean and atmosphere components. The primary two-way
interactions occur in the tropical area of the Pacific Ocean. The ocean component
has variations in tropical sea-surface temperature that are related to vertical
motion and changes in the depth of the ocean thermocline. The oceanic vertical
motions (upwelling) are associated with horizontal ocean currents forced by
atmospheric surface winds. The atmospheric component has variations in convec-
tive storm activity in the equatorial region caused by changes in the sea-surface
temperature. The convective activity influences the surface atmospheric pressure
and the associated surface winds. Processes involving horizontal advection,
30
upwelling, and wave propagation in the ocean are the primary determinants for
the time scale of the ENSO cycle.
In the climatological mean, the ocean temperature in the western equatorial
region of the Pacific is warmer than in the eastern equatorial region. This is
accompanied by a rainfall maximum, surface pressure minimum in the western
Pacific and surface trade wind flow in the atmosphere from east to west. During
El Niño events the temperature in the eastern Pacific is increased which enhances
rainfall in the east, reduces surface pressure difference between the east and west,
and reduces (or even reverses) the low-level atmospheric flow component from
east to west. In the opposite phase of the oscillation, termed ‘La Niña,’ the eastern
Pacific sea-surface temperature becomes colder than normal, reducing precipita-
tion in the region, enhancing the surface pressure difference between the east and
west, and accelerating the east to west flow component at low levels.
It is noteworthy that the atmospheric general circulation in the tropical
Pacific area has important east-west variations along with the north-south varia-
tions normally considered. The east-west variations are associated with the
‘Walker Circulation’ and the north-south variations are associated with the
‘Hadley Circulation.’
Primary indicators for ENSO are the anomalies in the sea-surface tempera-
ture of the eastern tropical Pacific Ocean (east of the international date line) and
averaged atmospheric surface pressure difference between the western and
eastern tropical Pacific Ocean regions (see Figure 1.17). The latter has tradition-
ally been represented by the monthly mean pressure differences between
Darwin, Australia and Easter Island in the eastern Pacific. Sometimes Tahiti, to
the west of Easter Island, is used instead of Easter Island. Both of these are south
of the equator. The surface pressure difference also provides a measure of the
east to west surface airflow and the stress force that it applies to the ocean
surface currents. During El Niño events the pressure difference between Darwin
and Easter Island decreases, the surface wind stress decreases and the eastern
Pacific sea-surface temperature increases.
Observation for the ENSO indicators described above are shown in Figure 2.4
for the period from 1949 to 1988. In the Figure it is easy to see the frequency of
occurrence of the El Niño and La Niña phases of the ENSO cycle. The major El
Niño event years are marked with arrows in the pressure difference graphs. Note
that the ENSO frequency increased in the 1990s with El Niño events in 1991,
1994, and 1997.
Earlier reference was made to the global influences of the ENSO cycle in
terms of correlated variations that have been observed at great distances (telecon-
nections) from the tropical Pacific area. Figure 2.5 shows some of these
ENSO-related anomalies (for the El Niño or warm phase), primarily for precipita-
tion. The sign of the anomaly is shown (wetter, dryer, or warmer) indicating the
length of time of duration and the time of year. The plus sign refers to the year
after the El Niño year.
There are negative precipitation anomalies in the west Pacific area and posi-
tive values in the central Pacific as the main precipitation area is displaced
eastward. Also, conditions are wetter than normal on the west coast of South
America. These wetter conditions and the warmer than normal sea-surface
temperatures along the west coast of South America have major impacts on
people living there, especially fishermen. The more distant impacts include
altered monsoon precipitation in India, altered rainfall in central and southern
Africa, reduced rainfall in Australia, reduced rainfall in Central America and the
northern part of South America, increased rainfall in the Gulf of Mexico area, and
warmer temperatures in Alaska and western Canada. Figure 2.5 also shows the
locations of Darwin and Easter Island, locations that have been used to define a
pressure gradient indicator for ENSO.
The ocean has a major role in decadal and longer time scales of variability of
climate. For these time scales, ocean circulations in the thermocline and deep
waters over vast regions of the ocean are important factors. There is observational
evidence of decadal variability in sea-surface temperatures in the Pacific and
31
Figure 2.4 — Two sets of

monthly-mean time series of
(a) a classical index of the
Southern Oscillation, i.e., the
pressure difference between Easter
Island and Darwin in mb (after
Wyrtki, 1982, updated with data
from Ropelewski, 1987;
(b) the zonal wind stress over the
central τ CEP
x equatorial Pacific
region (8°S-4°N, 160°E-130°W)
in units of Pa (Pa = 10 dyn
cm -2); and
(c) the sea-surface temperature
anomalies in the eastern
equatorial Pacific region
EEP
Ts (20°S-20°N, 180-80°W) in °C.
The upper set of three panels

cover the years 1949 to 1968, the
lower set covers 1969 to 1988.
The most important El Niño
events (i.e., when warm water
covers the entire eastern
equatorial Pacific Ocean)
according to Rasmusson and
Carpenter (1982, updated) are
indicated by arrows at the bottom
of (a). The thick solid lines
indicate a 12-month running
mean in (a) and 15-month
weighted means in (b) and (c)
(obtained by using a Gaussian-
type filter with eights 0.012,
0.025, 0.040, 0.061, 0.083,
0.101, 0.117, and 0.122 at the
central point). [from pages 424
and 425, Reference no. 5, with
32
Figure 2.5 — Schematic

representation of typical ENSO-
related precipitation anomalies
over the globe. Solid contours
enclose relatively dry regions
(light shading) and dashed
contours enclose relatively wet
regions (heavier shading). The
approximate period of extreme
conditions relative to the typical
El Niño (0) year is also shown for
the various regions (adapted from
Ropelewski and Halpert, 1987).
with permission of Atlantic Oceans. Oceanic processes that have a decadal time scale include advective
Springer-Verlag]. effects in the primary central gyre in the north Pacific and subduction processes
within the thermocline layer between the middle latitude and tropical regions. It
will require analysis with numerical models to sort out the role of these processes
as there is only a short record of observations for subsurface currents in the ocean.
Currents in the deep ocean layers below the thermocline owe their existance
to thermohaline density effects. Observational evidence has been obtained to
show that the deep ocean currents have a large scale structure and connect the
major ocean basins of the world. Figure 2.6 shows a schematic of this circulation
which is referenced as the ‘great ocean conveyor belt.’ The interconnection of the
deep ocean and surface waters brings the heat storage of the entire ocean into play
and produces century- and millennium-scale variability in sea-surface tempera-
ture and the atmospheric climate. A key part of the deep ocean circulation system
is the downwelling regions where dense water sinks to deep or even bottom layers
of the ocean in relatively small areas of rapid downward motion in the vicinity of
Greenland and Antarctica.
2.2.2.2 There are a number of important interactions of the cryosphere with the other
Cryosphere effects components of the climate system that can affect variability characteristics. None
of these introduce easily-defined time scales of oscillation, as is the case for the
ocean interactions. Ice cover is an indicator for changes in climate. The net
melting of mountain glaciers observed since 1850 is a clear example.
As earlier discussed, ice cover on the Earth’s surface has a dramatic feedback
effect on climate due to its high albedo. Briefly, an increase in ice cover by virtue
of increasing surface albedo causes less solar energy to be absorbed at the Earth’s

of the global ‘conveyor belt’
depicting global thermohaline
circulation (after Broecker, 1987).
with permission of Cambridge
University Press].
33
surface which helps to maintain the cool temperatures needed to sustain or

increase the ice cover. This positive feedback process is considered to be a factor
in the maintenance of previous ice age conditions.
There are several other interactive processes. The melting of ice over and near
the oceans in the polar regions can increase the supply of fresh water to the oceans.
This water will tend to stay at the surface because its low salt content tends to make
it less dense than sea water. Sufficient amounts of fresh water at the ocean surface
in the polar regions have the potential to reduce or even prevent the occurrence of
strong downward convective currents that supply cold water to the deep waters
and maintain the great ocean conveyor belt. Suppression or alteration of the great
ocean conveyor belt would have major global, long-term climate impacts.
Another interactive process involves the polar ice caps and the oceans. Most
of the ice in the cryosphere is found in the polar ice caps over Antarctica and
Greenland. As these ice areas lie on solid ground and are not floating in the ocean,
their melting will contribute directly to sea-level changes. It is estimated that if
these two ice fields melted completely, the sea level would rise more than 75
metres. It is also estimated that in the last major ice age (during the Pleistocene
about 18 000 years ago) sea levels were lower by about 100 metres. Such changes
in sea level would significantly affect the shape of ocean coastlines and the
impacts of ocean bottom topography leading to changes in the basic ocean circu-
lation. Interactions involving the cryosphere have time scales that can range
upwards to many thousands of years.
2.2.2.3 The biosphere, both on land and in the ocean, has a number of interactions that
Biosphere interactions can lead to climate variability. We are just beginning to understand some of the
aspects that could lead to variability cycles. Biosphere relationships with albedo
and greenhouse-gas concentrations provide two important factors for such
climate system variability. Many investigations have focused on the impacts of
human alterations of the terrestrial biosphere.
Albedo impacts on vegetation ground cover are shown in Table 1.1.
Vegetation can have albedos that are higher or lower than ground cover without
vegetation. It has been possible to show examples of the biosphere albedo inter-
active effects on the climate system through very simple models. The classical
example is the ‘Gaia’ model developed by Watson and Lovelock (1983) summa-
rized very nicely on pages 251-253 in Reference no. 1.
The ‘Gaia’ model has a world with three regions covered by black daisies (low
albedo), white daisies (high albedo), and Earth bare soil (intermediate albedo),
respectively. The growth and death rates of the daisies are prescribed functions of
regional temperature. The temperature in turn is determined from equilibrium
conditions for the overall global radiation budget (considering both the solar and
terrestrial radiation components) which also satisfies local energy balance for each
region. The local energy balance for each daisy type includes assumed rates of
energy transfer between the local areas depending on temperature differences.
The model defines an equilibrium solution in terms of the percentage of the area
covered by each type of daisy for a prescribed value of solar luminosity. No oscil-
lating solutions are produced in this simple model.
Daisy regions exist only for an intermediate range of solar luminosity values.
Values too low or too high give conditions which are too cold or too warm for
growth. For solar luminosity values at the low end of the range the black daisies
predominate. For luminosity values at the high end of the range the white daisies
predominate. The albedo of the daisy is a key factor in determining the tempera-
ture of the daisy’s growth environment.
The land-surface biosphere interacts directly with the greenhouse gases
(water vapour, carbon dioxide and methane) in the atmosphere. Water intake and
water vapour supply are part of the terrestrial biosphere life processes, and
methane is a product of decay. The marine biosphere interacts with the carbon
dioxide absorbed in the water (dissolved inorganic carbon) and converts it to solid
carbonates. The carbon dioxide gas can be transferred directly to and from the
34
atmosphere. The global carbon cycle presented earlier in Figure 1.13 shows
components for both the terrestrial and marine biosphere.
Time scales associated with biosphere interactions are short for responses of
the atmosphere to biosphere changes in comparison to those for the response of
the biosphere to changes in atmospheric conditions. The former involves changes
in albedo and surface energy and moisture transfer rates, whereas the latter
involves life cycles of plant life. The terrestrial biosphere response to the annual
cycle is clearly evident in the differences between summer and winter and
between wet and dry seasons. Changes in climate can cause changes in the types
of plants growing in a particular place and the gradual migration of the area where
a given plant species exists. A good example is the northward advance of the
boreal forests in Canada and Siberia in the last 10 000 years following the recent
ice age. Associated time scales vary from decadal to centennial to even longer. The
terrestrial biosphere is also affected by human activity (such as agriculture, forest
cutting, and urbanization) which introduces time scales dependent on the rate of
change due to human activity.
2.2.2.4 There are a number of low-frequency oscillations in the atmospheric flow which
Internal atmospheric processes are largely independent of interactions with other components of the climate
system or effects of external forcing. These have a variety of oscillation periods
and preferred latitudes for their existence, and may or may not have an impact
on climate. Many of them include nonlinear processes.
The stratospheric Quasi-Biennial Oscillation (QBO) is a variability in the
stratospheric flow field most apparent in the upper and middle stratosphere in
tropical latitudes. It has oscillations periods ranging from 22 to 34 months and
has minimal impacts on tropospheric climate. It is believed to result from the
interaction of vertically propagating gravity waves and the mean horizontal flow.
Another oscillation most pronounced in the tropical region is a tropospheric
intraseasonal oscillation, sometimes referred to as the ‘Maddan-Julian
Oscillation.’ This oscillation originates in the African area, develops as it moves
eastward over the Indian Ocean, and then diminishes as it reaches the eastern
Pacific Ocean. It has a period of oscillation ranging from 30 to 60 days. The
oscillation modulates tropical convective activity and is related to break periods
in the Indian monsoon. It results from interactions between eastward-moving
Kelvin waves in the atmosphere and convective activity.
A variability centered in the extratropical troposphere regions results from
nonlinear processes in the tropospheric jet streams. Nonlinear effects give rise to
‘index cycles’ and ‘weather regimes’ which involve persistent trough and ridge
patterns in the tropospheric jet streams and related modulations in extratropical
cyclone activity. The periods of oscillation are generally in the range of 10-20
days. A notable aspect of this variability is the development of a ‘blocking’ pattern
where a persistent and anomalous anti-cyclonic flow pattern prevents cyclones
and fronts from reaching given regions for a sustained period. This may cause a
drought situation to become very severe.
Nonlinear aspects of the atmospheric flow such as described for the extra-
tropics above seemingly lead to erratic variability characteristics. A focus on the
nonlinear aspects alone is described by mathematical chaos theory. Solutions
found by Lorenz (1990) for an idealized system of three components (a zonal jet
flow and two superimposed transient eddy components) showed some irregular
characteristics that resembled extratropical latitude flow variability.
His solutions for the strength of the zonal flow forced by a smoothly varying
annual cycle are shown in Figure 2.7 for a six-year period. The winter season is on
the left and right hand sides of the diagram and the summer season is in the
middle. The irregular oscillations appear to have some systematic aspects which
change as a function of season. The overall nature of the oscillation pattern
changes from year to year. In general chaos theory the variability is of ‘fractal’
nature meaning that it has similar complex structure for any time scale. Such
theory may be relevant to describing some aspects of climate variability.
35
Figure 2.7 — The variations of X

(dimensionless variable
representing the zonal flow
speed) with time, t (months) in a
6-year numerical solution of an
idealized three-component
nonlinear circulation model.
Each row begins on 1 January,
and, except for the first, each row
is a continuation of the previous
one. [from page 383, Lorenz,
1990, with permission of
Munksgaard].
2.3 For any time scale there are variations in the climate variables such as tempera-
OBSERVED CLIMATE ture, precipitation, and severe storms and in related climate system parameters
VARIABILITY such as sea-surface temperature and sea level. A few examples are shown here to
highlight magnitude, range, periodicity, and extremes in the variations. In the
examples you will see both systematic and irregular types of variability. It needs
to be emphasized that observational information itself has quantitative uncer-
tainty. The proper use of quantitative data requires being aware of their
uncertainties and what region and time scale they actually represent.
2.3.1 For the geological time scales a schematic for the Earth’s overall surface tempera-
SURFACE TEMPERATURE ture levels for the past 100 million years is shown in Figure 2.8. The curve shown
for the future (on the left side) is speculative and shows an early numerical model
estimate for effects where a doubling of CO2 was the only human impact consid-
ered. It is noted that climate change projections for temperature increases may be
less important than natural variations over time; however, the rate of change
might be larger in anthropogenic climate change compared to any natural rates
of change.
Primary features of temperature variations in the past 100, 1 000, 10 000,
and 100 000 years (up to about 1970) shown in Figure 2.9 reveal irregular varia-
tions at all these time scales. Note that the region covered and the data source
vary from panel to panel. Some of the recent variations relevant to the climate
change discussion are identified: the warm period ‘thermal maximum of 1940s
covering the 1930s-1950s; the particularly cold period ‘little ice age’ particularly
in the 18th and 19th centuries; and the ‘Younger Dryas cold interval’ about 10 000
years ago. The range for temperature variations increases with the time scale
which is consistent with the ‘red’ nature of the variance spectrum discussed at the
beginning of the chapter. For the last 15 000 years the range is 10°C in the middle
latitudes in the Northern Hemisphere (4-5°C for the global mean) whereas the
temperature range is only 1.5°C for the last 1 000 years over eastern Europe.
The details of surface global-mean temperature variations for the period
since 1860 (up to 2000), used as a reference for current climate change impact
studies, is shown in Figure 2.10. The values are actually deviations from the 1961-
1990 time period. Note that both yearly and smoothed depictions are used. The
level of uncertainty in the values is suggested by the plotted standard error bars
and the differences between two estimates, the dashed and solid curves, respec-
tively. Note that the standard error is larger for the earlier years.
Values for each hemisphere and the globe all show the general upward trend
in temperature. Based on this data it is now estimated that global mean surface
temperature has increased about 0.6 degrees C since the beginning of the 20th
century. It is also considered, with very high statistical confidence, that the last
decade (1900-1999) has been the warmest decade of the period. Furthermore,
36
Figure 2.8 — Rough schematic

comparison of possible future
greenhouse warming with
estimates of past changes in
temperature. Pleistocene glacial-
inter-glacial cycles are more
numerous than shown. The
characteristic amplitude of global
temperature change during
glacial-interglacial cycles is 3-4
K. Note that pre-Pleistocene
changes are not well fixed in
magnitude, but their relative
warmth is approximately correct.
Maximum warming in the
Cretaceous is based on estimates
by Barron and colleagues. Time
intervals in between have been
scaled accordingly (Crowley,
1989). [from page 671,
University Press].
Figure 2.9 — General trends in global climate for a variety of time scales. (a) Changes in the 5-year average surface
temperatures from about 1875 to 1970 averaged from instrumental records over the region 0-80°N (from Mitchell, 1963).
(b) Winter severity index for eastern Europe during the last 1000 years up to about 1970 (from Lamb, 1969).
(c) Mid-latitude Northern Hemisphere air temperature trends during the last 15 000 years based on changes in tree lines
[from La Marche (1974), marginal fluctuations in alpine and continental glaciers (from Denton and Karlén, 1973), and
shifts in vegetation patterns recorded in pollen spectra (from van der Hammen et al., 1971), (d) Northern Hemisphere air
temperature trends during the last 100 000 years based on mid-latitude sea-surface temperature and pollen records and on
worldwide sea-level records. (e) Fluctuations in global ice volume during the last million years as recorded in changes in
isotopic composition of fossil plankton in deep-sea core V28-238 (from Shackleton and Opdyke, 1973). [from page 256,
Reference no. 1, with permission of Academic Press].
37
Figure 2.10 — (a) to (c): (a) Northern Hemisphere

Combined annual land-surface 0.8
Optimum average (Folland et al., 2001)
relative to 1961 to 1990

air and sea-surface temperature
Area weighted average (adapted from Jones et al., 2001)
anomalies (°C) from 1986 to 0.4
Anomaly (°C)
1999, relative to the 1961 to
1999 mean value temperatures
(vertical solid bars) for Northern 0.0
Hemisphere, Southern
Hemisphere, and Globe −0.4
respectively; twice their standard
error (denoted by the thinner
−0.8
vertical lines with small 1860 1880 1900 1920 1940 1960 1980 2000
horizontal bars at the top and Year
bottom, where one end may be
partially obscured by the solid (b) Southern Hemisphere
bars); and time-smoothed curves 0.8
calculated using optimum

averages and area weighted 0.4
Anomaly (°C)
averages (solid and dashed lines,

respectively). Information is from
0.0
data bases at the U.K. Met.
Office and the Climatic Research
Unit at the Hadley Centre, both −0.4
in the United Kingdom. [from
page 114, Reference no. 7]
−0.8
1860 1880 1900 1920 1940 1960 1980 2000
Year
(c) Globe
0.8

0.4
Anomaly (°C)
0.0
−0.4
−0.8
1860 1880 1900 1920 1940 1960 1980 2000
Year
based upon indicators for temperature, it is considered likely that the increase in
temperature in the 20th century has been larger than that for any century in the
past 1 000 years.
2.3.2 Precipitation variability has quite different features from temperature variability.
PRECIPITATION Figure 2.11 presents precipitation variability for all of the main land areas of the
globe except for Antarctica for the past 90 years. Decadal variability is quite
evident as it was for temperature. The data suggests systematic variations of land-
area precipitation with latitude with increases in precipitation in the middle- and
high-latitudes of the Northern Hemisphere and decreases in the tropics. It is
difficult to see correlations between the temperature and precipitation variability
shown in Figures. 2.10 and 2.11, respectively. Also the precipitation does not have
a clear trend. Differences in precipitation variability from one area to another in
the tropics and subtropics is shown in Figure 2.12. Note that in northern Africa
multidecadal variations are quite evident whereas in the regions of Mexico and
India higher frequency oscillations are more evident.
38
Figure 2.11 — Changes in land- 60
Precipitation anomaly (mm)

surface precipitation averaged
over regions between 55°S and
85°N. Annual precipitation 30
departures from the 1961-1990
period are depicted by the hollow
0
bars. The continuous curve is a
smoothing of the same data.
[from page 28, Reference no. 3]. -30
-60
1900 1920 1940 1960 1980
Year
2.3.3 It is important to be aware of the variability in extreme events such as droughts,
SEVERE WEATHER floods and tropical storms. Drought conditions similar to those experienced in
Africa over several years in the 1980s can be devastating in semi-arid areas. Long-
term changes in water supply result in significant changes in areas covered by
semi-arid and arid conditions with important impacts on civilization. An inte-
grated picture of rainfall and evaporation in certain watershed regions may be
provided by lake level information such as for Lake Victoria, Tanzania and the
Great Salt Lake in the western United States. A 97-year record for lake level in the
Great Salt Lake shows a dramatic drop from 1925 to 1935 at the time of the very
dry period in the central and western part of the country and a 4 metre increase
during the wet 1982-85 period.
As a second example, climatic variations in severe storm activity for the past
100 years up to 1988 are shown for the case of tropical storms and hurricanes in
Figure 2.12 — Variations of tropical and subtropical land-surface precipitation anomalies based on the average of the
anomalies relative to 1961-1990 means from Hulme, 1991, Hulme et al., 1994, and the Global Historical Climate Network
(GHCN) [Vose et al., 1992; Eischeid et al., 1995]. Smooth curves are generated from nine-point binomial filters of the
annual anomalies. (from page 154, Reference no. 3).
39
Figure 2.13 — Time series of the

annual number of Atlantic
tropical cyclones reaching at least
tropical storm strength (open bar)
and those reaching hurricane
strength (solid bar) for 1886-
1988. The average numbers of
tropical storms and hurricanes
per year are 8.4 and 4.9,
respectively (adapted from
Neumann et al., 1981, updated).
[from page 448, Reference no. the Atlantic ocean area in Figure 2.13. The overall pattern in variations over time
5, with permission of Springer- is quite different from that for global surface temperature as shown above in
Verlag]. Figure 2.10.
2.3.4 Significant variations in sea-surface temperature exist at annual, interannual,

OCEAN CONDITIONS decadal and longer time scales which are important for atmospheric climate.
Pronounced interannual variability exists in the tropical Pacific sea-surface
temperature as part of ENSO described earlier in Chapter 1 (Figures 1.17 and 1.18)
and in section 2.2.2.1 (Figure 2.4) above. Oceanic variability at decadal scales has
been hard to detect and describe because of the short observational record.
Recently decadal variability in sea-surface temperature has been detected in
the central and northern Pacific Ocean area. In the north Atlantic the largest
oceanic decadal signal detected so far has been found in association with the
atmospheric North Atlantic Oscillation (defined by the atmospheric surface pres-
sure difference between Iceland and the Azores or Lisbon). The decadal
component of the North Atlantic Oscillation correlates with decadal variations in
sea-surface temperature, currents, salinity and sea ice in the north Atlantic. The
involvement of deep thermohaline currents can link this variability with other
parts of the ocean system.
Sea level itself is of interest to all shore land areas and ocean island inhabi-
tants. Long-term records for sea level show a general rising trend in the major
ocean regions in the range of 1-3 mm/yr along with decadal variations as large as
200 mm (see Figure 2.14).
The latest IPCC assessment (2001) concluded that the global mean sea-level
rise was between 10 and 20 cm in the 20th century and that this was mainly the
result of a rise in ocean temperature.
Figure 2.14 — Six long sea level

records from major world regions:
Takoradi (Africa), Honolulu
(Pacific), Sydney (Australia),
Bombay (Asia), San Francisco
(North America) and Brest
(Europe). Each record has been
offset vertically for presentational
purposes. The observed trends (in
mm/yr) for each record over the
20th century are, respectively,
3.1, 1.5, 0.8, 0.9, 2.0, and 1.3.
The effect of post-glacial rebound
(lowering relative sea level) as
simulated by the Peltier ICE-3G
model is less than, or of the order
of, 0.5 mm/yr at each site. [from
page 367, Reference no. 3].
40
CHAPTER 3
HUMAN IMPACTS ON THE CLIMATE SYSTEM
3.1 As discussed in Chapter 1, the primary human activities which cause climate
INTRODUCTION change are those which influence radiative transfer in the atmosphere and radia-
tion absorption on the Earth’s surface. Other human activities, such as heating
the atmosphere through combustion and changing surface wind flow by defor-
estation and building construction, are of secondary importance. There are four
specific impacts to be considered: increasing greenhouse gas concentrations in
the atmosphere, adding aerosols to the atmosphere, changing cloudiness and
changing surface conditions. The first two clearly have global effects. There is
incomplete understanding for climate impacts from changes in cloudiness due to
human activity. More observational and modelling studies are needed. The effect
could be large and global, particularly if the cloud changes are related to the
aerosols added by human activity. Land surface changes may have large impacts
on local climate; but have less impact globally as land covers only 29 per cent of
the earth’s surface.
There are several measures of the magnitude of human impact. An overall
measure is accomplished by comparison of current conditions of the atmosphere
and the Earth’s surface with those that existed before the industrial revolution
(considered to be before 1750). A measure used for gases and aerosols is the
current rate of increase in atmospheric concentration. This is expected to relate
directly to level of possible impact in the future. A convenient way to describe and
compare the overall impact of human-caused changes in greenhouse gases or
aerosols on radiation exchanges is by their ‘radiative forcing.’
Radiative forcing here is defined as the change in average net total radiation
in the planetary radiation budget at the top of the troposphere due to changes in
solar and infrared radiation with fixed vertical structure for temperature. Radiative
forcing can be used to describe the impacts of both anthropogenic and natural
changes in the physical system. A positive radiative forcing (increase in net down-
ward radiation) tends to cause average warming of the Earth’s surface and a
negative radiative forcing, average cooling. This definition was made by the IPCC
in 1995. As stated in that report: “For a range of mechanisms there appears to be
a similar relationship between global mean radiative forcing and global mean
temperature change. However, the applicability of global mean radiative forcing
to mechanisms such as changes in ozone or tropospheric aerosol (concentrations),
which are spatially very inhomogeneous, is unclear.” [p. 16, IPCC, 1995].
3.2 The natural atmosphere contains greenhouse gases (primarily water vapour,
ATMOSPHERIC GREENHOUSE carbon dioxide, ozone, nitrous oxide, and methane) which have a major impact
GAS ENHANCEMENT on determining temperatures in the atmosphere and at the Earth’s surface.
Human activity has provided additional sources for these and other gases that
have greenhouse-gas characteristics. The result is an enhanced greenhouse effect
which is expected to force increased temperature at the Earth’s surface and in the
lower atmosphere.
All greenhouse gases have sources and sinks which may include chemical
conversions in the atmosphere. The effect of human activity on concentration
levels in the atmosphere depends on the cumulative amounts added by human
activity and the strength of the sinks.
For each greenhouse gas constituent one can define an ‘adjustment time’
which describes the rate of reduction of a concentration enhancement. For
instance, if an unusual event were to suddenly increase carbon dioxide
concentration in the atmosphere by 100 parts per million by volume (ppmv), the
adjustment time may be defined as the time it takes to reduce the concentration
enhancement to 1/e of its initial value (where e is the base of natural logarithms
41
or about 2.71) or to roughly 37 per cent of its initial value, namely 37 ppmv. This
definition assumes that the rate of depletion of the gas concentration is
proportional to the concentration. Sometimes the adjustment time is equated to
the ‘lifetime’ of the constituent. The lifetime is defined as the total content of
the constituent divided by the current rate of removal. The lifetime definition is
used for stationary input of a tracer into the atmosphere.
3.2.1 Three of the five natural greenhouse gas constituents are undergoing well-docu-
NATURAL GREENHOUSE GAS mented increases in concentration due to human activity: carbon dioxide,
CONSTITUENTS methane, and nitrous oxide. The other two naturally occurring greenhouse gases,
water vapour and ozone, are less directly linked to human activity. The role of
ozone is small. The relevant concentration for ozone includes both tropospheric
and stratospheric components and not just the stratospheric concentration values
for which there has been an observed decrease in polar regions. Water vapour is
the dominant greenhouse gas and accounts for about 75 per cent of the overall
greenhouse effect. However, its modification results primarily from changes in
evaporation rates from the oceans due to temperature and surface wind variabil-
ity and not from human activity.
Carbon dioxide enhancement is the most important human impact on the
greenhouse gases. This enhancement accounts for more than half of the total
enhanced greenhouse effects due to human activity. A continuous observational
record for atmospheric carbon dioxide has been obtained since 1958 at Mauna
Loa, Hawaii and since 1957 at the South Pole. Concentrations prior to that time
have been measured from air trapped in ice in Antarctica. Both of these measure-
ments are considered to be representative of global mean concentration. Carbon
dioxide concentration is rather uniform in the atmosphere.
The change in concentration values from the pre-industrial period to about 1989
reveals that the rate of increase itself is increasing with time as shown in Figure 3.1. The
figure also shows estimates for the production of carbon dioxide from fossil fuel
combustion and cement manufacturing since 1860. On a linear scale the shape of the
Table 3.1 (below) — Average production curve would be similar to that shown for carbon dioxide concentration. The
annual budget of CO2 rate of production has increased by nearly a factor of 10 since 1900.
perturbations for 1980 to 1989. The overall human impact on carbon dioxide concentration has been to increase
Fluxes and reservoir changes of it from roughly 278 ppmv to 365 ppmv (the value in 1998), an increase of 87 ppmv or
carbon are expressed in GtC/yr, almost 31 per cent. The current level is estimated to be higher than at any time since
error limits correspond to an the last interglacial warming about 120 000 years ago. At the current rate of increase,
estimated 90 per cent confidence the present (1998) carbon dioxide concentration will double in less than 100 years.
interval. [from page 79, Estimates for the budget for the carbon introduced into the atmosphere
Reference no. 3]. by human activity from 1980 to 1989 are presented in Table 3.1. Shown are
IPCC 1992† IPCC 1994* IPCC 1995

Estimates for 1980s budget
CO2 sources
(1) Emissions from fossil fuel combustion and cement production 5.5 ± 0.5∆ 5.5 ± 0.5 5.5 ± 0.5§
(2) Net emissions from changes in tropical land-use 1.6 ± 1.0∆ 1.6 ± 1.0 1.6 ± 1.0§
(3) Total anthropogenic emissions = (1) + (2) 7.1 ± 1.1 7.1 ± 1.1 7.1 ± 1.1
Partitioning amongst reservoirs

(4) Storage in the atmosphere 3.4 ± 0.2∆ 3.4 ± 0.2 3.3 ± 0.2§
(5) Ocean uptake 2.0 ± 0.8∆ 2.0 ± 0.8 2.0 ± 0.8§
(6) Uptake by Northern Hemisphere forest regrowth not accounted for 0.5 ± 0.5 0.5 ± 0.5§
(7) Other terrestrial sinks = (3)–((4)+(5)+(6))
(CO2 fertilization, nitrogen fertilization, climatic effects) 1.7 ± 1.4 1.4 ± 1.5 1.3 ± 1.5
† Values given in IPCC (1990, 1992).

* Values given in IPCC (1994).
∆ Values used in the carbon cycle models for the calculations presented in IPCC (1994).
§ Values used in the carbon cycle models for the calculations presented here.
42
CHAPTER 3 — HUMAN IMPACTS ON THE CLIMATE SYSTEM
Figure 3.1 — (a) Global annual

emissions of CO2 from fossil fuel
combustion and cement
manufacturing expressed in GtC
year-1 (Rotty and Marland,
1986; Marland, 1989). The
average rate of increase in
emissions between 1960 and
1910 and between 1950 and
1970 is about 4 per cent per year.
Note: the ordinate scale is
logarithmic. (b) Atmospheric
CO2 concentration for the past
250 years as indicated by
measurements in air trapped in
ice from Siple Station, Antarctica
(squares, Neftel et al., 1985;
Friedli et al., 1986), and by
direct atmospheric measurements
at Mauna Loa, Hawaii
(triangles, Keeling et al., 1990).
Note: ppmv means part per
million by volume. [from
Watson et al., 1990, from page
323 in Reference no. 1, with
apportionments for the sources and reservoir dispositions. Fossil fuel combustion
and cement manufacturing account for more than 75 per cent of the total carbon
dioxide input. The remainder (less than 25 per cent) is due to net effects of defor-
estation and other land-clearing operations mainly in tropical areas. Of the total
carbon dioxide input into the atmosphere, almost half remains in the atmos-
phere, an estimated 30 per cent goes into the ocean, seven per cent into Northern
Hemisphere forest regrowth, and the remainder is assumed to go into other parts
of the terrestrial biosphere. Note that the remainder component in the budget
estimates has been reduced between the 1992 and 1995 IPCC reports. The 2001
IPCC report (Reference no. 7) shows that for the decade 1990-1999 the land
ecosystem uptake of CO2 is larger and is now nearly as large as the ocean uptake.
The estimates for the land and ocean were 1.4 and 1.7 GtC/yr respectively.
Two other ‘natural’ greenhouse gases that are increasing due to human activ-
ity are methane and nitrous oxide. As shown in Table 3.2, their relative increases
between pre-industrial times and 1990 are significant especially for methane
which has more than doubled in concentration since preindustrial times. Even
though the absolute values of the concentration of both these gases are less than
one per cent of that of carbon dioxide, the radiative effects of the human-caused
increases in these two gases together amount to nearly 40 per cent of that of
carbon dioxide.
43
Table 3.2 — Characteristics of some key greenhouse gases that are influenced by human activities a.
Parameter CO2 CH4 CFC-11 CFC-12 N2O
Pre-industrial atmospheric concentration 280 ppmvb 0.8 ppmv 0 0 288 ppbvb

(1750–1800)
Current atmospheric concentration (1990)c 353 ppmv 1.72 ppmv 280 pptvb 484 pptv 310 ppbv
Current rate of annual atmospheric accumulation 1.8 ppmv 0.015 ppmv 9.5 pptv 17 pptv 0.8 ppbv
(0.5%) (0.9%) (4%) (4%) (0.25%)
Atmospheric lifetimed (years) (50–200) 10 65 130 150
(from page 320, Reference no. 1).

a. Ozone has not been included in the table because of a lack of precise data.
b. ppmv = parts per million by volume; ppbv = parts per billion by volume; pptv = parts per trillion by volume.
c. The current (1990) concentration have been estimated from an extrapolation of measurements reported for earlier years,
assuming that the recent trends remained approximately constant.
d. For each gas in the table (except CO2), the ‘lifetime’ is defined here as the ratio of the atmospheric content to the total rate of
removal. This time scale also characterizes the rate of adjustment of the atmosphereric concentrations if the emission rates are
changed abruptly. Carbon dioxide is a special case since it has no real sinks, but is merely circulated between various reservoirs
(atmosphere, ocean, biota). The ‘lifetime’ of CO2 given in the table is a rough indication of the time it would take for the CO2
concentration to adjust to changes in the emissions.
3.2.2 The introduction of new greenhouse gases by human activity has had a notice-
NEW GREENHOUSE GASES able impact on the greenhouse effect, accounting for over 10 per cent of the total
human impact on the greenhouse effect. Many different gases have been intro-
duced. They are mainly halocarbons (compounds containing carbon together
with halogens such as chlorine, fluorine, bromine, and iodine) such as chloroflu-
orocarbons (CFCs) and hydrofluorocarbons (HFCs). As stated before, these
compounds were manufactured for use in refrigeration units, foaming agents and
solvents.
The halocarbons are strong greenhouse gases and their lifetimes are possibly
longer than those of the long-lived natural greenhouse gases. In addition, ultra-
violet radiation from the sun can disassociate their molecules and release chlorine
and bromine which will interact with and cause the destruction of stratospheric
ozone. The potential harmful effects were so obvious that international agree-
ments have already been put in place to reduce the production of these gases, e.g.
the 1987 Montreal protocol as part of the Vienna Convention to Protect the
Ozone Layer and its subsequent amendments.
A nearly-complete inventory of human impacts on the greenhouse gases is
presented in Table 3.3 (pages 46–47). This list includes a large variety of halocar-
bons and hydrocarbons. The entire list is presented to emphasize that there are a
large number of gases human impacts and to give some terminology for the halo-
and hydro-carbons. Note the numbers used to identify the halocarbon types in
the first column. The list does not contain data for ozone itself because much of
the change in ozone is not due directly to emissions from human activity but
from subsequent chemical reactions in the atmosphere. The list identifies: the life-
time of each gas; concentration change since the pre-industrial era; current
growth rate; and radiative forcing due to changes since the pre-industrial era. Note
that the concentrations of some of the gases are so small that no concentration is
listed. Several already show reductions in concentration through compliance with
international agreements.
Key information for the relative importance of the gases listed in Table 3.3
for greenhouse effects is given by the radiative forcing shown in the very last
column of the table. With this information alone one can easily sort out the small
subset of gases important for the enhancement of the greenhouse effect. Numbers
are omitted for radiative forcing if they are less than .001 Wm-2. As stated before,
the radiative forcing defines the change in the amount of radiative energy trans-
fer at the tropopause due to the changes in concentrations of the gas in the
atmosphere, in this case since the pre-industrial era. A positive sign means
44
increased energy in the downward direction which would lead to higher temper-
atures at the ground on average.
The gases CFC-11 and CFC-12 are responsible for most of the radiative
forcing due to the new greenhouse gases. They account for a total of 0.20 Wm-2
or 8 per cent of the total radiative forcing (+2.45 Wm-2) caused by human
enhancement of the greenhouse gases. Note from Table 3.2 that the percentage
rate of increase of these two halocarbons was much larger than for the naturally
occurring greenhouse gases before the Montreal Protocol was enacted. All of the
halocarbons together result in a radiative forcing of about 0.27 Wm-2 or 11 per
cent of the total radiative forcing in the early 1990s.
3.3 Fossil fuel combustion and biomass burning are the primary sources of aerosols
ATMOSPHERIC AEROSOL due to human activity. These sources produce both soot (particulate black carbon
ENHANCEMENT aerosols), gaseous sulphur dioxide and nitrogen oxides. The latter two are
3.3.1 partially transformed by chemical processes into sulphate and nitrate aerosols.
TYPES OF AEROSOLS Additional sources for human-produced aerosols include dust from changes in
land use. All of these aerosol products remain primarily in the troposphere (unlike
those produced by major volcanic eruptions) and thus are subject to rapid
removal by precipitation and settling processes. Table 3.4 summarizes source
strength, atmospheric loading, the radiative mass extinction coefficient and
radiative optical depth for the main aerosol constituents for both natural- and
human-produced components.
The mass extinction coefficient is a measure of the effectiveness of radiative
absorption and scattering per unit mass of aerosol, and the optical depth is a
measure of the overall reduction in radiation due to absorption and scattering
while passing through the atmosphere. The optical depth depends on both the
mass extinction coefficient and the total mass of the aerosol in a vertical column
of atmosphere. The optical depth gives a measure of the overall direct radiative
forcing on the atmosphere due to the aerosol.
The human production of sulphate aerosols can be measured directly by
sulphur dioxide emissions from fossil fuel combustion. The dominant source
regions are in the Northern Hemisphere. This emission has increased dramatically
over the past 130 years as shown in Figure 3.2. Concerns about other impacts such
as acid rain have led to a levelling off of production rates in recent years in some
areas such as Europe and north America.
70
Regional Anthropogenic Flux Northern Hemispheric
Asia
60
Europe
Anthropogenic Flux
Sulphur emissions (TgS/yr)
N America
50 USSR
40
30
20
Northern Hemispheric
Figure 3.2 — Natural and fossil Natural Flux
10
fuel combustion sources of SO2 in
the Northern Hemisphere (after
Dignon and Hameed, 1989; 0
1992). [from page 106, 1860 1880 1900 1920 1940 1960 1980
Reference no. 3]. Year
45
Species Lifetime Concentration Current Radiative

(ppbv) growth forcing
Year Uncert. 1992 pre-ind. ppbv/yr Wm-2/ppbv Wm-2
Natural and anthropogenically influenced gases

carbon dioxide CO2 variable 356 000 278 000 1 600 1.8 × 10-5 1.56
methane CH4@ 12.2 25% 1714 700 8 3.7 × 10-4 0.47
nitrous oxide N2O 120 311 275 0.8 3.7 × 10-3 0.14
methyl chloride CH3Cl 1.5 25% ~0.6 ~0.6 ~0 0
methyl bromide CH3Br 1.2 32% 0.010 <0.010 ~0 0
chloroform CHCl3 0.51 300% ~0.012 ~0 0.017
methylene chloride CH2Cl2 0.46 200% ~0.030 ~0 0.03
carbon monoxide CO 0.25 50-150 ~0 $
Gases phased out before 2000 under the Montreal Protocol and its amendments
CFC-11 CCl3F 50 10% 0.268 0 +0.000** 0.22 0.06
CFC-2 CCl2F2 102 0.503 0 +0.007** 0.28 0.14
CFC-113 CCl2FCCIF2 85 0.082 0 0.000** 0.28 0.02
CFC-114 CClF2CCIF2 300 0.020 0 0.32 0.007
CFC-115 CF3FCCIF2 1700 <0.01 0 0.26 <0.003
carbon tetrachloride CCl4 42 0.132 0 -0.0005** 0.10 0.01
methyl chloroform CH3CCl3 4.9 8% 0.135# 0 -0.010** 0.05 0.007
halon-1211 CBrCIF2 20 0.007 0 .00015
halon-1301 CBrF3 65 0.003 0 .0002 0.28
halon-2402 CBrF2CBrF2 20 0.0007 0
Chlorinated hydrocarbons controlled by the Montreal Protocol and its amendments

HCFC-22 CHCIF2 12.1 20% 0.100 0 +0.005** 0.19 0.02
HCFC-123 CF3CHCl2 1.4 25% 0 0.18
HCFC-124 CF3CHCIF 6.1 25% 0 0.19
HCFC-141b CH3CFCl2 9.4 25% 0.002 0 0.001** 0.14
HCFC-142b CH3CF2Cl 18.4 25% 0.006 0 0.001** 0.18
HCFC-225ca C3HF5Cl2 2.1 35% 0 0.24
HCFC-225cb C3HF5Cl2 6.2 35% 0 0.28
Perfluorinated compounds
sulphur hexafluoride SF6 3200 0.032 0 +0.0002 0.64 0.002
perfluoromethane CF4 50000 0.070 0 +0.0012 0.10 0.007
perfluoroethane C2F6 10000 0.004 0 0.23
perfluoropropane C3F8 2600 0 0.24
perfluorobutane C4F10 2600 0 0.31
perfluoropentane C5F12 4100 0 0.39
perfluorohexane C6F14 3200 0 0.46
perfluorocyclobutane c-C4F8 3200 0 0.32
Anthropogenic greenhouse gases not regulated (proposed or in use)

HFC-23 CHF3 264 45% 0.18
HFC-32 CH2F2 5.6 25% 0.11
HFC-41 CH3F 3.7 0.02
HFC-43-10mee C5H2F10 17.1 35% 0.35
HFC-125 C2HF5 32.6 35% 0.20
HFC-134 CF2HCF2H 10.6 200% 0.18
HFC-134a CH2FCF3 14.6 20% 0.17
HFC-143 CF2HCH2F 3.8 50% 0.11
HFC-143a CH3CF3 48.3 35% 0.14
HFC-152a CH3CHF2 1.5 25% 0.11
HFC-227ea C3HF7 36.5 20% 0.26
HFC-236fa C3H2F6 209 50% 0.24
HFC-245ca C3H3F5 6.6 35% 0.20
HFOC-125e CF3OCHF2 82 300%
HFOC-134e CHF2OCHF2 8 300%
trifluoroiodomethane CHF3I <0.005 0.38
46
Table 3.3 — Lifetimes for Notes:

radiatively active gases and This table lists only the direct radiative forcing from emitted gases. The indirect effects due
halocarbons. [from pages 92 to subsequent changes in atmospheric chemistry, notably ozone (see below), are not
and 93, Reference no. 3]. included. The Wm-2 column refers to the radiative forcing since the pre-industrial, and the
Wm-2/ppbv column is accurate only for small changes about the current atmospheric
composition (see Section 2.4 of Reference no. 3 and IPCC, 1994). In particular, CO2, CH4
and N2O concentration changes since pre-industrial times are too large to assume linear-
ity; the formulae reported in IPCC (1990) are used to evaluate their total contribution.
A blank entry indicates that a value is not available. Uncertainties for many lifetimes have
not been evaluated. The concentration of some anthropogenic gases are small and difficult
to measure. The pre-industrial concentrations of some gases with natural sources are diffi-
cult to determine. Radiative forcings are only given for those gases with values greater than
0.001 Wm-2.
@ Methane increases are calculated to cause increases in tropospheric ozone and stratos-
pheric H2O; these indirect effects, about 25 per cent of the direct effect, are not
included in the radiative forcings given here.
$ The direct radiative forcing due to changes in the CO concentration is unlikely to reach
a few hundredths of a Wm-2. The direct radiative forcing is hard to quantify.
** Gases with rapidly changing growth rates over the past decade, recent trends since
1992 are reported.
# The change in CH3CCl3 concentration is due to the recalibration of the absolute stan-
dards used to measure this gas.
Stratospheric ozone depletion due to halocarbons is about -2 per cent (globally) over the period
Table 3.4 — Source strength, 1979 to 1990 with half as much again occurring both immediately before and since; the total
atmospheric burden, extinction radiative forcing is thus now about –0.1 Wm-2. Tropospheric ozone appears to have increased
efficiency and optical depth due since the 19th Century over the northern mid-latitudes where few observational records are
to the various types of aerosol available; if over the entire Northern Hemisphere, tropospheric ozone increased from 25 ppb to 50
particles (after IPCC, 1994; ppb at present, then the radiative forcing is about +0.4 Wm-2.
Andreae, 1995; and Cook and
Wilson, 1996). [from page 104
in Reference no. 3].
Source Flux Global mean Mass extinction Global mean

(Tg/yr) column burden coefficient optical depth
(mg m-2) (hydrated) (m2g-1)
Natural
Primary
Soil dust (mineral aeorosol) 1500 32.2 0.7 0.023
Sea salt 1300 7.0 0.4 0.003
Volcanic dust 33 0.7 2.0 0.001
Biological debris 50 1.1 2.0 0.002
Secondary
Sulphates from natural precursors, as (NH4)2SO4) 102 2.8 5.1 0.014
Organic matter from biogenic VOC 55 2.1 5.1 0.011
Nitrates from NOx 22 0.5 2.0 0.001
Anthropogenic
Primary
Industrial dust, etc. 100 2.1 2.0 0.004
Soot (elemental carbon) from fossil fuels 8 0.2 10.0 0.002
Soot from biomass combustion 5 0.1 10.0 0.001
Secondary
Sulphates from SO4 as (NH4)2SO4 140 3.8 5.1 0.019
Biomass burning 80 3.4 5.1 0.017
Nitrates from NOx 36 0.8 2.0 0.002
47
3.3.2 Aerosols have a pronounced effect on solar radiation transfers in the atmosphere
RADIATIVE IMPACTS and a smaller one on terrestrial radiation transfers in the atmosphere. They cause
3.3.2.1 scattering and absorption of solar radiation with the result that less solar energy
Direct impacts gets to the Earth’s surface, more is absorbed in the atmosphere, and more is
backscattered to space. Only for the case of soot (black carbons) is there such
strong absorption of solar radiation in the lower atmosphere that solar radiation
backscattered to space is reduced, producing a positive radiative forcing. Aerosols
also enhance absorption and emission of terrestrial radiation and thus slightly
enhance the greenhouse effect. However, the effect on solar radiation is generally
the dominating factor so that the net aerosol effect is estimated to be opposite to
the greenhouse effect.
Since the lifetime of tropospheric aerosols is short, their concentration is not
uniform over the globe; they are concentrated over the land areas. A model simulation
for lower tropospheric sulphate aerosol concentrations in the 1980s is shown in Figure
3.3. (In Europe, the current values are much lower than in the 1980s). Figure 3.3
highlights the areas where another environmental effect of sulphates has occurred, the
acidification of rainwater leading to ‘acid rain.’ Acid rain became a serious problem in
places such as eastern north America, Europe and eastern Asia. The short lifetime of
aerosols implies that if the human sources were turned off, the human-produced aerosol
concentrations would decrease very fast, with a time scale of about a week.
In summary the overall direct radiative forcing from human-produced aerosols is
negative, estimated as -0.5 Wm-2. Estimates (from numerical models and not
observations) for the radiative forcing for the three main components, sulphate aerosols,
fossil fuel soot, and biomass burning are, -0.4 Wm-2, +0.1 Wm-2 and -0.2 Wm-2,
respectively. For comparison, the estimated peak values for radiative forcing of
stratospheric aerosols put there by volcanic activity at times during the last 130 years
have nearly ten times the current negative radiative forcing due to tropospheric aerosols
from human activity, as shown in Figure 3.4. The last major peak of roughly -3.5 Wm-2
was in 1991 due to the eruption of Mt. Pinatubo. However, volcano-produced negative
forcing lasts only for a few years and it occurs only sporadically. Over multi-year periods,
the mean volcanic forcing is comparable to anthropogenic aerosol.
3.3.2.2 Assessment of the overall impact of human aerosol enhancement is complicated

Indirect impacts by transformations that occur due to chemical processes with other gases and
with the water droplets in clouds. These transformations result in ‘indirect’
impacts. The impact on clouds and their optical properties is discussed below.
3.4 Since aerosol particles act as cloud condensation nuclei and freezing nuclei, they
CHANGE OF RADIATIVE have important impacts on the radiative properties of clouds. Aerosols may
EFFECTS OF CLOUDS change the drop-size distribution in clouds which can influence the optical prop-
erties. The magnitude of this effect depends on the properties of the aerosols
present which can vary from region to region. Nucleation leading to an increase
Figure 3.3 — Model simulation of

the annual mean sulphate (SO42-)
concentration at 900 mb.
Contours are shown at 25, 50,
100, 250, 500, 1000, and 2500
pptv. Concentrations over eastern
North America and eastern
Europe exceeded natural (non-
anthropogenic) levels by a factor
of 10. [from Langner and
Rodhe, 1991, from page 328,
48
Figure 3.4 — Variation of global

mean visible optical depth, and the
consequent radiative forcing
(Wm-2) resulting from
stratospheric aerosols of volcanic
origin from 1850 to 1993, as
estimated by Sata et al., 1993.
The radiative forcing has been
estimated using the simple
relationship given in Lacis et al.,
1992, where the radiative forcing is
-30 times the visible optical depth.
Reference no. 3].
in cloudiness and/or the prolonging of existing clouds are other possible influ-
ences of aerosols on clouds. Effects are believed to be quite different between ice
clouds and water clouds. The increase in high-level cloudiness due to jet aircraft
contrails is a specific example of increase in cloudiness due to the introduction of
aerosols and water vapour.
Additional observations and theoretical studies are needed before the
changes in the optical properties of clouds can be quantified. At the present time,
the IPCC assessment judges that, overall, the altered properties of clouds due to
aerosols produced by human activity will give a negative radiative forcing on the
global scale. Estimates of magnitudes range as high as -1.50 Wm-2 which would
be much larger than the direct aerosol impact and could offset the positive forcing
attributable to greenhouse gases by nearly a half.
3.5 It is recognized that humans have greatly altered the Earth’s land surface by
CHANGE OF RADIATIVE settlements, deforestation and agriculture. These changes have resulted in both
PROPERTIES OF THE LAND increases and decreases in local albedo values. Replacing forests with crops that
SURFACE are present only part of the year would likely increase albedo effects. Urbanization
could either increase or decrease the albedo depending on the reduction in trees
and on the materials used for roofs and streets.
These albedo changes clearly have a role in the local energy balance and
contribute to microscale climate changes such as the heat ‘islands’ over urban
regions. However, other factors such as evaporation, precipitation and wind flow
changes could have large impacts on local or even regional climate change. On a
global basis the change in the radiative properties of land surfaces is not consid-
ered to be a major factor in the energy balance.
3.6 Figure 3.5 provides an overall summary of the relative importance of the many
SUMMARY OF HUMAN ways in which human activity can alter radiative energy transfers in the atmos-
IMPACTS phere. Confidence in the quantitative values is indicated at the bottom. The
depiction includes ozone effects and, for comparison, solar variability effects due
to the sun spot cycle. Note that increases in tropospheric ozone are estimated to
give a radiative forcing value of +0.4 Wm-2 which is equal and opposite to the
forcing due to tropospheric sulphate increases. The cloud change impacts due to
aerosols are shown in terms of the range of estimated values of radiative forcing.
Further understanding is needed to establish a mid-range estimate as is done for
all the other radiative forcing components. No entry is made for volcanic effects,
which are quite variable in time, or for the impacts of changes in surface albedo
as these effects are considered important primarily on a regional scale. Much
research is needed to reduce these uncertainties especially with respect to the
indirect aerosol effect on clouds.
49
The global mean radiative forcing of the climate system

for the year 2000, relative to 1750
3
Halocarbons
Radiative forcing (Watts per square metre)
2 N2O Aerosols
CH4
Black
carbon from
Warming
CO2 fossil
1 Tropospheric Mineral Aviation-induced
fuel
ozone burning Dust Solar
Contrails Cirrus
0
Stratospheric Organic
ozone carbon Biomass Land-
Cooling
use
−1 Sulphate from burning Aerosol (albedo)
fossil indirect
fuel effect only
burning
−2
High Medium Medium Low Very Very Very Very Very Very Very Very
Low Low Low Low Low Low Low Low
Level of Scientific Understanding
Figure 3.5 — Global, annual mean radiative forcings (Wm-2) due to a number of agents for the period from pre-industrial
(1750) to present (late 1990s; ~2000). The height of the rectangular bar denotes a central or best-estimate value while its
absence denotes no best estimate is possible. The vertical lines capped with horizontal lines indicate an estimate of the
uncertainty range, for the most part guided by the spread in the published values of the forcing. The uncertainty range
specified here has no statistical basis. At the bottom a ‘level of scientific understanding’ index is accorded to each forcing.
This represents a subjective judgement about the reliability of the forcing estimate, involving factors such as the assumptions
necessary to evaluate the forcing, the degree of knowledge of the physical/chemical mechanisms determining the forcing, and
the uncertainties surrounding the quantitative estimate of the forcing. The well-mixed greenhouse gases are grouped together
into a single rectangular bar with the individual mean contributions due to CO2, CH4, N2O and halocarbons. The sign of the
effects due to mineral dust is itself an uncertainty. The forcing due to stratospheric aerosols from volcanic eruptions is highly
variable over the period and is not considered for this plot. All the forcings shown have distinct spatial and seasonal features
such that the global, annual-means do not yield a complete picture of the radiative perturbations. They are intended to give a
first-order perspective on a global, annual-mean scale and cannot be readily employed to obtain the total response to
forcings. It is emphasized that the positive and negative forcings cannot be added up and viewed a priori as providing offsets
in terms of the complete global climate impact. [from page 8, Reference no. 7].
50
CHAPTER 4
MODELLING CLIMATE CHANGE
4.1 This chapter has two basic purposes. The first is to summarize briefly some of the
INTRODUCTION fundamentals for numerical modelling of climate. This provides literacy for termi-
nology used in the discussion of climate models and some understanding of the
numerical modelling approach including its strengths and weaknesses. The
second is to present assessments of current climate model performance. This is
very important as most projections for future climate change are based on numer-
ical model solutions.
Mathematical simulation of the climate system is essential for understanding
climate change outcomes from human impacts such as those described in the last
chapter and for making estimates of future climate change. The climate system is
very complex with many interacting components and a large number of variables
and processes that need to be represented. A quantified description of the system
and its changes requires obtaining solutions to a large number of governing
mathematical equations.
The physical component systems are continuous in space and the scales of
phenomena that exist range from global to molecular. The governing equations
include partial differential and integral relationships. Many processes are not fully
described in the equations such as turbulence in the atmosphere and ocean,
precipitation growth in clouds, cumulus convection, radiation transfer in and
around clouds, and CO2 transfer processes in heterogeneous biosphere canopies.
The complexity of the system precludes obtaining general solutions by analytical
mathematical methods.
In order to obtain solutions, it is necessary to make approximations in both
the governing equations and the numerical methods. Approximations in the
equations are guided by understanding the relative importance of various
processes represented in the equations. Computer capacity limits the spatial reso-
lution available for the physical system in the numerical model. A climate model
must represent the entire globe, and the resolution limitation means that there is
a lower limit on the size of scales that can be explicitly represented in a model.
Current computer capacities limit overall atmospheric resolution to the order of
100 km in the horizontal for full climate system models. The effects of all smaller
scales must be parameterized, i.e. represented in terms of conditions which are
resolved by the model.
The most effective climate modelling work utilizes a hierarchy of models
including those which are simplified more than required by the computer
system. The simpler models make it easier to isolate physical processes in the
climate system and to give a general indication for overall impacts of climate
change forcing. The complete models provide the best quantitative estimates of
climate change on a detailed regional basis. Simple, one-dimensional or volume-
integrated models of the atmosphere with prescribed ocean, cryosphere,
biosphere, and land surface specifications, or with an interactive ocean using
simple diffusion, radiation and convection, and energy balance relationships are
important. They provide useful first indications of atmospheric changes due to
human enhancements of radiatively-active constituent concentrations in the
atmosphere. The model hierarchy also includes models that represent only a part
of the climate system to give more details on that part. Examples are regional
models which use information from the global models as external forcing
conditions.
It is essential that numerical climate models be calibrated by comparing solu-
tions with observational information. This should be done for as wide a range of
variables in as many locations as possible. A numerical model may give a reason-
able solution for surface atmospheric temperature, sea-surface temperature, or
51
sea-ice coverage while at the same time its solution for precipitation over a given
area can be very poor.
Obtaining and analysing the relevant observational information is a consid-
erable undertaking. Paleoclimatologists have made great progress in meeting the
challenge to make analyses of conditions from geological evidence for compari-
son with model predictions of past climates. For current-day analyses,
observational data from many different sources are merged together with internal
consistency provided by numerical model solutions. This process, termed data
assimilation, is used by many operational weather prediction groups in the world
today. As a result, the observational information used to compare with the model
prediction solutions may not be fully independent of the model.
The basic modelling strategy for climate change prediction has the following
key components. The first step is to understand the quality of model performance
based on comparison of model solutions with observations of existing conditions
for a variety of situations. The second step is to obtain model solutions where
anticipated future forcing of the climate system due to human activity is intro-
duced into the model. The final step is to interpret, evaluate, and establish the
level of uncertainty for the model predictions based on the understanding of
model performance.
4.2 The equations used to define the climate model are those associated with each of
BASICS FOR MODELING the five components of the climate system plus additional relationships for the
4.2.1 interactions between the components. Needless to say, the list of governing equa-
GOVERNING PHYSICAL tions is extensive. Formulation of basic equations is still under way particularly
EQUATIONS for the cryosphere, biosphere and land-surface components.
Many of the equations for the atmospheric and oceanic components have
been published in textbooks including the basic references listed in the introduc-
tion to these lecture notes. The governing equations include those which describe
mass continuity, Newton’s laws of motion for a fluid (with the vertical component
equation replaced by the hydrostatic relationship), the equation of state, first law
of thermodynamics, radiation transfer and, for the atmosphere, the hydrological
cycle. The equation of state for the atmosphere is that for a compressible gas
including water vapour effects. The equation of state for the ocean is that for an
incompressible fluid including salinity effects.
Equations for chemistry must be included with the atmospheric equations.
Chemistry is very important as it describes the concentrations, life cycles, and
interactions involving the greenhouse gases and aerosols, key forcing agents of
the climate system. An example of the important areas and processes involving
chemistry relationships in the climate system for water vapour is shown in Figure
4.1. This figure outlines the many chemical and physical interactions in the
climate system. No attempt is made to summarize here the governing equations
for all the chemical processes.
Equations for the land surface component in the climate system include
much more detail for variations in surface composition than are used in atmos-
pheric weather prediction models. Terrestrial biosphere conditions are closely tied
to the overall land surface condition. The low heat capacity of the surface means
that small changes in albedo can have large impacts on surface temperature. Thus,
governing equations are needed for ground cover canopy types and coverages,
and for the resultant albedo. Equations to describe details for hydrological
processes are essential both for the condition of the surface biosphere and for
evaporation rates into the atmosphere. Key parameters for discussion of the
Earth’s surface biosphere properties are the ‘resistances’ (resistance to transfer) for
water vapour and carbon dioxide. Comprehensive data on land cover types and
the details of vertical structure in the terrestrial biosphere are still needed to cali-
brate a full set of equations for the land surface.
A land surface model must represent and quantify an intricate array of
processes to define the key surface parameters (albedo, roughness length and soil
moisture) which can lead to large-scale climate system feedbacks. There are
numerous physical interactions related to each of these quantities from which it
52
CHAPTER 4 — MODELING CLIMATE CHANGE

of the physical and chemical
interactions involving water
vapour that might be included in
a comprehensive global climate
model. The key physical climate
parameters are shown in boxes.
Atmospheric chemical species are
enclosed in elipses. Chemical
mechanisms that may involve
water vapour are divided into
classes, identified at the bottom
of the figure, and the pathways
and modes of interactions are
indicated by arrows.
‘Photo’ encompasses those
chemical processes driven by
solar radiation or involving the
reaction of species produced when
ambient air is exposed to solar
ultraviolet radiation;
‘hetero’ refers to chemical
processes occurring in aqueous
solutions, principally in cloud
droplets, or on particle surfaces, is possible to argue both positive and negative feedbacks. It remains a challenge
particularly solid aerosols and ice to develop a land surface model that can resolve quantitatively the interaction
crystals; impacts.
‘anthropo’ to emissions and Representation of marine biochemical processes is important to define the
processes associated mainly with gas exchanges with the atmosphere for the radiatively-active gases such as carbon
human activities; dioxide, nitrous oxide, and dimethyl sulphide. Geological data document varia-
‘bio’ refers to processes related to tions in the marine biosphere associated with carbon dioxide variations that have
the assimilation and respiration accompanied large climate changes in the past. There is a wide range of processes,
of atmospheric constituents by including biosphere interactions, associated with the carbon cycle in the oceans.
living organisms; and Four forms of carbon are involved: gaseous carbon dioxide, its dissolved counter-
‘geo’ refers to chemical processes part (Dissolved Inorganic Carbon, DIC), solid organic compounds and solid
occurring on surfaces or in media inorganic compounds (carbonates). Ocean temperature and motions, and solar
at the interface between the radiation penetrating below the ocean surface are important factors in the oceanic
atmosphere and the oceans and carbon cycle.
land. [from page 213, The cryosphere has several components that must be represented by
Reference no. 6, with geophysical governing equations. For glaciers and ice domes on land, there are the
permission of Cambridge equations for motion that can be simplified because acceleration terms can be
University Press]. neglected. The motion of glaciers and ice fields is an important factor in deter-
mining coverage and total mass content for ice. Hydrodynamic equations are
needed to describe the motion of sea ice as it responds to ocean currents,
surrounding ice and atmospheric winds. There is an important instability hypo-
thesis (positive feedback mechanism) for the west Antarctic ice sheet that could
occur with an increase in sea-surface temperature. In this hypothesis, warming
temperatures increase the slippage of ice into the ocean which raises sea level thus
increasing further the slippage of ice into the ocean. It will be a challenge to repre-
sent this with deterministic equations.
The linkage between the ocean and atmosphere is central for the climate
model. It is necessary to accurately represent sensible and latent energy trans-
ports. The sensible energy transport depends on turbulent boundary layer
processes in both the ocean and atmosphere. Latent energy (water vapour) trans-
port depends on the boundary layer processes only in the atmosphere. In
addition, atmospheric forcing of ocean currents must be represented. This
requires proper specification of wind stress at the surface — another aspect of the
atmospheric boundary layer. Radiation transfer must take into account that solar
radiation penetrates into the ocean, thereby distributing the heating effect into
53
the upper layer of the ocean. The sensible and latent energy transport specifica-
tions involve approximations because of parameterization for turbulence in the
boundary layer. Because of such approximations in both the atmosphere and
ocean, special attention must be given to make sure that the vertical energy trans-
ports are the same on both sides of the oceanic-atmospheric interface.
4.2.2 The last section gave a brief overview of the large number of mathematical rela-
PARAMETERIZATION OF tionships needed for a complete climate model. It was noted that some of the
PHYSICAL PROCESSES relationships have yet to be developed. In the equations there are numerous
processes, such as turbulence, that cannot be described because of limitations in
the numerical model resolution. These processes are described as subgrid-scale
processes. Parameterization is the representation of the effects of the subgrid-scale
processes on the model-resolved conditions. It is based on the model variables
and specified proportionality constants.
The best spatial resolution attained in current climate models is on the order
of 100 km in the horizontal direction and up to 40 layers in the vertical direction
in both the ocean and atmospheric components. This means that many phenom-
ena which have important influences on the climate system cannot be fully
resolved explicitly and must be parameterized. Examples of these include:
• Radiation (both solar and terrestrial);
• Cumulus convection, including thunderstorms;
• Oceanic convection;
• Surface layer (including plant canopies) momentum, water vapour and heat
transfers;
• Planetary boundary layer momentum, water vapour and heat transfers;
• Turbulence in the free atmosphere and ocean;
• Precipitation processes;
• Atmospheric gravity waves over mountains and in the stratosphere;
• Weather fronts;
• Oceanic jet streams;
• Sea-ice thickness changes; and,
• Marine biosphere processes.
Improvement of parameterization is an essential part of improving climate
models. Parameterization schemes may be interrelated so that changing one
scheme may require simultaneously changing another to get overall improve-
ment in the model. An example is the interconnection between cumulus
convection parameterization and planetary boundary layer parameterization.
Changes in parameterization can have significant effects on the simulated
climate.
4.2.3 The mathematics required for solving the equations for the current comprehen-
MATHEMATICS sive climate system models are quite complicated in practice. Thus, many
scientists as well as other persons do not have a working knowledge of such
details and must rely upon specialists. Much of the modelling complexity arises
from dealing with the atmospheric and oceanic components of the climate
system. As noted before, the atmospheric and oceanic general circulation equa-
tions include a large number of nonlinear, partial differential equations including
integrals and do not have exact analytical solutions.
Different numerical approximation approaches have been used to represent
the partial differential equations as a finite set of algebraic equations that can be
handled by the computer. These approaches include finite-difference, spectral and
finite-element methods. Basic variables of the system such as pressure, temperature,
wind, water vapour and radiation are prescribed by a discrete and finite set of
numbers in the spatial and temporal domain of the climate system. These
numbers represent values of the variables in different ways. In the finite-differ-
ence method these number are grid point values for basic variables of the system.
In the spectral and finite-element methods the numbers are the amplitudes
(transform values) of specified continuous functions which when added together
describe the spatial variations of the basic variables. For each method of
54
representation appropriate algebraic expressions are derived for the spatial and
temporal partial derivatives and integral functions in the governing equations.
The ‘resolution’ of the model is an indicator of the number of grid point
positions or specified continuous functions that are used in the model. Resolution
increases as the number of the positions or functions used is increased. It is
expected that the accuracy of the solution will be improved as the resolution is
increased.
The finite-difference method is the easiest to understand and to use. Take, for
example, the algebraic expression used for the first derivative. The first derivative
(gradient) for a variable f in the x-direction at position x may be expressed as the
difference in value between f at the grid point position on the plus side of x (position
x+1) and f at the grid point position on the minus side of x (position x-1) divided by
the distance between these two points. For an integral over a given spatial range the
finite difference expression could be a summation of products of all f grid point values
within this range, each multiplied by the increment of distance for which that value
is representative. There are many choices for the finite-difference formulation
depending on the accuracy desired in the approximation.
Of the other two methods, the spectral method is more commonly used in
climate models. The concept of this approximation and definition of its resolu-
tion are illustrated in Figure 4.2. The thick line with right-angle bends, an
example of a physical variable functional relationship, is represented by sine and
cosine functions of appropriate amplitudes (transform values) so that the function
obtained by adding them together (represented by thin lines) is as close as possi-
ble to the original function (thick line). Results are improved if one increases the
resolution by increasing the number (K) of the sine and cosine functions added
together. Figure 4.2 shows results using 1, 2, 3, and 5 functions, respectively.
If the square function in Figure 4.2 were considered representative of a sharp
mountain on earth, note that the spectral approximation will give regions where
the ground level actually is lower than anywhere in the original representation.
Figure 4.2 — Spectral

approximations (thin lines) to the
top-hat function (thick lines) for
several different wave number
truncations. [from page 296,
University Press].
55
In all fairness, it should be noted that the finite difference method will also have
an obvious deficiency in this case. Namely, at each of the positions 0 and 0.5 on
the x-axis in Figure 4.2 only one value can be used for a grid point value so that
the vertical lines must be replaced by (steeply) sloping lines.
The mathematics for the solution of the spectral method equations are very
complex. Originally it was not even possible to use them for accurate representa-
tion of nonlinear or local phenomena because of extensive computer
requirements. However, new mathematical procedures such as the ‘fast Fourier
transform’ and the combination of the finite-difference method with the spectral
method have made it possible to incorporate the spectral method into the
comprehensive climate models. The ‘fast Fourier transform’ requires a grid point
array with grid point spacing specified according to the number of spectral
components in the model. This grid is called the transform, Gaussian, or equiva-
lent grid. Many climate models incorporate a mixture of the finite difference and
spectral formulations. In these cases the resolution may be described in terms of
either the number of spectral components or the spacing of grid points in the
equivalent grid.
There are a number of mathematical properties of the numerical model
equations that must be considered to produce a satisfactory climate model.
Numerical instability is the one feature that needs to be avoided at all cost. This
instability can result in very rapid and unrealistic increases in magnitudes of the
model variables which render the solution meaningless. This instability arises
from the mathematical formulation itself. Methods to prevent it include
decreasing the size of the time steps, adopting finite difference equation formu-
lations that have special energy-conserving properties, and adding smoothing
properties to the equation system to reduce amplitudes of the solution compo-
nents most likely to be unstable.
It is desirable to have mathematical formulations so that the model solutions
have similar properties to those of the partial differential equations. A basic
feature is the constancy of total mass and total energy expected if the only
physical process operating is transport (advection due to fluid motion) within the
model domain. Also, it is known that transport due to a uniform flow field will
move another dependent variable feature with the velocity of the uniform flow
and without distortion of the structure (shape and peak magnitude value) of the
other dependent variable. Numerical model errors with respect to these two char-
acteristics are referred to as ‘phase’ (position) and ‘amplitude’ errors.
4.2.4 Electronic computing machines which included stored-programme technology

COMPUTERS came into being in the late 1940s. This was followed by a rapid growth in the
capability of single-processor machines starting in the early 1950s and still
continuing today. Application to weather prediction was one motivating factor in
this development. The first operational numerical weather prediction model
began operation in 1955. Speed of single-processor machines, one measure of
computer power, has increased by many orders of magnitude from the first elec-
tronic machines to the present time. Figure 4.3 shows the increase in electronic
computer speeds from the mid-1950s to 1990. The pace of increase in computer
speed has increased even more since 1990 and current (1998) single-processor
machines have reached sustained speeds as high as 600 million operations per
second (600 megaflops) [Hammond, 1998, private communication].
Advances in computer power continued in the 1980s with the development
of parallel-processor vector machines wherein single-processor machines were
combined to work on parts of the same problem simultaneously. This intercon-
nection capability has made it possible to achieve computer system speeds as high
as 20 billion operations per second by 1998. Projects are now under way in the
United States and Japan to increase computer system peak speeds to 32 trillion
operations per second (32 teraflops) by the year 2004 [private communication H.
Grassl,1998 with strategy document for ECMWF 1999-2008]. At the same time
communications and networking capabilities have led to systems that may be
physically separate but can combine to give large computer capability.
56
Figure 4.3 — Trend in single

processor computational
performance up to 1990 (after
Worlton, 1987, personal
communication). [from page
288 , Reference no. 6, with
University Press].
General circulation modelling became an important use of computer power

by 1960. Expansion of such models to climate models and the subsequent growth
of climate modelling work has continued the pressure to expand computer power
even further.
4.3 Both simple and comprehensive numerical climate models have been used exten-
CURRENT CLIMATE sively in climate change research. Simple climate models based on (oceanic)
MODELS AND THEIR Upwelling Diffusion and Energy Balance (UD/EB) have played a key role in sensi-
PERFORMANCE tivity studies for climate change. These UD/EB models represent the atmosphere
4.3.1 and ocean by several large domains respectively. Mean temperatures are found
INTRODUCTION which balance the energy transfers between the domains for a prescribed radia-
tive forcing in the atmosphere. Energy transfers within the ocean domains are
represented by simple upwelling and diffusion processes. Descriptions of the
UD/EB model used in climate change studies appear in publications by Wigley
and Raper (1987, 1992).
It has been possible to get solutions for global mean temperatures in the
UD/EB model that correspond to comprehensive climate model results by adjust-
ing the structure and parameter values in the UD/EB model. Since the UD/EB
models have much smaller computer requirements than the comprehensive
57
models, it has been possible to perform a large number of sensitivity experiments

to quantify the range in global temperature changes due to uncertainties in radia-
tive forcing effects and future concentrations of aerosols and greenhouse gases.
These results have been comparable to the results for global mean temperatures
from the comprehensive climate models.
Comprehensive climate models containing representation of both the ocean
and atmosphere have been developed by over two dozen research groups worldwide
for climate change research and assessments. These models differ in details of model
formulation including resolution and parameterization. The representation of the
ocean is comprehensive including currents, temperature and salinity distributions,
sea ice, turbulent mixing processes, radiation transfer and bottom topography.
Systematic efforts have been made to compare model simulations with current and
past observed conditions and with each other. Such model intercomparisons help
us to isolate and understand the impact of different model approaches and to docu-
ment the range of uncertainty in climate model simulations.
Three kinds of model evaluation have been made. The first is an evaluation
of the overall full climate model where the components are interactive. The
second approach is to examine individual components of the climate model. The
third is to look at the sensitivity of the model results to its formulation, boundary
conditions and parameterizations. These three approaches are presented in the
next three sections. The final part of the section on model sensitivity summarizes
the model factors which are currently believed to be most responsible for the
uncertainty in climate model simulations.
4.3.2 Sixteen comprehensive climate models presented in the 1995 IPCC report
OVERALL CLIMATE MODEL included full two-way interactions between the atmospheric and oceanic compo-
EVALUATION nents have provided a basis for overall climate model evaluation and
4.3.2.1 intercomparison. These evaluations have been based on the climate produced by
Current climate conditions the models, all with the same forcing parameters corresponding to current condi-
tions. Extended time simulations ranging in length from 100 to 1000 years have
been made to establish what was considered to be equilibrium climate conditions.
Model results have been compared to observations and to each other.
The sixteen groups are listed in Table 4.1 along with a few key descriptors for
each model such as:
(a) Resolution of the atmospheric general circulation model component (AGCM).
Terminology for horizontal resolution is given by the latitude-longitude grid
point spacing for the grid point models and with numbers representing the
number of spectral components for the spectral models (R and T refer to rhom-
boidal and triangular, respectively, descriptors for the details of the spectral
components). The vertical resolution is given by the number of layers written
after ‘L.’ Note that the model summary given in the 2001 IPCC report (Reference
no. 7) shows improved resolution for many of these models.
(b) Resolution of the Ocean General Circulation Model Component (OGCM).
Terminology is the same as for the atmospheric component. (Note: only grid
point formulations are used since the spectral approach is not practical for a
domain broken up by continents.).
(c) Descriptions for the sea-ice and land-surface components.
(d) Indication if ‘flux corrections’ are used at the ocean-atmosphere interface.
The use of flux corrections has been necessary for some of the models to offset
discrepancies in the vertical momentum, heat, and water vapour fluxes at the
ocean-atmosphere interface which could cause an unnatural drift away from
observed climate conditions. It is anticipated that the need for flux corrections
will decrease as models are further improved.
A number of variables from these climate simulations are compared with
observations and with each other in the group of the 11 climate models that had
completed the simulation at the time of the 1995 IPCC report. The variables are
surface air temperature, precipitation, Northern Hemisphere snow cover, sea-ice
cover, mean sea-level pressure, surface heat flux over the ocean, and strength of
the north Atlantic ocean thermohaline circulation. Comparisons are made in
58
Table 4.1 — Coupled model control simulation for current climate. [from page 236, Reference no. 3].
Group Model No. Country AGCM OGCM Sea Ice@ Flux Land-surface$ Initial Notes§
Resolution Resolution correction+ scheme state^
BMRC 1 Australia R21 L9 3.2° × 5.6° L12 T none B E d

CCC 2 Canada T32 L10 1.8° × 1.8° L29 T H, W, T BB E d
CERFACS 3 France T42 L31 1° × 2° L20 none B E
COLA 4 USA R15 L9 3° × 3° L16 T none Crs (SSiB) d
CSIRO 5 Australia R21 L9 3.2° × 5.6° L12 T/R H, W, τ, T Crs (CSIRO) E d
GFDL 6 USA R30 L14 2° × 2° L18 T/Dr H, W B E d
GISS 7 USA 4° × 5° L9 4° × 5° L13 T none C I d (1)
GISS 8 USA 4° × 5° L9 4° × 5° L16 T none C E (2)
IAP 9 China 4° × 5° L2 4° × 5° L20 T H, W B E NPOGA
LMD/OPA 10 France 3.6° × 2.4° L15 1° × 2° L20 T none B (SECHIBA)
MPI 11 Germany T21 L19 5.6° × 5.6° L11 T H, W, τ, T B E d E1/LSG
MPI 12 Germany T21 L19 2.8° × 2.8° L9 T/R H, W, τ, T Crs (ECHAM) E d
E2/OPYC
MRI 13 Japan 4° × 5° L15 (0.5–2°) × 2.5° L21 T/Dr H, W B E d
NCAR 14 USA R15 L9 1° × 1° L20 T/R none B I d
UCLA 15 USA 4° × 5° L9 1° × 1° L15 none fixed wetness NPOGA
UKMO 16 UK 2.5° × 3.8° L19 2.5° × 3.8° L20 T/Dr H, W Crs (UKMO) U d
@ T refers to ‘thermodynamic’ and R to ‘dynamic’ sea ice with rheology. Dr stands for ‘free drift’ sea ice.
+ H, W, τ, T stand for flux adjustment of heat, fresh water, surface stress and ocean surface temperture, respectively.
$ B refers to a ‘simple bucket”; BB refers to ‘modified bucket’; C includes canopy processes; rs denotes inclusion of stomatal resistence. The name of the land-surface scheme
is given in parentheses.
^ The method of initialising the coupled model is indicated by E for an equilibrium of the coupled system, U for equilibrium of the upper ocean and I for initial conditions specified from
available observations.
§ NPOGA: stands for ‘no polar ocean with global atmosphere’ and reflects models which either exclude the polar oceans and ice by specifying climatological values or which
control polar deep ocean quantities with a relaxation to observed values. E1 and E2 refer to the ECHAM1 and ECHAM2 AGDMs, (1) and (2) refer to two versions of the GISS
model, and LSG and OPYC refer to ocean models. d indicates that data from this stimulation are included in Reference no. 3.
59
terms of global means or global coverage, spatial distributions, and zonal aver-
ages. Mean values were shown for all variables. Also examined is surface
temperature variability on monthly, seasonal, interannual and decadal time
scales.
Some comparisons of these climate descriptors are shown in the following
table and two figures with information on observed values: Table 4.2 for global
average surface air temperature and precipitation for December-February and
June-August; Figure 4.4 for zonally-averaged surface temperature fields for
December-February and June-August; and Figure 4.5 for zonally-averaged mean
precipitation fields for December-February and June-August.
These figures illustrate the differences between models. For global mean
temperature, differences are on the order of several degrees Celsius (Table 4.2)
which is in the same range as the overall temperature increases discussed for
global climate change. For zonally-averaged surface temperature, model errors
tend to be larger at higher latitudes (Figure 4.4).
Global mean precipitation simulation values vary from observed values by
up to 30 per cent (Table 4.2). Spatial variations in model-simulated precipitation
differences from observed values tend to be largest in the tropical latitudes.
For the models using surface heat flux corrections over the oceans, the
typical values required to give equilibrium consistent with current climate condi-
tions were on the order of tens of watts per square metre. These values are larger
than any associated with radiative forcing associated with climate change. Since
1995 great progress has been made to reduce or eliminate the need for surface
heat flux corrections (Reference no. 7).
These examples show that model errors and variability between models in
simulations for current climate exceed changes expected with climate change. The
proper use of models for climate-change determination is to examine the changes in
model climate when radiative and other forcing associated with climate change are
added, instead of comparing the model prediction of future climate directly to current
observed climate. It is assumed that as long as the overall climate conditions simulated
by the model are close to observed conditions, then the modelled changes will be valid
indicators for climate change in the real world.
The variability in the model-simulated climates has also been examined. This
is an important but more difficult aspect of the climate system to simulate. In one
study it was shown that the modelled intermonthly standard deviation of lower
tropospheric temperature was larger than observed in the tropics and smaller than
observed in the southern high latitudes. In another study it was found that the
model simulation of the ENSO oscillations tended to underestimate the
magnitudes of the oscillations. Decadal variability also has been found in model
simulations for current climate conditions having amplitude increases with lati-
tude consistent with observations.
Table 4.2 — Coupled model Surface air Temperatures Precipitation

simulations of global average (°C) (mm/day)
temperature and precipitation. DJF JJA DJF JJA
[from page 238, Reference no. 3].
BMRC* 12.7 16.7 2.79 2.92
CCC 12.0 15.7 2.72 2.86
COLA 12.6 15.5 2.64 2.67
CSIRO 12.1 15.3 2.73 2.82
GFDL 9.6 14.0 2.39 2.50
* models without flux GISS (I) * 13.0 15.6 3.14 3.13
adjustment MPI (LSG) 11.0 15.2
MPI (OPYC) 11.2 14.8 2.64 2.73
Here the observed surface air MRI 13.4 17.4 2.89 3.03
temperature is from Jenne NCAR* 15.5 19.6 3.78 3.74
(1975) and the observed UKMO 12.0 15.0 3.02 3.09
precipitation from Jaeger (1976). Observed 12.4 15.9 2.74 2.90
60
Figure 4.4 – (a) The zonally-averaged difference of 11 coupled models’ surface air temperatures from observations for
December-February; (b) as in (a) but for June-August. Units °C. [from page 239, Reference no. 3].
Figure 4.5 – (a) The zonally-averaged precipitation rate from 11 coupled models and that from observations according to
Jaeger (1976) (solid line) for December-February; (b) as in (a) but for June-August. Units mm/day. (See Fig 4.4(a) for model
identification.) [from page 241, Reference no. 3]
In conclusion, as already judged in the 1995 IPCC report (Reference no. 3),
climate models have been shown to simulate satisfactorily as a group the large-
scale features of current climate. The IPCC report also concludes that the:
“different coupled models simulate the current climate with varying degrees of
success, and this affects the confidence that can be placed in their simulations of
climate change.”
4.3.2.2 Testing climate models on past climate situations (paleoclimate) makes it possible
Past climate conditions to evaluate model performance for a wider range of conditions of forcing of the
climate system. Chapter 2 reviewed variability of past climates in terms of surface
61
atmospheric temperature. Changes since the last glacial maximum (about 20 000
years ago) have been documented from ice cores, tree rings, fossil pollen, moun-
tain glaciers, ancient soils, closed-basin lakes, and sediments in lakes and oceans.
These records have provided enough information for climate in terms of time
changes of spatial patterns for temperature and precipitation so that validation of
the general comprehensive climate models is possible.
The general approach has been to examine model-simulated equilibrium
climates applying fixed forcing conditions for given times in the past. By this
means the evolution of climate changes corresponding to changes in the forcing
conditions has been examined. It has only been possible to evaluate the general
climate regimes produced by the model and not the details of processes and local
variations because of the limitations in the paleoclimatic data.
The validation of climate models for paleoclimate situations has proven to
be useful for testing and understanding climate models. This has led to the use of
more sophisticated climate models which include interactions with simple
‘mixed-layer’ oceans and the terrestrial biosphere. The establishment of a compre-
hensive Paleoclimate Modelling Intercomparison Project (PMIP) will provide even
more understanding of model performance in the future.
4.3.3 Modelling of the atmospheric component is the most comprehensive and devel-
EVALUATION OF CLIMATE oped of all the climate system components. Atmospheric general circulation
MODEL COMPONENTS models have been developed over a 40-year period in connection with weather
4.3.3.1 prediction and climate modelling. More than two dozen model versions exist,
Atmospheric component and these have been subject to extensive intercomparison such as in the interna-
tional Atmospheric Model Intercomparison Project (AMIP) [Gates, 1992]. In the
simulations, sea-surface temperature is normally prescribed with climatological
mean annual cycle values. This means that surface atmospheric temperature over
the oceans is constrained to be close to climatological values because of the strong
influence of ocean temperature on the temperature of the atmospheric surface
layer.
Model simulations for surface air temperature over land, precipitation, and
Figure 4.6 — The zonally- sea-level pressure have the same general quality as obtained in the coupled
averaged zonal wind (ms-1) at climate models discussed in the last section. Simulated tropospheric temperature
200 hPa as observed (black line) is close to that observed, but there is a tendency for the models to be too cold at
and as simulated by the AMIP lower levels in the troposphere in the tropics and too warm in the tropical lower
models for (a) DJF and (b) JJA. stratosphere. The associated zonally- and seasonally-averaged zonal winds at the
The mean of the models’ results is upper troposphere represent the observations well. The mean zonal wind values
given by the full white lines, and found by averaging all the model results together have errors of less than 5 ms-1
the 10, 20, 30, 70, 80 and 90 at all of the latitudes including the winter hemisphere latitudes where the jet
percentiles are given by the stream magnitudes reach 40 ms-1. There are differences among the models for jet
shading surrounding the model stream speeds at given latitudes also in the order of 5 ms-1. Figure 4.6 shows the
mean. The observed data are observed and AMIP-model distribution percentiles for the December-January and
from Schubert et al. (1992). June-August zonal wind speed mean values.
[from page 250, Reference no. 3].
62
Model simulations for cloudiness are not so good. Cloudiness is an impor-

tant component in a climate model because of its effects on radiation transfer. On
a zonally- and seasonally-averaged basis, errors in cloud cover and differences
between models have magnitudes in the order of 20 per cent. Generally the
models tend to underestimate the cloudiness in the winter and summer seasons
in low and middle latitudes, and overestimate cloudiness in polar regions.
The quality of the simulated radiation transfers is a key to the overall climate
of the atmosphere and the capability of the atmosphere model to simulate prop-
erly the effects of climate change forcing.
An important measure is the amount of terrestrial radiation emitted to space
from the atmosphere-Earth system. In terms of zonally- and seasonally-averaged
values, the models simulate outgoing terrestrial radiation values that generally
differ by no more than an order of 10 Wm-2 from observations at any latitude. It
is noted that the models are able to represent the local minimum of outgoing
terrestrial radiation values near the equator. This minimum results from the
effects of deep convective clouds in tropical areas. This convective activity is an
important component for the general circulation of the atmosphere.
The overall effect of the clouds on the radiation budget of the atmosphere is
a key factor in the climate system. Clouds affect both the solar and terrestrial radi-
ation transfer, and reflect solar radiation which leads to a reduction of input
energy; whereas they absorb terrestrial radiation which leads to a reduction in
output energy. These effects offset each other so that the net radiative forcing due
to clouds can be positive or negative.
Model simulations for the net radiative forcing due to clouds are compared
to observations in Figure 4.7. The overall latitudinal structure is reproduced by the
models fairly well. However, differences at individual latitudes such as those:
1) between the average of the model results and observed conditions; and
2) among the model results typically range between 10 and 20 Wm-2. Global and
annual mean values of the differences are less because of offset between plus and
minus values. Nevertheless, values are comparable to those for global mean radia-
tive forcing due to the greenhouse gases. This shows that cloud cover is an
important factor for model simulations for climate. The most obvious discrepancy
in model performance is the systematic underestimation of negative radiative
forcing due to clouds in the tropical and subtropical latitudes (30°S to 30°N) in
both the December-February and June-August periods and the overestimation of
the negative forcing near 60° latitude in the summer hemisphere.
Overall the atmospheric general circulation model simulations show real-
Figure 4.7 — As in Figure 4.6 istic variability for atmospheric surface temperature. The observed and
except for net cloud radiative simulated diurnal ranges in surface temperature over land are similar, although
forcing (Wm-2). The observational simulated values are too large over the high northern latitude areas in January
estimates are from ERBE data for and over the northern continents and deserts in July. The simulated annual
1985-88 (Harrison et al., 1990). cycle for surface temperature over the continents is generally reasonable
[from page 251, Reference no. 3]. although there is an overestimation of the seasonal amplitude in the drier
63
climate regions. There are large differences among the models for the seasonal
amplitudes in the higher latitudes.
Variability associated with processes internal to the atmosphere is also
present in the models but with amplitudes smaller than observed. This is true
both for extratropical variability due to synoptic storm activity and for the
Madden-Julian 30-60 day oscillation in the tropics.
Finally, the recent trends in surface temperature have been approximately
represented in atmospheric model simulations for the past 45 years using
observed values of sea-surface temperature and, in some cases, observed carbon
dioxide and tropospheric aerosol changes. The results obtained by five modelling
groups for mean surface temperature over land areas compared to each other and
to observations are shown in Figure 4.8.
4.3.3.2 Ocean general circulation models have been developed by a number of research
Ocean component groups. The highest resolution versions of these models have horizontal resolu-
tions as small as 20 km and vertical resolution up to 60 layers (as of 1995). At
present, computer size limits the resolution that can be used in coupled ocean-
atmosphere models.
Ocean general circulation models have been able to simulate interannual
variability of the type related to ENSO. In the coupled ocean-atmosphere models
the ENSO-like variability is much weaker and much more regular than in observed
conditions. Some ocean general circulation models have been able to simulate the
Rossby waves and equatorial Kelvin waves that are part of the ENSO variability
cycle. The newest coupled models show ENSO rather near to reality both in
frequency and intensity. Decadal variability has also been simulated in ocean
basins where deep water formation occurs (due to downward convective motions
of cold water).
As concluded in the 1995 IPCC report [Reference no. 3], ocean general circu-
lation models are able to portray realistically the large-scale structure of the
oceanic gyres and the main features of the thermohaline circulation. Primary defi-
ciencies are in the representation of mixing processes and the structure and
strength of the western boundary currents, the simulation of meridional heat
transport, and the simulation of convection and subduction. Accordingly, a
number of problem areas and deficiencies remain to be addressed, these include:
(a) Representation of geometry and bathymetry;
(b) Parameterization of subgrid-scale processes as convection, mixing, and mesoscale
eddies;
(c) Errors in defining surface forcing by the atmosphere;
(d) A thermocline that is too deep and too diffuse;
(e) Weak poleward heat transport;
(f) Distortion of upper ocean and deep boundary currents; and
(g) Temperature and salinity errors in deep water.
An important challenge, adding to the list above, for further improving and
understanding ocean general circulation modelling is obtaining observational data for
the oceans. Observations for the ocean are insufficient in coverage especially in the deep
waters for comprehensive description of currents and thermohaline structure and for
their variability. In many parts of the ocean the observational record is insufficient to
define the key decadal and longer-term variability characteristics.
4.3.3.3 The land-surface component is dependent on the parameterization for a number

Land-surface component of processes that involve water vapour, evaporation, liquid water distribution and
storage, soil moisture, heat and momentum transfer, ground temperature, and
radiative exchanges. The parameterizations depend in part on resolved variables
in the atmospheric model, so that deficiencies in the atmospheric model will
influence the performance of the land-surface parameterizations.
Differences between parameterization methods have been shown by the
Project for Intercomparison of Land Surface Parameterization Schemes (PILPS). In
this project 20 land-surface models were compared for computed values of surface
heat and moisture fluxes and liquid water runoff and drainage. All of the models
64
Figure 4.8 — The observed global

annual 1.5 m land air OBSERVED
temperature for 1949 to 1993
0.4 UKMO Fixed CO2
from the Jones (1994) dataset
(shaded) and the corresponding UKMO Changing CO2
Temperature Anomaly ( C)
O
modelled 1.5 m air temperature UKMO Changing CO2 + SO4
deviations from the 1950 to BMRC
1959 average for: (i) the average
of four simulations with the
0.2 GSIRO
UKMO model without progressive COLA
changes in radiative forcing; JMA
(ii) the average of four
simulations with changing CO2 ;
(iii) the average of four models 0.0
with a representation of
tropospheric aerosols; and (iv) the
averages from four other models
with no changes in radiative
forcing. [from page 258, -0.2
Reference no. 3]. 1950 1960 1970 1980 1990
YEAR
were subjected to the same observed atmospheric forcing inputs for rainfall and
surface radiation, and for low-level temperature, moisture and winds. One year of
atmospheric observations at Cabauw, the Netherlands, was used for the forcing
inputs and for verification of the land-surface quantities computed. Each model
was run for as many annual cycles as necessary to obtain a near-constancy for the
annual averages of the computed quantities.
General results are shown in Figure 4.9. The variations have certain correla-
tions. For instance, an increased sensible heat flux relates to decreased latent heat
flux implying that the total of sensible plus latent heat flux has less variation.
Similarly, the increase in runoff relates directly to decreases in evapotranspiration.
Nevertheless, the range in sensible and latent heat fluxes individually was on the
order of 20 Wm-2 with values roughly centered on the observed amounts. The range
in runoff and evaporation amounts was in the order of 300 mm (for a year). Note
that for the runoff and evapotranspiration, results are included for several cases
where the land surface model was forced by results from atmospheric models (refer-
enced AMIP/PILPS). The generally wider spread of the later points illustrates the
error contribution from the atmospheric model.
Another intercomparison experiment was made for the annual cycle of soil
moisture using 13 land-surface models and observational data from an experi-
mental site in southern France. The range for computed soil moisture amounts
among the simulations was about 100 mm. The range was the greatest in the
summer when the schemes tended to underestimate soil moisture.
4.3.3.4 Sea-ice models have been developed that can simulate sea-ice thickness and
Cryosphere component coverage. An example of capability is shown in Figure 4.10. In current climate
(sea-ice models) models, only very approximate sea-ice models have been used. Major simulation
discrepancies exist in climate models such as the ice edge being too far south in
the Atlantic Ocean and not far enough north in the southern ocean.
It is necessary to develop more comprehensive sea-ice models for climate
modelling. Sea-ice affects ocean salinity and oceanic convection that sends
surface water down to the ocean depths. Furthermore, sea-ice transport results in
important heat transport. There is some evidence from preliminary studies that
incorporating sea-ice dynamics reduces the climate (change) sensitivity of the
climate model. Recently the Sea-Ice Model Intercomparison Project (SIMIP),
which was undertaken within the Arctic Climate System Study (ACSYS) has
completed an intercomparison of sea-ice models and promoted a ‘viscous-plastic’
sea-ice model for climate models [Lemke et al., 1997].
65
Figure 4.9 — (a) The annually (a)

averaged latent and sensible heat
fluxes predicted by the PILPS 15
land-surface schemes. A single
year’s observations from 10
GISS
Cabauw, The Netherlands, were
used for as many annual cycles
5 ECHAM
as was required for each land- ISBA
MOSAIC
Sensible heat flux ( m–2)

surface scheme to conserve energy UGAMP2
UKMO
BATS
SWAP BASE
(≤3 Wm-2) and water (≤3 mm/yr). 0 SECHIBA2 SSIB VIC
SEWAB CLASS
CAPSLLNL CSIRO9
(b) As in (a) but for the annual CABAUW (obs) PLACE
NMC CAPS
totals of evaporation and runoff –5
SWB
plus drainage. The AMIP/PILPS SPONSOR
models’ 10-year mean values are –10

the weighted average for the
closest GCM grid point and as
many of the surrounding eight –15
grid points that are designated as

BUCK
land. (The AMIP precipitation –20
totals are given in parentheses
and differ from those prescribed
–25
for the off-line forcing.) [from 25 30 35 40 45 50 55 60
page 259, Reference no. 3]. Latent heat flux (Wm–2)
(b)
500
CCC (1191)
400 ECHAM
UGAMP2
Runoff + Drainage (mm)
SWAP GISS BMRC (1102)
SECHIBA2 UKMO
ISBA SEWAB
300 CAPSLLNL MOSAIC
SSIB BATS UKMO (786)
VIC CSIRO9 (784)
CLASS BASE
CABAUW (obs) CSIRO9
NMC
MPI (717) PLACE
200 CAPS
SWB
SPONSOR
100
PILPS (776) BUCK
AMIP (precipitation)
0
300 400 500 600 700 800
Evapotranspiration (mm)
4.3.4 The climate model is a very complex entity. Improvement in its solutions requires
SENSITIVITY OF CLIMATE improving many of the subparts. Improvements cannot be made in isolation as
MODELS TO MODEL the parts affect each other, and the final solution includes approximations that
FORMULATION, BOUNDARY balance each other. A parameterization scheme that is improved in isolation may,
CONDITIONS, AND in fact, reduce the quality of the solution of the model system because it changes
PARAMETERIZATION inputs to other parameterization schemes and that could offset existing compen-
sations within the model.
The key strategy is to focus on improving the parts of a model system which
contribute the most uncertainty to the model simulations. This avoids giving
66
Figure 4.10 — (a) Simulated, and

(b) observed sea-ice concentrations
(in per cent) in the Weddell Sea for
September 1987 (Fischer and
Lemke, 1994). [from page 268,
Reference no. 3].
excess attention to improving parts of a model which would not make much
difference to the quality of the solution. An important part of model evaluation
is to identify the aspects of a model most responsible for its simulation deficien-
cies, i.e., ‘model sensitivity.’ The 1995 IPCC report (pp. 271-274, Reference no. 3)
lists five key areas of sensitivity given below.
(a) Water vapour content has a large range of values in the troposphere. Because of
Representation of water vapour this, numerical approximation errors for advection of water vapour can lead to
large errors in water vapour distribution. Such errors can lead to errors not only
in cloud coverage and precipitation but also in latent heating and radiation trans-
fer. Special numerical techniques are needed to deal with this problem.
(b) Model resolution in both the atmosphere and the ocean model is considered to
Model resolution have important impacts on simulation quality. Resolution affects both the
numerical errors in the resolved phenomena of the model and the behavior of
parameterization schemes. It may be hard to separate these two effects. It is
appropriate to consider a resolution that varies in space such as is usually done
with vertical resolution near the atmosphere-ocean interface and near the
tropopause in the atmosphere.
(c) The sensitivity of climate model solutions to the cumulus convection parameteri-
Convection and clouds zation scheme has already been demonstrated. Cumulus convection
parameterization quality is very important in tropical regions as it affects precipita-
tion, cloud cover, cloud radiative forcing, and the overall hydrological cycle.
Climate model solutions are also sensitive to microphysical processes for all clouds.
(d) The earlier discussion showed the wide range of solutions given by current land
Land surface processes surface model parameterization methods. The overall climate solution may be less
sensitive to the details of the land surface because land covers only 29 per cent of
the Earth’s surface; however, the land surface is where people live so it will be
67
necessary to have more detailed description of climate change effects for the land
areas.
(e) In order to offset the effects of chaos sensitivity due to the nonlinear character of the
Initial conditions and surface fluid dynamic equations for the atmosphere and ocean, it is considered necessary to
boundary conditions employ ensemble simulation techniques to reduce the sensitivity to initial condi-
tions. This is considered important for aspects of the climate such as interannual
variability. It is believed that the boundary conditions at the ocean surface play an
important role in the stability of a simulated thermohaline circulation in the ocean.
Of the points listed above, it is felt that clouds, convection, the hydrological
cycle, and land surface processes — points (c) and (d) above — are the areas of
largest uncertainty in climate models.
4.3.5 The 2001 IPCC report summarized recent advances in modelling studies. These
UPDATE FROM 2001 IPCC have improved even further confidence in the ability of models to predict future
REPORT (Reference no. 7) climates. These advances include increased computer power to enable higher reso-
lution models and multiple runs (ensembles) of models, better coupled model
simulation of lower frequency events such as the El Niño-Southern Oscillation
(ENSO), and model calibration for climate changes using paleoclimate simulations.
68
CHAPTER 5
CLIMATE PREDICTION
5.1 Climate prediction covers many ranges of time. Strategies for handling these can
INTRODUCTION be organized into three time-scale categories: short-range (covering seasonal and
interannual time scales), medium-range (covering time scales larger than interan-
nual up to a century), and long-range (covering time scales on the order of 10 000
years) [see Bengtsson in Chapter 23 of Reference no. 6]. Prediction strategy can
build on experience of day-to-day weather prediction, and it must also consider
the nonlinear and chaotic characteristics of the atmospheric and oceanic compo-
nents in the climate system. ‘Predictability,’ defined as the length of time for
which the model can provide useful, deterministic information, is greatly influ-
enced by the nonlinear characteristics of the system.
5.2 Predictability limits in numerical weather prediction models have been measured
PREDICTABILITY by the growth in differences, or conversely the reduction in correlations, between
the model forecast and the corresponding observational information. Commonly
the atmospheric variable examined has been the 500 mb height field. Typically,
the growth of model errors in time is exponential when the errors are very small
compared to overall spatial variability. The error growth gradually decreases to
zero as error magnitudes reach values similar to overall spatial variability.
Conversely, the correlation between the time-dependent spatial patterns
(anomaly structure) in the model field and those in the observations will decrease
with time and eventually approach zero.
The predictability of the model can be measured by the length of time the
above correlation remains greater than a specified base value. In setting the base
value, an effort is made to distinguish model prediction skill in the forecast from
other factors which might happen to give a correct forecast. For instance, a
persistence prediction (predicting continuation of existing conditions) can be
correct some of the time. The skill in the model forecast is represented by the
increase in anomaly structure correlations between the model forecast and obser-
vations over those obtained with a persistence or climatological forecast.
The predictability of a numerical weather prediction model depends on the
accuracy of its numerical formulation, the accuracy of the initial conditions, and
the rate at which smaller scales unresolved in the model actually affect the resolv-
able scales. No matter how well the first two conditions are met, predictability will
always be limited by the third factor in a geophysical fluid system because the
model can never resolve all scales. Even in the geophysical fluid system itself there
is no complete determinism from an initial state because of the inherent uncer-
tainty in the initial state and of forcing effects. Lorenz (1969) gave a graphic
example of the concept of predictability limits by stating that it could be argued
that the flapping wings of a single butterfly could completely alter the details of
the entire atmospheric system if given a sufficiently long time.
For weather prediction, the inherent limit to predictability has been esti-
mated to be on the order of two to three weeks based on theoretical studies by
Lorenz (1969, ob cit.). The predictability achieved by weather prediction models
for the entire region north of 20°N has improved greatly over the past 30 years as
shown in Figure 5.1. From data sources such as those shown in Figure 5.1, it has
been estimated that the ‘useful’ predictability skill achieved by models increased
from slightly more than half a week in the mid-1960s to nearly ten days in the
1980s.
Weather prediction models have become so accurate for one-day forecasts
that it is now possible to approximate closely the predictability limits to which
weather prediction models may be compared. The procedure proposed by Lorenz
(1982) compares forecasts of one-day difference in length for each day in the
69
Figure 5.1 — The predictive skill

in numerical weather prediction
and its improvement from (a) the
mid-1960s to (b) the late 1980s.
Shown is the decrease with time
of the correlation coefficient
between predicted and observed
500-mb geopotential height
anomalies for 12 two-week
predictions in (a) and 90 ten-day
predictions in (b). The correlation
coefficients were computed using
all grid-point values north of
20°N. The thick solid lines show
the mean correlation values
averaged over all individual
cases, whereas the shaded areas
give a measure of the range of the
individual cases (i.e., about 5 per
cent of the predictions lie below
the shaded area and 5 per cent
above it). For comparison, the
persistence curve (dashed) forecast period. Likewise this procedure specifically compares the two-day and
indicates the no-skill forecast. one-day forecasts for tomorrow, the three-day and two-day forecasts for the day
The growth in (useful) predictive after tomorrow, the four-day and three-day forecasts for the day after that, and so
skill from less than one week in forth. This provides a predictability upper-limit estimate containing only one-day
the mid-1960s to about ten days forecast errors for the whole period of time. Figure 5.2 shows an example of fore-
in the 1980s is evident. The data casts for a winter season. This procedure provides estimates on how close the
in (a) are from Miyakoda et al., ten-day forecast is to the best that could be obtained (the solid and dashed lines,
1972, and in (b) from respectively, in Figure 5.2).
unpublished ECMWF statistics Predictability considerations for climate are quite different from those for
(courtesy L. Bengtsson). [from weather. Climate prediction is made for time periods longer than the two to three
page 461, Reference no. 5, with week weather predictability limits. This is possible because climate prediction is
permission of Springer-Verlag]. for statistical descriptions of the weather conditions, such as monthly means,
instead of instantaneous conditions. Climate predictability extends to periods
longer than two or three weeks; this is especially true in situations influenced by
slowly varying oceanic conditions such as ENSO.
However, nonlinear processes add complexity to climate predictability.
Specifically, there may be more than one climate for a given set of forcing param-
eters. In such a case, the climate is not unique and is termed ‘intransitive.’ If the
climate is unique, it is termed ‘transitive.’ The first case is an example of chaos
theory where ‘multiple attractors’ exist. The ‘attractor’ is mathematical terminol-
ogy that may be considered a frame of reference for a ‘given climate.’ The theory
goes on to describe how the shift from one ‘attractor’ (or climate) to another may
be random and unpredictable.
It is not known what, if any, aspects of climate are intransitive. However, this
characteristic would imply that a small additional forcing, such as that due to
human activity, could potentially cause a general shift to a new climate pattern
which is not necessarily reversible even if the small additional forcing is removed.
Clearly, this is an important area for research.
An ‘ensemble forecasting’ modelling strategy has been developed to deal
with numerical predictability in situations which are highly sensitive to nonlin-
ear processes. In this method a large number of numerical simulations are made
for a forecast period, using the same model. The simulations differ slightly only
in details of the initial conditions or forcing effects. Then statistics for model fore-
cast variability can be determined as a function of space and time. These statistics
show where the model solution is more variable and therefore less reliable for esti-
mating values. The statistics also provide information on the probability of
occurrence for specific forecast states.
70
This approach has been used for operational weather prediction in which up
to 50 separate forecasts may be made for a given forecast period. It is just as valid
for climate models. Ensembles of up to ten forecasts are currently used in short
range climate prediction (seasonal forecasts). The intercomparison tests among
climate models, such as those discussed in Chapter 4, serve a similar function.
However, in such intercomparison experiments, it is not possible to separate inter-
nal nonlinear processes from effects due to differences in model formulation and
parameterization.
Short-term climate forecasting (seasonal to interannual time scales) is a direct

extension of deterministic long-range weather forecasting. Currently, long-range
weather forecasting has been extended to seasonal forecasting by focusing on the
pronounced ENSO variability in the tropical Pacific ocean area and its relation-
– Winter precipitation in central Chile;

– Central England summer precipitation (based on Atlantic Ocean sea-surface
temperature);
– Tropical Pacific and Indian Ocean sea-surface temperatures;
– Northern tropical Atlantic sea-surface temperatures;
– Atlantic tropical storm activity;
– Southern Oscillation index time series;
– Summer monsoon rainfall in central-east China;
– Tropical Pacific island precipitation;
– Canadian wintertime temperature and precipitation; and
– U.S. temperature and precipitation.
A range of methods is often used in seasonal forecasting. For example, a
number of groups currently supply seasonal forecasts for rainfall in north-east
Brazil. These forecasts use January conditions for sea-surface temperature, partic-
ularly in the Atlantic Ocean, to predict rainfall anomalies in the following spring.
Methods used include statistical linear regression, discriminant analysis and
dynamic approaches. Correlation skills are generally in the 0.9 range for this
seasonal forecast.
The long-range forecasts for tropical Pacific sea-surface temperatures show
skill. As an example, forecasts from a coupled ocean-atmosphere model at the
Climate Prediction Center, U.S. National Centers for Environmental Prediction
(NCEP), are presented in Figure 5.3. These are mean values of ensemble forecasts
for sea surface temperature anomalies in the eastern tropical Pacific for the area
between 5°N and 5°S latitude and longitudes ranging from 150° to 90°W for the
‘NIÑO’ index and 170° to 120°W for the ‘NIÑO3.4’ index. Forecasts made three,
six, and nine months in advance are compared with observed values. Note that
the forecasts for late 1997 included the large El Niño event that did actually occur.
Since January 1998 global seasonal forecasts have been made available from
the European Centre for Medium Range Weather Forecasts (ECMWF) on the
Internet at the web site: <http://www.ecmwf.int/html/seasonal/info/info.html>.
An example of their six-month forecast for the NIÑO-3 index anomaly during the
onset of the major 1997/98 El Niño event is shown in Figure 5.4. The monthly
forecast ensemble is based on forecasts made three times a week, which gives an
ensemble of 12 to 15 members for each month.
Forecasts for climate conditions in other parts of the world which have some
observed correlation with the ENSO generally have limited, but useful, accuracy
over subregions of the forecast area. The results of an experimental programme to
predict seasonal rainfall over northern Africa one month prior to the season are
are a good example. (A joint programme among operational weather prediction
Figure 5.3 — NCEP coupled- units in England, France, and ECMWF.) An ensemble prediction method was
model, monthly-mean SST
anomaly forecast time series for
Niño-3 and Niño-3.4 forecasts
from 1994 to 1998 for three, six,
and nine month lead times in the
upper, middle, and lower panels
respectively. These monthly
values were used in three-month
mean SST anomaly forecast
products. The predictions
represent the mean of three
ensemble-mean forecasts, each for
one of the three most recent
months, respectively, and each
produced by forecasts from two to
three individual one to two-week-
apart initial conditions per
month. [from page 13, Ji. et al.
(1997)].
72
CHAPTER 5 — CLIMATE PREDICTION
Figure 5.4 — Monthly sea-surface NIÑO 3 anomaly

temperature (SST) ensemble
predictions for the eastern tropical 30oN
20oN
Pacific (the NIÑO-3 region) for ~ 2
Nino
10oN
the onset of the 1997-98 El Niño.
Latitude
0 ~ 4
Nino ~ 3
Nino
Six-month predictions from a o
*
10 S *
coupled ocean-atmosphere model o
20 S ~ 1
Nino
o
(fine lines) compared with 30 S
140 E 180 E 180 180 W 140 W 120 W 100 W 80 W
observations (heavy line) are Longitude
shown for the November 1996 to

September 1997 time period.
Predictions are initiated over an
extended period from November
1996 to March 1997. Three
forecasts are made each week
(only one is plotted in order to
avoid clutter) which gives an
ensemble of 12 to 15 members
each month. Note that the onset
of El Niño in April 1997 is well
forecast. The subsequent
evolution is also well forecast,
although one can see a tendency applied to the 15-year period from 1979 to 1993. Correlations between the ensem-
to underpredict the amplitude of ble mean rainfall and the observed rainfall were made by Harrison, et al. [1997].
the event. [from ECMWF, 1998]. After calibrating the ensemble forecast variance, they found that in some areas the
observed rainfall was within the ensemble range for most of the 15 years while in
others this was the case for only ten of the 15 years.
In summary, useful seasonal prediction is now possible for selected regions
of the globe. The prediction is generally for anomalies in monthly mean temper-
ature or precipitation. The skill level in some cases may be sufficient to be of value
as an operational product for the public. It is important that operational meteo-
rologists become acquainted with possible benefits for their area of service.
The World Meteorological Organization has instituted a Climate
Information and Prediction Services (CLIPS) project which will make short-range
climate forecast information available to the world meteorological community. It
is important to understand the strengths and limitations in forecasting to be able
to integrate CLIPS most effectively with national service and policy.
5.4 Forecasting climate variability for decadal to century time scales does not deal
MEDIUM-RANGE CLIMATE with specific variability phenomena but rather with simulation of overall climate
FORECASTS processes and changes due to external forcing that causes climate change. It is
expected that the modelling of all components of the climate system will have an
important role in forecast quality. For such forecasts, major variations in some of
the external forcing conditions, for example changes in the major ice sheets that
occurred in the last ice age, may be excluded.
It is important that the climate models used for simulating this time scale
have reasonable equilibrium characteristics for the control state. Climate drift due
to energy transfer discrepancies must be counteracted with ‘flux adjustments’ as
discussed in Chapter 4.
For this scale of forecasting and certainly for long-range climate prediction,
the possible intransitive nature of the ocean thermohaline circulation can present
a challenge. There is evidence that a significantly different circulation existed in
the Atlantic Ocean around 11 000 years ago. The increase in fresh water runoff
from north America into the north Atlantic caused oceanic subduction and down-
ward convection to stop. This in turn prevented the formation of cold deep ocean
waters and reduced currents that provide the northward transport of heat by the
oceans. This resulted in a lowering of atmospheric temperatures in the northern
polar latitudes. Modelling studies have shown that this large change in circulation
could have occurred on a decadal time scale.
73
Calibration of medium-range climate forecast products is provided by testing

the model on the climate of the recent past, for a period for which sufficient
observational data are available. An example was shown in Figure 4.9. A model
simulation of global mean surface temperature for the entire period of climate
change since the pre-industrial era is shown in Figure 5.5. In this simulation the
greenhouse gas and aerosol forcing is based on the observational estimates of the
concentration of these forcing constituents. Note that the model forecast for
overall trends is reasonable; however, decadal variability, especially in the period
from 1920 to 1960, is not forecast well.
5.5 Forecasting for time scales of 10 000 years and greater requires climate models
LONG-RANGE CLIMATE with the full representation of the components of the climate system. Effects due
PREDICTION to the deep circulations in the oceans (with time scales of 1 000 years or more)
and major changes in the cryosphere (glaciers and ice fields) must be included. A
number of processes, such as the oceanic thermohaline circulation, may cause the
climate to be intransitive. This would mean that there is not a unique climate for
a specified climate forcing, so that the model prediction would depend on the
initial conditions.
Model studies to simulate past climates have demonstrated that climate
models can represent different climate conditions. However, these studies have
not reproduced all of the process cycles in the evolution of the climate system. As
discussed by Bengtsson [p. 721, Reference no. 6], the time scales that must be
handled for long-range prediction range from hours for the atmosphere to
centuries and millennia for the cryosphere, ocean, and astronomical forcing.
Bengtsson goes on to describe a modelling technique originally suggested by
Hasselmann (1988) whereby the ‘slow system’ representing the land ice, deep
ocean, and astronomical forcing are solved separately from the ‘fast system’
consisting of the atmosphere, land surface and the upper ocean. An equilibrium
condition would be found in the fast system for a given state of the slow system.
The forcing effects of the fast system due to this equilibrium condition would be
maintained as constant on the slow system for a time duration comparable to the
time scale for changes in the slow system. Periodically a new equilibrium would
be computed for the fast system to be consistent with changes in the slow system.
5.6 The primary climate simulation models are global. Currently available computer
PREDICTABILITY FOR power limits the resolution of the global climate models to a few hundred kilo-
REGIONAL CLIMATE meters in the horizontal. This means that local climate conditions and variations
cannot be represented. Yet the local conditions are those which relate most
directly to climate-impact assessments. Local variations and changes in climate
1.5
Global temperature change (°C)
Equivalent CO2 only

Equivalent CO2 and the direct effect of sulphate aerosol
Observed
1.0
Figure 5.5 — Simulated global

annual mean warming from
1860 to 1900, allowing for 0.5
increases in equivalent CO2 only
(dashed curve) and allowing for
increases in equivalent CO2 and
the direct effects of sulphates 0.0
(flecked curve) (Mitchell et al.,
1995). The observed changes are
from Parker et al. (1994). The
anomalies are calculated relative -0.5
to 1880-1920. [from page 297, 1860 1880 1900 1920 1940 1960 1980
74
are much larger than averages measured over large spatial scales. Below, an
example of the limitations of the global climate models for regional climate
prediction is presented followed by a brief discussion of two methods that can
provide useful predictive information on local climate from climate models:
statistical ‘downscaling’ and higher resolution regional models.
5.6.1 Analysis of regional results in global simulation models demonstrates the large
GLOBAL CLIMATE MODELS variability in comparing model results with each other and with observations.
Nine modelling groups examined seasonal climatologies for surface temperature
and precipitation in control runs intended to represent existing current condi-
tions. Comparisons were made for seven regions over land areas with dimensions
of very roughly 2 000 km by 2 000 km. They were from central north America
(CNA), south-east Asia (SEA), the Sahel (SAH), southern Europe (SEU), Australia
(AUS), northern Europe (NEU), and east Asia (EAS). Results shown in panels b, d,
f, and h of Figure 5.6 illustrate the large difference between the global models.
5.6.2 The statistical method has the following two-step approach: 1) Development of
STATISTICAL DOWNSCALING statistical relationships between local climate variables and large-scale predictors from
TECHNIQUE the global climate model; and 2) Application of such relationships to the output from
the climate models to estimate local climate characteristics. This approach has been
quite successful in relating weather-prediction model simulations to local surface
weather conditions. The method requires a large number of realizations where model
output is related to corresponding observed surface conditions in order to determine
the appropriate model predictors and statistical coefficients.
This method can be applied to any predicted parameter that has a physical
relationship to the model variables, as long as there is an observational record for
(a) Winter, Temperature (b) Winter, Temperature (c) Summer, Temperature (d) Summer, Temperature
7.0 7.0 20.0
20.0
CO2 – Control Control – Observed
Temperature difference (°C)

Temperature difference (°C)
Temperature change (°C)

Temperature change (°C)
6.0 6.0 15.0

15.0
5.0 5.0 10.0

10.0
4.0 4.0 5.0

5.0
3.0 3.0 0.0

0.0
2.0 2.0 –5.0

–5.0
1.0 1.0 –10.0

–10.0
0.0 0.0 –15.0

–15.0 CNA SEA SAH SEU AUS NEU EAS CNA SEA SAH SEU AUS NEU EAS
CNA SEA SAH SEU AUS NEU EAS CNA SEA SAH SEU AUS NEU EAS
Region Region Region Region
(e) Winter, Precipitation (f) Winter, Precipitation (g) Summer, Precipitation (h) Summer, Precipitation
60.0 200.0
60.0 200.0
Precipitation (% of Observed)

Precipitation (% of Observed)
Precipitation (% of Control)
CO2 – Control
Precipitation (% of Control)
150.0 Control – Observed 40.0

150.0
40.0
100.0
100.0
20.0
20.0
50.0
50.0
0.0
0.0 0.0
0.0
–20.0
–20.0 –50.0
–50.0
–40.0 –100.0
–40.0 –100.0 CNA SEA SAH SEU AUS NEU EAS CNA SEA SAH SEU AUS NEU EAS
CNA SEA SAH SEU AUS NEU EAS CNA SEA SAH SEU AUS NEU EAS
Region Region Region
Region
MRI(p) GFDL(g) BMRC(a) NCAR(r) UKMO(t ) MPI(x ) NCAR(q ) UKMO(s) MPI(m)
Figure 5.6 — Differences between averages at time of CO2 doubling and control run averages (CO2 –Control) and difference
between control-run averages and observed averages (Control–Observed) as simulated by nine AOGCM runs over seven
regions. (a), (b) Temperature, winter; (c), (d) temperature, summer; (e), (f) precipitation, winter; (g), (h) precipitation, summer.
Units are °C for temperature and percentage of control run or observed averages for precipitation. In (f) and (h) values in excess
of 200 per cent have been reported at the top end of the vertical scale. In (e) values in excess of 60 per cent have been reported
at the top end of the vertical scale. CNA=central north American, SEA=south-east Asia; SAH=Sahel; SEU=southern Europe;
AUS=Australia; NEU=northern Europe; EAS=east Asia. [from page 338, Reference no. 3].
75
calibration. The method can provide results for situations where small spatial
structure is expected, such as for temperature and precipitation in regions with
large topography. One example of application is the prediction of sea level at tidal
gauges in Japan from model-simulated sea-level pressure anomalies [Maochange
et al., 1995].
5.6.3 The regional modelling approach uses output from the global climate model to
REGIONAL CLIMATE MODELS provide initial and boundary conditions for a regional climate model. These
regional models are forced ‘one way’ by the global climate model, i.e. the global
climate model determines forcing conditions for the regional model but the
regional model does not, in turn, influence the global model.
These models can have a much higher resolution and can incorporate phys-
ical processes not in the global model. Regional climate models have included
coupling to lake models, dynamic sea ice models, coastal ocean models and
ecosystem models.
Experiments using regional climate models for present-day climate experiments
have shown the following results for regional climate models [see pp. 340-341,
Reference no. 3]. These models had horizontal resolutions ranging from 15 to 125 km.
(a) Realistic synoptic events have been simulated with small biases in temperature
and precipitation when initial and boundary conditions were provided by obser-
vations. Biases were in the range of a few °C for temperature and 10-40 per cent
for precipitation.
(b) Performance was degraded when the quality of the model forcing (i.e. specifica-
tion of initial and boundary conditions) was reduced by using general circulation
model simulations instead of observations to force the regional climate model.
(c) Simulations produced more realistic detail than in the global climate models used
to force the regional models; however, regionally-averaged values could be more
or less realistic than those in the driving climate model.
(d) Models performed better at mid-latitudes than in tropical regions.
(e) Model performance improved as the resolution of the driving global climate
model increased.
(f) Seasonal as well as diurnal temperature ranges were simulated reasonably well.
(g) Validation data from adequately dense observational networks was lacking, espe-
cially in mountainous areas.
5.7 The 2001 IPCC report lists ten highlights in modelling advances since the 1995
UPDATE HIGHLIGHTS ON IPCC report. These are listed below.
CLIMATE MODELING • Coupled models can provide credible simulations of both the present annual
FROM THE 2001 IPCC mean climate and the climatological seasonal cycle over broad continental
REPORT scales for most variables of interest for climate change. Clouds and humidity
remain sources of significant uncertainty but there have been incremental
improvements in simulations of these quantities.
• Confidence in model projections is increased by the improved performance of
several models that do not use flux adjustment. These models now maintain
stable, multi-century simulations of surface climate that are considered to be
of sufficient quality to allow their use for climate change projections.
• There is no systematic difference between flux-adjusted and non flux-adjusted
models in the simulation of internal climate variability. This supports the use
of both types of model in detection and attribution of climate change.
• Confidence in the ability of models to project future climates is increased by
the ability of several models to reproduce the warming trend in 20th century
surface air temperature when driven by radiative forcing due to increasing
greenhouse gases and sulphate aerosols. However, only idealised scenarios of
sulphate aerosols have been used.
• Some modelling studies suggest that inclusion of additional forcings such as solar
variability and volcanic aerosols may improve some aspects of the simulated
climate variability of the 20th century.
• Confidence in simulating future climates has been enhanced following a system-
atic evaluation of models under a limited number of past climates.
76
• The performance of coupled models in simulating the El Niño-Southern

Oscillation (ENSO) has improved; however, the region of maximum sea-surface
temperature variability associated with El Niño events is displaced westward
and its strength is generally underestimated. When suitably initialised with an
ocean data assimilation system, some coupled models have had a degree of
success in predicting El Niño events.
• Other phenomena previously not well simulated in coupled models are now
handled reasonably well, including monsoons and the North Atlantic
Oscillation.
• Some palaeoclimate modelling studies and some land-surface experiments
(including deforestation, desertification and land cover change) have revealed
the importance of vegetation feedbacks at sub-continental scales. Whether or
not vegetation changes are important for future climate projections should be
investigated.
• Analysis of, and confidence in, extreme events simulated within climate
models is emerging, particularly for storm tracks and storm frequency.
‘Tropical cyclone-like’ vortices are being simulated in climate models, although
enough uncertainty remains over their interpretation to warrant caution in
projections of tropical cyclone changes.
77
CHAPTER 6
OBSERVATIONS FOR LONG-TERM CLIMATE MONITORING
6.1 Observations are a critical component in the discussion, understanding, and iden-
INTRODUCTION tification of climate change. There are uncertainties in the observational data
which contribute to the overall uncertainty about climate variability both natural
and anthropogenic. Many persons might assume that the current observational
network is sufficient. A global observing system has been in place for years to
support operational weather prediction; it forms part of WMO’s World Weather
Watch (WWW) programme. In recent decades, the observations have been
substantially reinforced by observations from new systems such as satellites.
Ocean measurements are an indispensable part of the global climate observing

system. Measurements are needed to provide spatial and temporal descriptions of
temperature, salinity and currents. Measurements for sea-surface temperature
have been comprehensive; however, measurements for conditions below the
ocean surface are currently insufficient. A fully four-dimensional observational
system is needed for the ocean as exists for the atmosphere.
There are serious deficiencies in, and critical issues for, the observational system
for climate monitoring for the atmosphere, as well as the ocean. It is important
that these deficiencies and critical issues should be understood and addressed. In
recognition of the needs for climate monitoring, a new international programme
was established by the Second World Climate Conference in 1990, the Global
Climate Observing System (GCOS). Local weather observations provide a key
foundation for climate-information systems. Meteorologists throughout the
world need to understand the importance of long-term and well-documented
local observations for climate monitoring.
There are several underlying principles for a climate monitoring system.
First, the system must be underpinned by the scientific community in terms of
development, calibration, and monitoring. This is true even for automated
observing systems. Second, it must be understood that climate observations have
requirements that go beyond those for weather observing. Third, the observa-
tional records need to be long-term and have a consistent, homogeneous, and
documented frame of reference to be able to detect trends in climate conditions.
This chapter leads off with elaboration on the key specific principles for
long-term climate monitoring. Examples of the status of selected observations
relevant to climate change are then presented followed by discussion of strategies
for improving long-term climate monitoring. The primary reference for this
chapter is Reference no. 2 listed in the introduction.
6.2 The list of ten principles below covers the details for assuring that observations
KEY PRINCIPLES FOR will be of value for long-term climate monitoring. They apply to observations
LONG-TERM CLIMATE taken anywhere in the world. Staff meteorologists, weather station managers,
MODELING meteorological service directors, and the appropriate government agencies
need to become aware of and understand the importance of these principles.
The principles apply to both small and simple and large and elaborate
observing operations. The list is presented in complete form on pages 86-87 in
Reference no. 2.
(a) The effects on the climate record due to changes in instrumentation, observing
System changes practices, observation location, etc., must be known before implementing such
changes. These effects can be determined by a period of overlapping measure-
ments between the old and new system or by comparison of the old and new
systems with a standard reference. Sites chosen for in-situ measurements should
78
CHAPTER 6 — OBSERVATIONS FOR LONG-TERM CLIMATE MONITORING
have expectations of long and uninterrupted use and expectations that there will
be little change of the nearby physical environment over time.
(b) Processing algorithms for determining data values from the instrument system
Processing algorithm description must be well documented and archived with the original data.
(c) Information on instrument, station, platform history, changes in sampling time

Observing system description and local environmental conditions, and all other factors relevant to interpreta-
tion of the data should be recorded as part of the observing routine and archived
with the original data.
(d) Observations that have a long, uninterrupted record should be maintained and
Length of record kept uniform (homogeneous) in terms of measurement procedures. Long-term for
space-based measurements is measured in decades, but long-term records for
more conventional measurements may be a century or more.
(e) Calibration, validation and system maintenance should be used to provide a

Climate record homogeneity constant frame of reference for the climate record.
(f) Some form of ‘low-technology’ back-up to ‘high-technology’ observing systems

System backup should be put in place to safeguard against unexpected operational failures.
(g) Highest priority should be given to the design and implementation of new
Observing system priorities climate observing systems for; 1) data-poor regions; 2) variables and regions sensi-
tive to climate change; and 3) key measurements with inadequate spatial and
temporal resolution.
(h) Long-term climate requirements should be made known to the designers and
Network design engineers at the outset of designing a new network.
(i) A long-term association and cooperation commitment is needed between the

New observation system research group whose needs require developing a new instrument system and the
development group that will eventually handle the system in operational mode. A clear plan
for the transition from research to operational applications should be made.
(j) It is essential to have data management systems that facilitate access, use and
Data management interpretation of the data. Data management should have freedom of access, low
cost, and user-friendly interfaces (directions, catalogues, browsers, metadata on
station history, algorithm accessibility, documentation, etc.). International
cooperation is very important.
6.3 The solar radiation entering the Earth’s atmosphere is the primary input forcing
STATUS OF SELECTED factor for the climate system and its measurement is essential for analysis of
OBSERVATIONS CRITICAL climate. Variations in the magnitude of this irradiance have a direct impact on
FOR CLIMATE CHANGE equilibrium energy levels in the climate system. Satellite systems are optimal for
6.3.1 its measurement because they are above the atmosphere. However, measurement
OBSERVATIONS FOR BASIC of the solar radiation made by multiple satellite systems may show differences
FORCING FACTORS from system to system that are nearly as large as 10 Wm-2 . A difference of 10 Wm-2
6.3.1.1 is quite significant. It exceeds by quite a bit the 1-2 Wm-2 variation in irradiance
Solar radiation due to the sun spot cycle as discussed in Chapter 2. A solar irradiance difference
of 10 Wm-2 is equivalent to a mean radiative forcing factor of 1.75 Wm-2 if the
Earth’s sphericity (a factor of 0.25) and albedo (a factor of 0.7) are taken into
account. The estimated overall change in radiative forcing due to human-
produced greenhouse gas enhancements is of the same order of magnitude.
Clearly, great care is required to obtain a valid (homogeneous) long-term climate
record of solar radiation intensity with calibration to a fixed reference.
A new 20-year record has been produced with appropriate calibration correc-
tions which has less variability and clearly depicts the 11-year sunspot cycle (see
Figure 6.1).
79
Figure 6.1 – Composite total

solar irradiance for 1978 to
1997. The whole time series is
adjusted to the Space Absolute
Radiometer Reference (SARR)
which does not improve absolute
accuracy, but allows comparison
of repeated space experiments
with the same radiometer. [from
Fröhlich and Lean, 1988, with
permission of Robert B. Lee III
and Kluwer Academic
Publishers].
6.3.1.2 Carbon dioxide concentration in the atmosphere has been measured at Mauna
Greenhouse gases Loa Observatory, Hawaii and the South Pole from around 1957 onward. This has
6.3.1.2.1 provided consistent long-term, single-point records of an important climate-
Carbon dioxide forcing parameter. In climate change study it is necessary to expand the
measurement programme to improve our understanding of the global carbon-
cycle budget. This requires measuring spatial variations in carbon dioxide
concentration in the atmosphere to identify the details of sources and sinks with
respect to the biosphere and ocean. Investigation of the sources and sinks of
carbon dioxide will also be facilitated by having measurements for spatial varia-
tions of O2 concentrations. The extremely high accuracy needed for the O2
measurements has restricted these records to the last few years at only a few
stations.
6.3.1.2.2 Both vertical and horizontal distributions of ozone need to be measured in order
Ozone to understand implications for climate change forcing. Total ozone (in a vertical
column) has been measured by surface-based Dobson spectrophotometers and by
satellites. It has been necessary to calibrate Dobson measurements and to adjust
for changes in the calibration factor for satellite measurements. Now it is impor-
tant to develop measurement systems to determine ozone concentrations in the
troposphere. There are also satellite instruments measuring the vertical profiles of
ozone.
6.3.1.2.3 Water vapour is the most important greenhouse gas, accounting for roughly 80
Water vapour per cent of the total greenhouse effect. Water vapour concentration is expected to
increase if temperature increases because of enhanced evaporation from natural
sources. It will be important to observe variations in water vapour concentrations
on a global basis to understand observed climate change trends and to validate
climate model simulations.
Current water vapour measurements are not as accurate as those for other
basic variables of the atmosphere. The slow response to relative humidity of the
sensors in radiosondes means that vertical variation of water vapour are smoothed
out. In addition, processing procedures have changed over time. For instance,
starting in 1993 the calculation for relative humidity from VIZ sondes, which are
part of the global network, was modified to include calculation of humidity values
below 20 per cent and upward adjustments in humidity values over 80 per cent.
The net effect was to increase the measured value of total water vapour and to add
an artificial discontinuity to the time-series record for water vapour. The number
80
of different kinds of radiosonde systems currently used in the global observing

system also makes it necessary to apply calibration adjustments in the data
processing.
Satellite observations are good for showing horizontal variations of humidity;
however, vertical variations are smoothed out even more than with radiosonde
measurements. Intercomparison of satellite measurements with in-situ
(radiosonde) measurements will be required to assure the validity of trend studies
[page 70, Reference no. 7].
6.3.1.3 Improved measurements are needed for aerosol concentrations in the atmos-
Aerosols phere. Such measurement systems must include monitoring the aerosol type and
its size spectrum. Turbidity measurements alone are not sufficient. Monitoring of
source magnitudes including those from biomass burning will be essential for
understanding the budgets for aerosols. It is expected that remote sensing from
satellites will be essential for making the required observations.
6.3.2 Cloud observations are critical for understanding recent and future climate
OBSERVATIONS FOR change. The measurements need to include not only total cloud cover, but also
FEEDBACKS FROM CLIMATE quantative data on the level and composition of the cloud. Ice clouds have quite
SYSTEM COMPONENTS different radiative properties than water clouds. Continuity in the climatology
6.3.2.1 record for clouds has been degraded by changes in observational methods.
Clouds Viewing clouds from the Earth’s surface is quite different from viewing them from
space. Conversion to automated cloud-observing systems introduces a major
discontinuity into the climate record.
The International Satellite Cloud Climatology Project (ISCCP) has been estab-
lished to construct a valid climatology of cloud coverage. Observational records from
the 1980s had shown considerable uncertainty in the observations even when made
by satellite systems. Reductions in percentage cloud cover of approximately three per
cent over a seven-year period were found for both nimbostratus and deep convective
clouds. The changes were primarily in the periods when there was conversion from
one satellite system to another. This conversion effect has now been eliminated by
reprocessing the data. Cloud data are now considered reliable for studies of shorter
term and regional variations of clouds, even if we cannot monitor long-term trends.
Attempts have been made to use surface solar radiation measurements to
give an indication of cloud cover. However, the two are poorly correlated because
of measurement system changes for both methods and because of the interference
effect of air pollution.
6.3.2.2 Measurements of temperature, salinity and currents are required to understand

Oceans oceanic processes and overall how much heat the oceans will store or give up in a
climate change scenario. It is essential to measure the current systems in the deep
ocean as well as in the upper ocean (to depths of several hundred meters) to
describe the heat transports within the ocean which affect sea-surface temperature.
Measurements in the deep ocean will make it possible to monitor the Atlantic ther-
mohaline circulation which is a key factor for identifying decadal-scale shifts in
the ocean circulation. Measurements of temperature in the deeper ocean may
provide a detection of climate change less obscured by seasonal cycle and shorter
period variability. For example Figure 6.2 shows a warming trend between 1957
and 1992 that has been observed in subsurface north Atlantic waters at 24°N and
is most pronounced between depths of 0.7 and 2.5 km with values up to 0.5°C.
In limited areas, comprehensive ocean-measurement networks exist. An
example is the network to support the research, detection, and prediction of El
Niño, the Tropical Ocean Atmosphere (TAO) array in the tropical Pacific Ocean.
This network of moored instrument systems provides temperature and current
measurements in the upper levels of the ocean across the Pacific between 10°S and
10°N. However, that network does not involve deep ocean water. For decadal-scale
variability, long observational records will be required in both the Pacific and
Atlantic ocean areas. In the Atlantic Ocean area, the decadal-scale variability
involves deep water conditions which need to be thoroughly measured.
81
Sea-level observations are needed for several reasons. First, spatial sea-level
variations relate to ocean currents and provide information useful for under-
standing ocean dynamics. Spatial sea-level variations can be measured from
satellites. Second, sea-level changes can have potentially deleterious impacts on
coastal regions. Long term in-situ monitoring is needed to isolate changes due to
climate change from many other factors that can affect sea level in order to make
projections for future impacts.
6.3.2.3 Routine long-term soil moisture observations are woefully scarce over the world.
Surface hydrology Only a few countries such as the Russian Federation have observations that
extend over many decades. Understanding climate change processes over land
areas will require such information.
6.3.2.4 Remote sensing from satellites together with calibration from local surface obser-
Surface land cover vations is providing an important monitoring of land surface changes. This
information helps to define changes in the land surface forcing of the atmosphere
and the response of the terrestrial biosphere to climate variations. It is important
to maintain this observing activity for monitoring and studying climate
change.
6.3.2.5 Satellite observations are important for monitoring snow and ice cover over land,
Cryosphere and sea-ice extent over the oceans. However, many of the records to date are
undocumented with regard to processing procedures and are of short duration. It
will be important to give more attention to data quality and continuity for
climate monitoring purposes.
82
6.3.3 The surface temperature of the Earth is a key descriptor for climate. Surface
OBSERVATIONS FOR CLIMATE station reports are important for providing the observational data as satellites are
RESPONSES incapable of measuring the details of temperature in the surface boundary layer
6.3.3.1 (the lowest several meters in the atmosphere). There are numerous deficiencies in
Surface temperature these observational records. Over land, the density of reporting stations varies by
continent. Africa, Central America, and South America have large areas which are
not adequately covered. Figure 6.3, which presents mean maximum temperatures
reported by surface stations, shows the uneven spacing in station distribution. To
compound the problem, the number of reporting stations is decreasing even in
those areas that are already sparsely represented. As shown in Figure 6.4, the
decrease was more than 10 per cent from 1989 to 1994.
There are many factors which introduce non-uniform spatial and temporal
biases into land surface temperature data. These include changes in instrumenta-
tion, instrument shelters, station location and time of observations. Changing an
observation site from a city to a nearby airport can cause systematic changes in
the temperature due to the ‘heat island’ effect of the urban area. In many cases,
the principles for long-term climate prediction presented in Section 6.2 were not
observed. A common omission has been the overlapping of measurements
between old and new observing systems when the observing system is changed.
Many of the temperature biases introduced by these changes are of magnitudes
similar to those expected with climate change. Thus, it is important to define
these biases and remove them from the climate record.
Sea-surface temperatures have been measured extensively over large parts of
the oceans from ships that voluntarily make temperature measurements
(voluntary observing ships and ships of opportunity), as well as from former ship
weather stations, and from buoy systems. Ships of opportunity and voluntary
observing ships have provided considerable data for more than 100 years in
some parts of the world. However, the method of measurement has changed over
the years from sampling with buckets of water drawn from the ocean to
sampling of water coming into the engine intake. Efforts have been made to
adjust data to a common frame of reference. However, information is lacking
about the details of the measurement so it has been difficult to make precise
corrections to historical data. A good review is provided by Parker, et al., in
Reference no. 2 [pp. 429-470].
Satellites provide virtually complete coverage of the oceans for sea-surface
temperature measurement. It is a challenge to combine this data with that
obtained by in-situ measurements because the satellite is measuring temperatures
at a very thin surface layer (‘skin’ temperatures) whereas in-situ measurements are
for a deeper surface layer. In addition satellite measurements are affected by
atmospheric turbidity which requires adjustments to the data. The recent volcanic
eruption of Mt. Pinatubo caused biases as large as 1°C in the satellite temperature
data.
Figure 6.3 — Climatological

stations with 1961-90 normals
data for mean maximum
temperature. All station data has
been received by the Climate
Research Unit at the University
of East Anglia from National
Meteorological Agencies. A total
of 6 632 locations are shown.
[from page 418, Reference
no. 2, with permission of
Kluwer Academic Publishers].
83
Figure 6.4 — Number of

‘CLIMAT’ messages (containing
monthly temperature summaries)
received at the UK Met. Office
through the Global Telecommu-
nications Systems for the period
of 1977-1993. [from page 72,
permission of Kluwer Academic
Publishers].
6.3.3.2 Rainfall is a very important climate parameter for all forms of life on land.
Precipitation Nevertheless, its measurement is very inadequate. Precipitation has much local-
scale variability which makes it difficult to find measurements representative for
an area. Some of this variability is due to topographic effect. Rain gauge meas-
urements are quite sensitive to rain gauge design, to wind conditions and to
whether the precipitation is rain or snow. For snow conditions gauges can seri-
ously underestimate precipitation amounts. Changes in rain gauge design and
observing practices which have been common throughout the world result in a
major task for producing homogeneous climate data. Figure 6.5 shows a sampling
of changes that have been made throughout the world with estimates on the
biases introduced into the data.
Satellite and radar sensing systems provide additional information on precip-
itation, and this information should become better quantified with time. However,
these data are available only for short times and the radar data has limitations in
the observations of diurnal cycle variability. It will be essential to calibrate and
combine gauge and remotely-sensed data to get long-term climate records. The
gauge data will remain important as calibration for the remote-sensed data.
For reasons mentioned above, there is considerable uncertainty in climate
records of precipitation. Nevertheless, there are first estimates of precipitation
trends as reported by IPCC [pp. 152-156, Reference no. 3]. For example, analysis
of data suggests an increase in precipitation in high northern latitudes during this
century and an increase of winter precipitation in northern mid-latitudes. It will
require considerable effort to obtain climate records for precipitation that will be
useful for identifying trends due to climate change.
6.4 A number of steps are being taken to improve long-term climate monitoring. A
STRATEGIES FOR few are briefly presented here.
IMPROVING LONG-TERM
CLIMATE MONITORING There are ongoing efforts to gather data from archives throughout the world and
6.4.1 to render it into a form that allows it to be easily accessed and incorporated into
DATA RECOVERY AND the climate data files. Digitizing hand written records is an important aspect of
RECALIBRATION this work. For all such data it is also necessary to study existing metadata (docu-
(‘REHABILITATION’) mentary records about site changes, exposure details, etc.) to correct for artificial
biases in the data. WMO has its Data Rescue programme (DARE), to help in this
area.
6.4.2 Construction of analysis (spatial distribution maps) from data is important for
REANALYSIS describing and understanding processes in the climate system. The process of
analysis also helps to fill in where observational data is missing and to provide the
information database needed for numerical model studies.
Numerical models themselves are important components of the analysis
process. Adjustments in the patterns that evolve in numerical models as observa-
tional information is introduced achieves a ‘four dimensional data assimilation’
whereby best-estimate and internally consistent initial analyses for prediction
models are produced from observational data. This has been the practice since the
advent of operational Numerical Weather Prediction (NWP) in the 1950s. The
improvement of NWP models has improved the quality of the analysis fields
representing observations.
84
Figure 6.5 — Time-varying biases

of precipitation measurement for
various countries (from Karl et
al., 1993. [from page 77,
permission of Kluwer
Academic Publishers].
Changes of the model formulation over time have led to variations in the
quality of the analysis and have introduced artificial time-varying biases into the
data. In order to obtain an observation analysis record that has a uniform (homo-
geneous) frame of reference in time, reanalysis has been conducted using the
same numerical prediction model. For the purposes of NWP, a reanalysis has been
carried out to obtain a consistent dataset of atmospheric observations over a
significant portion of the world from around 1957.
General reanalysis is currently being done by two programmes. In the United
States, the Environmental Modelling Center of the National Centers for
Environmental Prediction (NCEP) together with the National Center for
Atmospheric Research (NCAR) are using a forecast model that became operational
in 1994 to reanalyse initial conditions for daily data at many levels in the atmos-
phere. Their plan is to go back to 1957. In Europe, the European Centre for
Medium-Range Weather Forecasts (ECMWF) is also conducting a reanalysis of
daily data at many atmospheric levels using one of their recent comprehensive
85
forecast models. Their reanalysis period starts in 1979. These reanalysis projects
are of central interest to the World Climate Research Programme (WCRP). The
reanalysis data will greatly facilitate long-term climate monitoring. Further
reanalysis information is given in WMO (1997).
6.4.3 The increase of in-situ measurements for greenhouse gases, aerosols and ozone is
INCREASING THE NUMBER OF considered a relatively low-cost enhancement of great value for observations
MEASUREMENTS directly related to climate change. In particular, increases in the number of places
where flask measurements for CO2 concentration are obtained are needed to
provide validation points for future satellite measurements. More in-situ measure-
ments of aerosol chemical and physical characteristics would greatly benefit
determination of its radiative properties. An increased number of balloon-borne
measurements of ozone and water vapour in the stratosphere would provide
important validation data for future satellite measurement.
6.4.4 Satellites will play a key role in new measurements for climate monitoring. Many
NEW MEASUREMENTS SYSTEMS new systems are being developed for satellites to improve measurements of the
radiation budget, clouds, trace gases, surface changes on Earth and so forth.
Careful attention will have to be given to calibration of the data and overlapping
measurements between different satellite systems.
The Earth Observing System (EOS) programme of the U.S. National
Aeronautics and Space Administration (NASA) is a good example of the advance-
ment in remote sensing from satellites. The EOS programme is developing over 20
new satellite systems to obtain or improve measurements of many climate system
parameters (NASA, 1995). The list includes: solar radiation (total irradiation and
ultraviolet component only); atmospheric water vapour vertical distribution;
images of the land surface, water, ice and clouds; Earth radiation budget; topog-
raphy of the sea surface and ice sheets; flux of trace gases (including carbon
dioxide) at the air-sea interface; global aerosol distribution; cloud properties such
as optical thickness; cloud heights; planetary boundary layer heights; global
distributions of numerous trace gases and aerosols in the upper troposphere,
stratosphere and mesosphere; location and radiant energy of lightning flashes;
precipitation rate; cloud water; sea-surface temperature; soil moisture; angular
solar reflectance functions from the top of the atmosphere, clouds and the Earth’s
surface; biological processes (chlorophyll concentration, vegetation productivity,
etc); tropospheric pollution; and sea-surface winds.
Innovative approaches for temperature measurements are being considered.
One idea is to monitor global temperature changes by measuring the Earth’s elec-
trical fields. Overall electric potential variations with height are related to the
number and intensity of thunderstorms in the world. To the extent that thun-
derstorm activity relates to temperature, measurement of electric potential would
provide a measure of temperature conditions. Another idea is that the measure-
ment of temperatures at great depths (like 600 m) below the surface of the Earth
could provide signals for long-term temperature changes. This is already being
done in the ocean as shown in Figure 6.4. Finally, the travel time for acoustic
waves in the ocean over great distances may provide useful information on
temperature changes in the deep ocean since the sound speed depends, in part,
on temperature.
86
CHAPTER 7
MODELLING, DETECTION AND ATTRIBUTION OF RECENT

AND FUTURE CLIMATE CHANGE
7.1 The projections for future climate change are continually updated as climate
INTRODUCTION models are improved observational records are expanded, understanding of the
climate system is improved and estimates of the climate forcing due to human
activity are refined. All of these areas involve complex considerations and uncer-
tainties. A large number of research groups throughout the world are focused on
improving the projections for future climate through the coordination efforts of
the Intergovernmental Panel on Climate Change (IPCC), the World Climate
Research Programme (WCRP) and other substantive programme activities of the
World Meteorological Organization (WMO) and the International Council of
Scientific Unions (ICSU).
The results published at any one time will be superseded by new results. The
differences between successive conclusions may be confusing to those outside the
community working on climate change. One must appreciate that the changes in
conclusions are incremental and part of a coordinated approach to finding answers.
Intercomparisons among many model experiments involving long-term
simulations are required to establish a meaningful understanding of their charac-
teristics, sensitivities and uncertainties for any given specification of human
climate-forcing impact. As a result, only a limited number of forcing scenarios can
be modelled and it takes time for the gains made in climate-change projections to
become apparent as models and forcing specifications are improved.
This chapter provides a perspective on the scope and type of conclusions that
are being reached. The material presented is based primarily on the material in
chapters 6 and 8 of Reference no. 3. Current work and conclusions are overviewed
here with full recognition that the details will be different in the future. Some
updates from chapter 9 of the 2001 IPCC report (Reference no. 7) are also
included.
This chapter includes discussion of recent climate change because that is the
direct antecedent for calibration and understanding of projections for the future.
It first presents modelling results and then moves to the detection (analysis of
observations) and attribution (understanding the causes) for climate change.
Attribution is the key factor for isolating human-produced effects from natural
changes in the climate system.
7.2 Figure 3.5 in Chapter 3 identifies the primary anthropogenic impacts on the
MODEL RESULTS FOR climate system up to the present. Initial model simulation studies considered only
CLIMATE CHANGE the carbon dioxide component. More recent model simulation studies have
7.2.1 included the sulphate aerosol component. Note that in the description of these
RECENT CLIMATE CHANGE experiments, the carbon dioxide concentration used may be an equivalent
concentration to represent all of the greenhouse gases.
Model simulations for recent global mean temperature changes in the last
45 and 120 years have already been shown in Chapters 4 and 5, respectively.
The predicted global mean temperature increases when the observed increases in
both carbon dioxide and sulphate aerosol are included in the model (see Figure
5.5). The sulphate aerosol itself leads to a cooling effect which partially offsets
the warming due to carbon dioxide. Earlier model studies which had not
included sulphate aerosols gave larger values for predicted warming. The general
upward trend appears quite realistic when compared with observations. The
observed temperature record shows much more variability within the 1860-1990
period than is simulated in the model. However, the model result has been
smoothed over time. Yearly mean model values would show similar variability.
87
Source DJF* MAM* JJA* SON* Year
Simulated, increase in equivalent CO2 since 1900 –0.46 –0.35 –0.08 –0.34 –0.29
Simulated, aerosol forcing and equivalent CO2 –0.43 –0.27 –0.16 –0.32 –0.27
increase since 1990
Observations of recent change –0.28 –0.17 –0.19 –0.36 –0.19
(1981 to 1990 mean less 1951 to 1981 mean)
* DJF = December, January, February; MAM = March, April, May; JJA = June, July, August; SON = September, October, November.
Table 7.1 — Changes in diurnal Diurnal temperature range decreases are predicted by theory when the green-
range of 1.5 m temperature house gases and sulphate aerosols are increased. Model results show a decrease in
averaged over seasons and the diurnal temperature range since 1900 over Northern Hemisphere continental areas
annual cycle. The simulated and as observed increases in carbon dioxide and sulphate aerosols are included in the
observed values are averaged over model. The observations also show decreases in diurnal temperature range (see Table
the regions where observations are 7.1). However, other processes also affect diurnal temperature ranges such as cloudi-
available. The observed data are ness and surface evaporation. The relative importance of all these effects is still
from Horton (1995) and the uncertain.
simulations from Mitchell et al. In summary, model simulations generally have suggested that the human
(1995). The changes in greenhouse production of carbon dioxide and sulphate aerosols, starting with the industrial
gas forcing (represented by an era, has already caused climate change in terms of global warming. Recent numer-
equivalent increase in CO2) and ical model simulations have given estimates for the global-mean warming that
direct sulphate aerosol forcing are range from 0° to +1.6°C, as shown in Table 7.2. These results imply that such
those estimated to have occurred changes are currently ongoing and that the atmosphere and the climate system
since 1900. Note that the observed currently are not in equilibrium. As reported in the 2001 IPCC report (Reference
changes are available only over the no. 7) recent studies further confirm this conclusion.
latter half of this period and are the This pre-existing change and current lack of equilibrium must be considered
difference between the mean for when designing model experiments for future climate change that start from
1981 to 1990 and the mean for present conditions. The year 1990 has often been used as the initial time for
1951 to 1980. [from page 295, future climate-change simulations. If the climate model is defined as being in
Reference no. 3]. equilibrium before starting the future climate-change experiment, the initial rate
of simulated temperature increase will be erroneously suppressed as the model
develops the radiative imbalance conditions. This is called the ‘cold start’
Study Sensitivity to Direct aerosol Temperature response Temperature response of

doubling CO2 (°C) forcing (Wm-2) of equilibrium due equilibrium to combined aerosol
to aerosols (°C) and CO2 forcing since 1900 (°C)
S1 Roeckner et al. (1995) 2.8 -0.7 –0.9* 0.5

S2 Taylor and Penner (1994) 5.2# –0.9 –0.9 0.6+
S3 Mitchell et al. (1995a) 5.2 –0.6 –0.8 1.6
S4 Le Treut et al. (1995) 3.9 –0.3 (direct) –1.6$ 0.0$
–0.8 (indirect)
Table 7.2 — Equilibrium global S1 and S3 use the aerosol distribution of Langer and Rodhe (1991) and represent aerosols as an
mean response to the increase in increase in surface albedo. S2 derives the sulphate loading from a coupled atmosphere sulphur
greenhouse gases and sulphate cycle model and includes an explicit radiative scattering treatment of aerosols.
aerosol concentrations over the * Assuming 40 per cent increase of CO2 gives 50 per cent of the warming due to
20th century. The experimental doubling, and subtracting this value from the combined forcing experiment in the final
designs in the four studies differ, column.
so some of the entries are derived # Assuming 25 per cent increase in CO2 gives 29 per cent of the warming due to doubling.
under specific assumptions + Using a 25 per cent increase in CO2, whereas S1 and S3 use a 40 per cent increase to
defined below. [from page 293, allow for changes in all greenhouse gases.
Reference no. 3]. $ The forcing used includes both the estimated direct and indirect forcing. In the last
column, CO2 was increased by 25 per cent. Substantially higher sensitivities are found if
a colder simulation is used. Although the global mean temperature change in S4 is zero,
the model gives a cooling in the Northern Hemisphere and a warming in the Southern
Hemisphere.
88
CHAPTER 7 — MODELING, DETECTION, AND ATTRIBUTION OF RECENT AND FUTURE CLIMATE CHANGE
(a)
(b) (c)
Gross forcing (Wm–2)
Temperature response
Net forcing (Wm–2)

T
Actual a Actual
Actual Cold
Model start Model
Model
1900 1990 1900 1990 1900 1990
Year Year Year
Figure 7.1 — Schematic diagrams of radiative forcing and temperature response, showing the effect of neglecting the effect of past
forcing (the ‘cold start’ problem). Panel (a) shows forcing due to increases in CO2, with a rate of increase rising gradually to 1990 as
observed, and maintained at one per cent/yr thereafter (left curve) and from a one per cent/yr increase starting abruptly in 1990, as in
idealized experiments (right curve). Panel (b) shows temperature response to the forcing in (a). The upper curve, which is the response
in the case with the gradual initial increase as observed, has been transposed vertically to zero at 1990 (dashed curve) to highlight the
initial slow response in the case of an abrupt increase used in idealized experiments (lower curve). The difference between the two
curves is known as ‘the cold start’ and is an artefact of the experimental design. Panel (c) shows the net forcing (which allows for the
increased loss of radiation to space as the model warms) in the case with a gradual start to the forcing. The lower curve shows the net
forcing (upper curve) in the idealized case. Note the net heating at 1990 which maintains the warming of the ocean in the upper curve
in (b). To heat 300m depth of water by 0.3°C/decade (typical of the AOGCM experiments in climate change assessments) requires a
net heating of 1.5 Wm-2 (cf. 4 Wm-2 for a doubling of CO2) which takes several decades to build up with a 1 per cent/yr increase in
CO2. [from page 313, Reference no. 3].
problem. It is estimated that the lag effect on predicted temperature changes will
last for several decades, a time scale determined by the adjustment time of the
oceans. Rates of temperature change would be underestimated during this time as
shown in Figure 7.1.
7.2.2 A range of climate model predictions have been made. The largest group dealt
FUTURE CLIMATE CHANGE with greenhouse-gas impacts alone. In the description of these experiments, as
noted before, reference may be made to the changing concentration of carbon
dioxide alone, or to an equivalent carbon-dioxide concentration, that is calcu-
lated to represent the effect of all the greenhouse gas changes attributable to
human activity. Experiments have been carried out to examine the effects of
doubling equivalent carbon dioxide concentrations either all at once or gradually
with rates of increase varying from 0.25 per cent per year to 4 per cent per year.
Note that the current observed rate of increase of carbon dioxide itself is about
0.5 per cent per year, and the overall increase from the pre-industrial period to
1990 has been 26 per cent (see Table 3.2). A current rate of 1.0 per cent per year
3
Global temperature change (°C)
Fig. 7.2 — Comparison between

2
several AOGCM simulations
(climate sensitivities between 2.1
and 4.6°C) and two versions of Coupled AOGCMs
the simpler UD/EB-type models 1 GISS (k)
CSIRO (d)
MRI (p)
(with climate sensitivities of UKMO (t1)
UKMO (t2)
2.5°C and 2.2°C). All models GFDL (j)
BMRC (a)
were forced with one per cent/yr UKMO (s)
0 NCAR (r)
(compound) increase of COLA (c)
atmospheric CO2 concentration UD/EB models
Section 6.3
from equilibrium or near- Section 7.5.3
equilibrium in 1990. [from page -1

0 20 40 60 80 100
300, Reference no. 3]. Year from start of experiment
89
for equivalent carbon dioxide is close to reality. More recently, experiments have
included sulphate aerosols and other greenhouse gases more explicitly. Other
experiments have looked at factors of two and four for carbon dioxide increases.
Projections have been generally made to about the year 2100, although some
studies have gone to the year 2500. Many predictions have started with 1990
conditions; others have gone back to pre-industrial time, eliminating the ‘cold
start’ problem.
7.2.2.1 Results from experiments with fully-coupled ocean-atmosphere climate models

Mean conditions have shown large differences in the simulations for climate change impacts even
for globally-averaged surface temperature. A comparison was made among ten
models for the case where (equivalent) carbon dioxide concentration was
increased at the rate of one per cent per year starting from 1990 values and the
model conditions were initially in equilibrium. This rate of increase gives a
doubling of CO2 in about 70 years.
Results for global surface temperature are shown in Figure 7.2. Also shown
are results from the simpler type of climate model labelled (UD/EB) discussed in
chapter 4. The warming after 70 years ranges from 1.5° to 3.8°C. This variability of
model response has led to a calibration descriptor for climate models called
‘climate sensitivity.’ Climate sensitivity is defined as the increase in the equilib-
rium value of global mean surface temperature produced by the model with a
doubling of carbon dioxide concentration. The climate sensitivity value is larger
than the values shown in Figure 7.2 after 70 years because the 70-year value is not
an equilibrium value. It would be necessary to run the model for many years with
the CO2 concentration held constant at its doubled value in order to get the equi-
librium value of temperature.
Differences between the model results are even more dramatic for spatial
patterns and local/regional temperature change values. Figure 7.3 shows a
comparison of annual-mean temperature spatial patterns between two of the
models shown in Figure 7.2 at the time the CO2 reaches a doubled value. Note
that the increases tend to be larger over land than over water and larger at higher
latitudes in both models. However, differences in the details between the models
can be large as is seen over northern Africa.
Some experiments for CO2 concentration doubling and with sulphate
aerosol effects included have simulated mean temperature decreases in regional
areas in both summer and winter seasons. These areas included parts of China and
the United States where aerosol concentrations were high. However, the uncer-
tainty in results for regional areas is large, as shown by the range of values in the
model intercomparisons in Figure 5.6 and Figure 7.4. Differences in predicted
changes of seasonal mean temperatures ranged up to 5°C. For the same simula-
tions, predictions of seasonal mean precipitation changes differred by up to two
mm/day. In some cases, different models even predicted changes of opposite sign.
90
Figure 7.4 — Simulated regional

changes from 1880-1889 to
2040-2049 (experiments x,y) or
from pre-industrial to 2030-2050
(experiments w,z). Experiments x
and w include greenhouse gas
forcing only, whereas y and z also
include direct sulphate aerosol
effects. The x, y, w and z
indicators are in the model list at
the bottom of the figure. (a)
Temperature (December to
February); (b) Temperature (June
to August); (c) Precipitation
(December to February); (d)
Precipitation (June to August); (e)
Soil moisture (December to
February); (f) Soil moisture (June
to August). CNA=Central North
America; SEA=South East Asia;
SAH=Sahel; SEU=Southern
Europe; AUS=Australia. [from
page 306, Reference no. 3].
Seasonal soil moisture predictions were also in varying directions with differences
as large as three cm.
Note that the model simulations discussed in the previous paragraph cover
different time periods and have different CO2 concentration variations than do
the model simulations discussed in the two preceding paragraphs. This does not
change the overall characteristics discussed here.
Climate-change assessments have been made with the climate models using
standardized projections for the changes in anthropogenic forcing of the climate
system. For the 1995 IPCC reports six emission scenarios established in the 1992
IPCC report were used in the model studies. The assumptions for these scenarios
(referenced IS92a-f) are described in Table 7.3. The scenarios cover the period from
1990 to 2100 and provide examples of low (IS92c), medium (IS92a), and high
(IS92e) impacts.
A new set of 40 emission scenarios (referenced SRES for ‘Special Report on
Emission Standards’) was established in 2000 (Nakić enović et al., 2000) to define
future projections for anthropogenic emissions. Of these, 35 scenarios which
contain data on the full range of gases required for climate modelling, define the
full set of scenarios used for climate-change projections. A subset of these was
used in studies for the 2001 IPCC report. These scenarios generally define the
range of possibilities in terms of demographic, economic and technological devel-
opment. A group of six ‘marker’ or ‘illustrative’ scenarios from this set serves to
describe the primary range of uncertainty for future projections. These six scenar-
ios are highlighted in Table 7.4 (from Chapter 9, Reference no. 7). Basically the A
categories have larger emission outcomes than the B categories. The three subsets
of the A1 category (A1FI, A1T, and A1B) refer to fossil fuel intensive, non-fossil
fuel sources, and reliance not on only one energy source cases, respectively. The
projections for total radiative forcing out to 2100 for the six SRES ‘marker’ cate-
gories along with three of the scenarios examined in the IPCC 1992 studies (i.e.
IS92c, IS92a, and IS92e for low, medium and high impacts, respectively) are
shown in Figure 7.5.
Note that the IPCC Working Group II, which examined impacts, adapta-
tions, and mitigations of climate change, has adapted revised anthropogenic
forcing scenarios based on anticipated new efficiencies for anthropogenic
energy production systems that will cut back on increases in carbon dioxide
production. See the 1995 IPCC Working Group II report (pp. 47-50, Reference
91
92
Table 7.3 — Summary of Assumptions in the Six IPCC 1992 Alternative Scenarios†. (from page 12, IPCC, 1992a).
Scenario Population Economic growth Energy supplies†† Other††† CFCs
IS92a World Bank 1991 1990-2025: 2.9 per cent. 12.000 EJ conventional oil. Legally enacted and internationally Partial compliance with Montreal
11.3 B by 2100. 1990-2100: 2.3 per cent. 13.000 EJ natural gas. agreed controls on SOx, NOx and Protocol. Technological transfer
Solar costs fall to $0.075/kWh. NMVOC emissions. results in gradual phase out of
191 EJ of biofuels available at CFCs, also in non-signatory
$70/barrel. countries by 2075.
IS92b World Bank 1991 1990-2025: 2.9 per cent. Same as ‘a’. Same as ‘a’ plus commitments by Global compliance with
11.3 B by 2100. 1990-2100: 2.3 per cent. many OECD countries to stabilize scheduled phase out of Montreal
or reduce CO2 emissions. Protocol.
IS92c UN Medium Low Case 1990-2025: 2.0 per cent. 8.000 EJ conventional oil. Same as ‘a.’ Same as ‘a’.
6.4 B by 2100. 1990-2100: 1.2 per cent. 7.300 EJ natural gas.
Nuclear costs decline by 0.4 per
cent annually.
IS92d UN Medium Low Case 1990-2025: 2.7 per cent. Oil and gas same as ‘c’. Emissions controls extended worldwide CFC production phase out by
6.4 B by 2100. 1990-2100: 2.0 per cent. Solar costs fall to $0.065/kWh. for CO, NOx, NMVOC and SOx. Halt 1997 for industrialized countries.
272 EJ of biofuels available at deforestation. Capture and use of Phase out of HCFCs.
$50/barrel. emissions from coal mining and gas

production and use.
IS92e World Bank 1991 1990-2025: 3.5 per cent. 18.400 EJ conventional oil. Emissions controls (30 per cent Same as ‘d’.
11.3 B by 2100. 1990-2100: 3.0 per cent. Gas same as ‘a’. pollution surcharge on fossil energy).
Phase out nuclear by 2075.
IS92f UN Medium High Case Same as ‘a’. Oil and gas same as ‘e.’ Same as ‘a’. Same as ‘a’.
17.6 B by 2100. Solar costs fall to $0.083/kWh.
Nuclear cost increase to $0.09/kWh.
† The assumption for the 1990 Scenario A are described in IPCC (1990) Annex A. pp. 331-339.
†† All scenarios assume coal resources up to 197 000 EJ. Up to 15 per cent of this resource is assumed to be available as $1.30/gigajoule at the mine.
††† Tropical deforestation rates (for closed and open forest) begin from an average rate of 17.0 million hectares/year (FAO, 1991) for 1981-1990, then increase with population until
constrained by availability of land not legally protected. IS91d assumes an eventual halt of deforestation for reasons other than climate. Above-ground carbon density per hectare varies
with forest type from 16 to 117 tons C/hectare, with soil C ranging from 68 to 100 T C/ha. However, only a portion of carbon is released over time with land conversion, depending
on type of land conversion.
A1. The A1 storyline and scenario family describe a future world of very rapid economic growth, global population that peaks
in mid-century and declines thereafter, and the rapid introduction of new and more efficient technologies. Major underlying
themes are convergence among regions, capacity building and increased cultural and social interactions, with a substantial
reduction in regional differences in per capita income. The A1 scenario family develops into three groups that describe alter-
native directions of technological change in the energy system. The three A1 groups are distinguished by their technological
emphasis: fossil intensive (A1FI), non-fossil energy sources (A1T), or a balance across all sources (A1B) (where balanced is
defined as not relying too heavily on one particular energy source, on the assumption that similar improvement rates apply to
all energy supply and end-use technologies).
A2. The A2 storyline and scenario family describe a very heterogeneous world. The underlying theme is self-reliance and preser-
vation of local identities. Fertility patterns across regions converge very slowly, which results in continuously increasing
population. Economic development is primarily regionally oriented and per capita economic growth and technological change
are more fragmented and slower than in other storylines.
B1. The B1 storyline and scenario family describe a convergent world with the same global population, that peaks in mid-
century and declines thereafter, as in the A1 storyline, but with rapid change in economic structures toward a service and
information economy, with reductions in material intensity and the introduction of clean and resource-efficient technologies.
The emphasis is on global solutions to economic, social and environmental sustainability, including improved equity, but
without additional climate initiatives.
B2. The B2 storyline and scenario family describe a world in which the emphasis is on local solutions to economic, social and
environmental sustainability. It is a world with continuously increasing global population, at a rate lower than A2, intermediate
levels of economic development, and less rapid and more diverse technological change than in the B1 and A1 storylines. While
the scenario is also oriented towards environmental protection and social equity, it focuses on local and regional levels.
Table 7.4 — The Emissions no. 4) for a description of the low CO2-emitting energy supply system (LESS)
Scenarios of the Special Report on scenarios.
Emissions Scenarios (SRES). (from The uncertainty in climate-model predictions for climate change is large. The
page 554, Reference no. 7). sources for this uncertainty include the uncertainty in the forcing to be prescribed
for the system (as shown in Figure 7.5), and uncertainty due to model formula-
tion (as demonstrated by the range in climate sensitivities in the model
intercomparisons shown above in Figure 7.2). Note that the uncertainty in the
forcing specification is even larger than that due to the model formulation.
As a result of these uncertainties, it has been necessary to examine a wide
range of forcing scenarios. A simple reference climate model of the UD/EB type
described in Chapter 4 has been used to make an overall initial analysis for the
many cases to determine general characteristics of climate change impacts
because of limitations in the computer resources needed to use the comprehen-
sive climate models for so many experiments. The comprehensive climate
(a) 10
A1FI
9 A1B
A1T
Figure 7.5 — Estimated 8 A2
B1
historical anthropogenic 7 B2
Forcing (Wm−2)
radiative forcing (Wm-2) IS92a

6 IS92c Model ensemble
followed by projections for future IS92e all SRES
envelope
forcing based on the six SRES 5
marker scenarios and three of the
4
scenarios from the 1992 IPCC
(IS92a, IS92c, IS92e). Also 3
shown is the total range of 2 Bars show the
variation (shaded area) for the range in 2100
1 produced by
35 SRES scenarios that could be everal models
used for climate modeling 0
studies. (from Figure 9.13a on 1800 1900 2000 2100
page 554 in Reference no. 7). Year
93
Figure 7.6 — Simple model (b) 7

results for the estimated historical A1FI
A1B Several models
anthropogenic global-mean 6 A1T all SRES
envelope
Temperature Change (°C)

surface temperature change (°C) A2
followed by projections for future 5 B1
B2 Model ensemble
all SRES
changes based on the six SRES IS92a envelope
marker scenarios and middle- 4
range of the scenarios from the
1992 IPCC (IS92a) using a 3
simple climate model tuned to
seven ocean-atmosphere climate 2
models. Also shown is the total
Bars show the
range of variation (shaded area) 1 range in 2100
produced by
for all SRES. (from Figure 9.13b several models
on page 554 in Reference no. 7). 0
1800 1900 2000 2100
Year
models are used to examine the details of climate change for selected cases. The
UD/EB model was calibrated using the comprehensive climate models and set to
represent a model with ‘climate sensitivity’ of 2.5°C. The global-mean surface
temperature change predictions out to 2100 for the scenarios used in Figure 7.5
(except for IS92a and IS92e) are shown in Figure 7.6. These figures also show the
full range of variability ‘envelopes’ for the full set of 35 SRES scenarios described
above.
The UD/EB-type model has also been used to examine other possible impacts
due to projected future increases in the greenhouse gases. As an example, possible
impacts on the thermohaline circulation in the oceans were studied in experi-
ments that required 1000-year model simulations (see Stocker and Schmittner,
1997). Use of the UD/EB-type model made it possible to perform the set of very
long simulations needed for the study. Stocker and Schmittner showed an
example where the thermohaline circulation change depended on the rate of
increase of greenhouse-gas concentration in the atmosphere. A slow rate of
increase to a final enhanced value caused the thermohaline circulation to be
reduced. A rapid rate of increase to the same final enhanced value resulted in a
total cessation of the thermohaline circulation.
Studies with more sophisticated models reported in the 2001 IPCC report
(Reference no. 7) still show considerable variation among models for the same
forcing scenario, a doubling of CO2 in 70 years. See Figure 7.7 for a comparison
of the results from 19 models for both temperature and precipitation changes.
The change in global mean temperature ranges from +1.1 to +3.1°C and the
change in global mean precipitation ranges from -0.2 to +5.6 per cent. This vari-
ability is similar to that reported in the 1996 IPCC studies shown in Figure 7.2
before.
7.2.2.2 There have been climate-change analyses for many specific aspects of the climate
Variability system other than mean conditions. Studies have generally been with individual
climate model simulations and not the whole group of climate models. Thus,
results are quite preliminary.
The decrease of diurnal variability over continental areas has already been
discussed. For longer-term variability, it is sufficient to reproduce the essence of
the summary comments from page 330 in the 1995 IPCC report (Reference No. 3).
(i) Experiments with different model configurations indicate that zonal-mean,
mid-latitude, intermonthly temperature variability may be reproduced, but
there are no consistent results in regard to changes in persistent anomalies
called ‘blocks,’ one of the contributors to intermonthly variability.
Generalization of climate change impacts from these experiments is difficult
since different definitions of blocking have been used and different models
show different changes of geographical patterns of blocking.
94
ARPEGE/OPA2 CSIRO Mk2 GFDL_R15_a HadCM3

BMRCa CSM 1.0 GFDL_R30_c IPSL-CM2
CCSR/NIES DOE PCM GISS2 MRI1
CCSR/NIES2 ECHAM3/LSG GOALS MRI2
CGCM1 ECHAM4/OPYC HadCM2 Mean
(ii) One study shows that intermonthly and interannual variability differ-
ences between GCMs with a simple mixed-layer (ocean model) and those
coupled to a full ocean model are larger than changes in either type of
model due to increased CO2 alone, pointing to the importance of using
a model with some representation of ENSO-like phenomena (mixed layer
ocean models cannot represent ENSO processes).
(iii) ENSO-like variability in Sea-Surface Temperatures (SSTs) found in several
Atmospheric-Ocean General Circulation Models (AOGCMs) still exists
with increased CO2 conditions. The variability shows either little
change or a slight decrease in the eastern tropical Pacific Ocean.
(iv) Precipitation variability associated with ENSO events increased in the
climate change simulations, especially over the tropical continents. It
was suggested that this could be associated with the mean increase of
tropical SSTs.
(v) Several models indicate enhanced interannual variability of area-aver-
aged summer rainfall in the South Asian monsoon.
(vi) In a number of simulations decadal and longer time-scale variability
obscured the signal of climate change in the rates of global warming and
patterns of zonal mean temperature change. In one case multi-century
variability in ENSO phenomena was as large as the mean change caused
by CO2 increase. This indicates that natural climate variability on long-
time scales will continue to be problematic for CO2 climate change
analysis and detection.
7.2.2.3 Extreme event statistics are an important aspect of climate. Predicting such events
CHANGES IN EXTREME requires predicting changes in the probability distributions. Changes in variabil-
EVENTS ity will strongly affect the occurrence of extreme events and, in some cases, even
more than would result from changes in mean values. For instance, decreasing
diurnal variability would be expected to decrease the probability of extreme
temperature events (hot and cold). Global climate models do not have the reso-
lution needed to predict most extreme events. Techniques such as statistical
downscaling and high resolution regional climate models are needed. These tech-
niques were discussed earlier. A summary of some observed and modelled changes
listed in the 2001 IPCC report (Reference no.7) are shown in Table 7.5.
95
Confidence in observed changes Changes in Phenomenon Confidence in projected changes

(latter half of the 20th century) (during the 21st century)
Likely Higher maximum temperatures and Very likely
more hot days a over nearly all land areas
Very likely Higher minimum temperatures, fewer Very likely
cold days and frost days over nearly
all land areas
Very likely Reduced diurnal temperature range Very likely
over most land areas
Likely, over many areas Increase of heat index b over land areas Very likely, over most areas
Likely, over many Northern Hemisphere More intense precipitation events c Very likely, over many areas
middle to high latitude land areas
Likely, in a few areas Increased summer continental drying Likely, over most mid-latitude
and associated risk of drought continental interiors. (Lack of
consistent projections in other areas)
Not observed in the few analyses available Increase in tropical cyclone peak wind Likely, over some areas
intensities d
Insufficient data for assessment Increase in tropical cyclone mean and Likely, over some areas
peak precipitation intensities d
a
Hot days refers to a day whose maximum temperature reaches or exceeds some temperature that is considered a critical threshold
for impacts on human and natural systems. Actual thresholds vary regionally, but typical values include 32°C, 35°C or 40°C.
b
Heat index refers to a combination of temperature and humidity that measures effects on human comfort.
c
For other areas, there are either insufficient data or conflicting analyses.
d
Past and future changes in tropical cyclone location and frequency are uncertain.
Table 7.5 — Estimates of Some of the examples below from the 1995 IPCC report (Reference no. 3)
confidence in observed and show where changes in extreme weather events were inferred from overall
projected changes in extreme weather or wind pattern changes in the experiments with CO2 global warming.
weather and climate events. As in the previous section, these examples are intended to illustrate areas of inter-
(from page 575, Reference no. 7). est and not areas where conclusions of high certainty have been reached.
7.2.2.3.1 One source of extreme wind events in the middle latitudes is synoptic-scale
Wind storms. The occurrence and intensity of such storms relates to baroclinic field
intensity and moisture supply. Changes in middle-latitude storm intensity and
tracks have been examined in some climate-change simulations. There has been
some evidence in the models of storm tracks being displaced poleward and
storms being more intense as global warming occurred.
Tropical cyclones are the primary cause of extreme wind events (as well as heavy
rain events) in tropical and subtropical regions. Some global climate models can
simulate aspects of tropical cyclone occurrence. However, analysis of global warming
simulations has not shown any clear-cut patterns of changes in frequency, area of
occurrence, time of occurrence, mean intensity, or maximum intensity of tropical
cyclones. Recall that the observations for Atlantic tropical storm activity have shown
no systematic change in the past 50 years (see Figure 2.13).
The 1995 IPCC report goes on to highlight the following problems for clari-
fying climate change impacts on tropical cyclones:
(i) Tropical cyclones cannot be simulated adequately in present general circula-
tion climate models.
(ii) Some aspects of ENSO are not simulated well in general circulation climate
models.
(iii) Other large-scale changes in the atmospheric general circulation which could
affect tropical cyclones such as jet stream activity cannot yet be discounted.
(iv) Natural variability of tropical storms is very large, so small trends are likely
to be lost in the noise.
7.2.2.3.2 The effects of doubling the concentration of CO2 in climate models have been
Temperature analysed in terms of changes in daily maximum and minimum temperatures.
Changes of up to 10°C were found in regions over land areas. The larger changes
96
were related to modelled alterations in snow cover, soil moisture, and cloud
cover.
An analysis of climate-model simulations for Victoria, Australia, in a global
warming experiment, clearly documented the large change in occurrences of
extremes associated with a small change in the mean value. In a low-warming
scenario where the mean temperature increased about 0.5°C in that area, there
was a 25 per cent increase in summertime days with temperatures over 35°C and
a 25 per cent decrease in wintertime days when the temperature went below 0°C.
The probability that five consecutive days would exhibit such extreme high or low
temperature conditions also showed notable changes.
7.2.2.3.3 A warmer climate is expected to have a more active hydrological cycle as

Precipitation increased evaporation generally leads to higher water vapour content in the
atmosphere. It is believed that this would lead to increases in rainfall, including
extreme rainfall events. Recent model studies for doubled CO2 cases have shown
an increase in the intensity of single precipitation events along with an overall
precipitation increase with temperature. In general, the resolution in the climate
models is not good enough to represent the atmospheric convective elements
that actually cause heavy rains. Model experiments for doubled CO2 cases have
shown both decreases and increases in rainfall for areas that normally have much
rain. In some cases, the predicted rainfall increased while, at the same time, the
number of days with rain events decreased.
In one experiment where the mean precipitation decreased by 22 per cent in
a southern Europe area, the frequency of occurrence for 30-day dry spells more
than doubled in the summer.
It will take considerably more research and model experimentation to clarify
the expectations for mean and extreme precipitation associated with climate change.
7.2.3 The discussion in Chapter 4 on the aspects of current climate models which
REDUCING MODEL contribute the most uncertainty to model simulations provides the basis for devel-
UNCERTAINTIES AND oping the model needed to improve climate change assessments. The 1995 IPCC
IMPROVING CLIMATE CHANGE report [pp. 345-348, Reference no. 3] discussed nine areas considered to be most
ESTIMATES important for improving global climate models; these are summarized below.
(a) Cloud modeling It is important to improve the parameterization of cloud formation and dissipa-
tion, as radiative energy transfer is very sensitive to cloud cover. This will require
improvements in microphysical parameterization to better represent the ice phase
in cloud, particle size distribution and the type of precipitation (rain or snow).
Particle size distribution and ice phase components are all important for deter-
mining the radiative properties of clouds, particularly for solar radiation. The
proper simulation of snow is important for impacts on surface albedo and ice field
growth. Furthermore, improvement is needed in parameterization for cloud scale
dynamics to include deep convection and turbulence effects. Additional observa-
tional information will be necessary for this work.
(b) Ocean component Improvement of resolution is essential to improving the ocean component. It is
felt that the horizontal resolution needs to be reduced to much less than a 1°
latitude-longitude grid to resolve the smaller-scale eddies that influence the circu-
lation with even smaller grid spacing in tropical areas. Currently, some ocean
general circulation models have grid spacing as small as 1/6° in both latitude and
longitude. Improving the thermohaline circulation simulation is necessary for
representation of the dynamics of the full ocean and long time-scale interactions
provided by the oceans in the climate system. Observation of the temperature,
salinity and motions in the deeper ocean areas will be required for this modelling
improvement work. Finally, parameterization for sub-grid scale processes needs to
be improved.
(c) Flux adjustments To maintain appropriate balances in the coupled ocean-atmosphere system, it is
desirable to avoid the need for using flux adjustments at the ocean-atmosphere
97
interface in the climate models. The use of a flux adjustment makes it more diffi-
cult to interpret variability in the model simulations; many of the major modelling
centers no longer use it, because the realism of their climate models has improved.
(d) Longer periods for More general availability of ensembles of 100- to 1000-year (or even longer) simu-
simulations lations will help to calibrate and validate climate models using past-climate
variations. It is expected that computer capacity will continue to grow to make
this possible.
(e) Sea ice component Sea-ice modelling should include motion effects to properly represent dynamic
and thermodynamic feedbacks to the ocean and atmosphere.
(f) Land surface processes A fully interactive land-surface component should be incorporated into the global
climate models used for climate-change assessments particularly for land areas.
This will require model formulations that represent land-surface structure (land-
surface type) and functions (processes within the land-surface features).
(g) Radiation computation It is necessary to improve the radiation computational scheme so that, in partic-
ular, the water vapor and aerosol effects on solar radiation are represented better.
(h) Global carbon cycle The oceanic and land-surface components of the carbon cycle should be incorpo-
rated into the climate models. This incorporation depends on other improvements
in the oceanic and land-surface components mentioned above. In the ocean, deep
circulation effects are important. It will be necessary to obtain better observations
for the deep ocean currents and carbon chemical components.
(i) Tropospheric chemistry The radiative effects of tropospheric sulphates, dust, ozone and other greenhouse
gases (CFCs, methane, and nitrous oxide) should be incorporated into the climate
model. This will require modelling the appropriate chemical processes for these
substances and predicting the size distribution for the aerosol particles.
7.3 A key challenge for climate change assessment is to determine how much of the
DETECTION AND observed changes in climate are, in fact, due to the effects of human activity
ATTRIBUTION FOR (anthropogenic factors) and to describe these aspects. It is important to make the
CAUSES OF RECENT climate change issue understood in order to gain public interest for taking appro-
CLIMATE CHANGE priate responses to limit climate change and to deal with impacts on society.
7.3.1 Meteorologists need to be able to give a clear response to the question: “has
INTRODUCTION climate change already occurred?”.
It is recognized that humans have already had significant impact on envi-
ronmental conditions, particularly since the beginning of the industrial era. The
discussion about climate change has included changes up to the present as well as
those anticipated in the future. Considerable attention is now being given to
showing that climate change is already in progress. The task has two parts: first to
isolate signals due to anthropogenic influences from the background ‘noise’ of
natural variability (detection), and second to define the specific causes of these
effects (attribution).
The studies used to detect climate change and to attribute its causes have
become increasingly sophisticated. The approaches, from simplest to most
complex, may be described by four ‘stages’ listed below:
Stage 1: Examine global or hemisphere-mean values of a single atmospheric descriptor,

(most commonly used has been annual-mean surface temperature);
Stage 2: Examine spatial patterns of a single atmospheric descriptor, (again, most

commonly used has been surface temperature averaged over a season or a year,
although three dimensional temperature patterns have been used);
Stage 3: Same as Stage 2 but the data (model or observed) have been filtered in space and/or
in time to make the anthropogenic component signal more detectable; and,
98
Stage 4: Simultaneously use more than one atmospheric descriptor in the analysis made
for any of the first three stages.
Model results are often used as signals for the attribution of the data. It is
possible to isolate the causes for model signals by model-sensitivity studies where
hypothesized causal factors are varied one at a time. Results can be extrapolated
from observed conditions where there is correspondence of model signals to
observed change signals.
7.3.2 Recent progress in detection and attribution studies has been made possible by a
RECENT PROGRESS number of advances. First, more realistic climate simulations have become avail-
able from improved climate models. Anthropogenic forcing factors such as for
sulphate aerosols have been added. The first experiments had carbon dioxide or
general greenhouse-gas forcing only. Second, more and longer control simula-
tions have provided more reliable statistics for natural variability. Third, more
sophisticated statistical analysis techniques have been developed.
Nevertheless, there are many challenges and uncertainties that must be dealt with
for the detection and attribution of climate change. Correlations of climate variations
with natural forcing factors may appear sufficient to explain all of the observed
variability and trends. For instance, Friis-Christensen and Lassen (1991), found a
remarkable correlation between the length of the solar sunspot cycle and Northern
Hemisphere land temperature anomalies from 1860 to 1985 (see Figure 7.8).
Subsequent study has suggested that although solar impacts are important in
climate variations, they do not explain most of the recent warming trend (Kelly
and Wigley, 1992). Uncertainties about natural variability exist because of limita-
tions in the observational data as discussed in Chapter 6. Instrument data with
sufficient quality to make variability estimates for the atmosphere go back as far
as 150 years, but the coverage over the globe during the early years was quite
limited; it still has deficiencies. Paleoclimatological data are difficult to interpret
and very incomplete in coverage of space and time. Very long climate-model
control-case simulations have helped to fill in the gaps for natural variability. The
uncertainties of climate model simulations for climate change signals have
already been discussed.
Discussion of several recent studies can help to illustrate accomplishments in
detection and attribution. Many studies have been made at the Stage-1 level using
globally-averaged surface temperature observations. The likelihood that observed
trends in the observational record for global-mean surface temperature could
result from natural variability has been evaluated. In this analysis it was necessary
to use variability statistics from climate models because observational data for
variability was insufficient.
The observed global-mean temperature trend for the past 100 years was
compared with values expected from linear trends of natural variability charac-
teristics defined by three climate models. Linear trends were fitted to a number of
different assembled time-series segments (overlapping ‘chunks’) taken from multi-
century model simulations made by these models with no anthropogenic
radiative forcing changes. This procedure produced a statistical distribution of
possible trends that could be found in the model-simulation data for time scales
Figure 7.8 — Variations in solar-

cycle length and Northern
Hemisphere temperature
anomalies. The two plotted
variables parallel each other
quite remarkably (from Friis-
Christensen and Lassen, 1991,
with permission). [from page
186, Hoyt and Schatten, 1997,
with permission of Oxford
University]
99
Figure 7.9 — Significance of

observed changes in global mean,
annually averaged near-surface
temperature. The solid line gives
the magnitudes of the observed
temperature trend (°C/year) over
the recent record — i.e. over 10
years (1984 to 1993), 20 years
(1974 to 1993), etc. to 100 years.
Observed data are from Jones and
Briffa (1992). Model results
(dashed lines) are from three
AOGCM control integrations: the
GFDL control run (Stouffer et al.,
1994), the first 600 years of the
1000-year ECHAM-1/LSG control
integration (Hasselmann et al.,
1995), and the first 310 years of
the UKMO control run (Mitchell ranging from 10 to 100 years. The observed temperature trend for various lengths
et al., 1995). Linear trends were of record ending at the present time was then compared with the distribution of
fitted to overlapping ‘chunks’ of trends found in the model data for the same length of record (see Figure 7.9). The
the model temperature series, results show that for any record longer than 20 years, the observed warming trend is
thus allowing sampling higher than the level that might be expected from natural variability alone, for more than
distributions of trends to be 95 per cent of the natural variability possibilities. This, of course, is not certain proof
generated for the same 10- to of climate change, but the results indicate probability is low that natural processes
100-year time scales for which alone could account for the observed trend. The conclusion may be exaggerated
observed temperature trends were because the model variability an underestimate of natural variability in the real
estimated. The 95th percentiles of climate system. The italicized part in the sentence above is an answer that can be
these distributions are plotted given to the question: ‘has climate change already occurred?’
with dashed lines for each model Sensitivity of climate variation to forcing can provide information on how
control run and each trend anthropogenic forcing may account for observed climate variations. Again with
length. [from page 423, attention to globally-average surface temperature, studies with simple climate
Reference no. 3]. models have identified aspects of observed variations that can be attributed to
human forcing effects. Figure 7.10 shows the matching between observed and
modelled temperature obtained by adding anthropogenic greenhouse gases and
aerosol forcing to a model system. In this case, an upwelling diffusion-energy
balance model (UD/EB) was used. The adjustment factor is the ‘climate
sensitivity’ of the model. Results suggest that there is some correspondence of
overall trends in the observations to those attained in the model.
In the Stage 2 level of analysis, spatial patterns of change are evaluated. This
represents a more rigorous analysis and one that can provide more confidence in
the attribution of the observed changes to human-produced climate change. The
patterns are called ‘fingerprints.’ They may involve the total value of the variable
or structures associated with variable magnitudes that have certain internal
coherency such as Empirical Orthogonal Functions (EOFs).
An example of one fingerprint is the vertical structure for temperature in the
atmosphere. Enhanced greenhouse forcing, such as that due to increased carbon
dioxide, causes a warming in the troposphere and a cooling in the stratosphere.
Forcing due to solar radiation increase would be expected to increase temperature
at all levels. Figure 7.11 shows the vertical pattern of temperature changes due to
changes in anthropogenic forcing from the pre-industrial age to the present in
both observations and model simulations. The observed pattern of change has a
vertical structure qualitatively similar to that expected from an enhanced green-
house effect and is not of the form expected for solar radiation increase.
Specifically, a common pattern of stratospheric cooling and tropospheric warming
is evident in the observations and in both model experiments. In the model data,
this pattern primarily reflects the direct radiative effect of changes in atmospheric
CO2. Temperature changes in both the observations and in the experiment with
combined CO2+aerosol forcing also show a common pattern of hemispherically
100
Figure 7.10 — Observed changes (a) 0.75

in global mean temperature over OBS
1861 to 1994 compared with ∆T2x = 1.5oC
0.50
Temperature anomaly (oC)

∆T2x = 2.5oC
those simulated using an
∆T2x = 4.5oC
upwelling diffusion-energy
balance climate model. The 0.25
model was run first with forcing

due to greenhouse gases alone 0.00
(a), then with greenhouse gases
and aerosols (b), and finally with
greenhouse gases, aerosols, and -0.25
an estimate of solar irradiance
changes (c). The radiative -0.50
forcings were the best-guess 1850 1880 1910 1940 1970 2000
values recommended in this 0.75
(b)
report. Simulations were carried OBS
out with climate sensitivities ∆T2x = 1.5oC
0.50
(T2X) of 1.5, 2.5 and 4.5°C for ∆T2x = 2.5oC

the equilibrium CO2-doubling ∆T2x = 4.5oC
temperature change. [from page 0.25

424, Reference no. 3].
0.00
-0.25
-0.50
1850 1880 1910 1940 1970 2000
(c) 0.75
0.50
0.25
0.00
-0.25
-0.50
1850 1880 1910 1940 1970 2000
Time (years)
asymmetric warming in the low- to mid-troposphere, with reduced warming in

the Northern Hemisphere. This asymmetry is absent in the CO2-only case.
Each forcing component has its own fingerprint. To the extent that finger-
prints are different, it is possible to identify causes for patterns of change seen in
both observations and model simulations. For example, horizontal fingerprints
for surface-temperature climate change expected with greenhouse-gas enhance-
ment would have large amplitude over polar and continental areas such as shown
in Figure 7.3. In contrast, the horizontal fingerprint for aerosol effects would more
likely have large amplitudes over industrialized continental areas and resemble
the aerosol concentration variations shown in Figure 3.3.
Analysis of the correspondence of observed or modelled patterns of climate
change with the fingerprints defined for each forcing component can be done by
statistical methods. An example is the EOF approach where the statistically ‘most
dominant’ one or two component structures (EOFs) are identified for the patterns
101
Figure 7.11 — Modelled and (a) 50

observed changes in the zonal- 100
mean, annual-average temperature 200
Pressure (hPa)
structure of the atmosphere (°C).
300
(a) Model results are from
equilibrium response experiments
performed by Taylor and Penner 500
(1994), in which an AGCM with a
mixed-layer ocean was coupled to a
700
tropospheric chemistry model and
forced with present-day atmospheric
850
concentration of CO2 (b) and 60N 45N 30N 15N 0 15S 30S 45S 60S
by the combined effects of present-
day CO2 levels and sulphur (b) 50
emissions (c). Model changes are 100
expressed relative to a control run 200

Pressure (hPa)
with pre-industrial levels of CO2 300

and no anthropogenic sulphur
emissions. Observed changes (c) are
radiosonde-based temperature 500
measurements from the data set by
Oort and Liu (1993) and are
700
expressed as total least-squares
linear trends over the 25-year period
850
extending from May 1963 to April 60N 45N 30N 15N 0 15S 30S 45S 60S
1988 (i.e., °C/25 years). Dark
shading at and above the 150 hPa (c) 50
100
pressure level highlights regions of
cooling. Dark shading generally 200
Pressure (hPa)
below the 200 hPa level highlights 300

warming except for the region at
300 hPa and 60°N which has
cooling. For further details refer to 500
Santer et al. (1995). [from page
428, Reference no. 3].
700
850
60N 45N 30N 15N 0 15S 30S 45S 60S
-1.5 -0.9 -0.3 0.3 0.9 1.5
-1.8 -1.2 -0.6 0 0.6 1.2 1.8
and fingerprints along with an amplitude factor for each. Then correspondence of
the overall patterns can be measured by examining the amplitude factors.
In conclusion, the overall assessment of the scientific community involved
with the 1995 Intergovernmental Panel on Climate Change report concerning the
existence of climate change was that: ‘the balance of evidence suggests a discernible
human influence on global climate’. The detection of and attribution for climate
change will continue to receive attention from many scientists. One can expect
many new results with more definitive conclusions in forthcoming years. The
subsequent IPCC assessment of 2001 (Reference no. 7) had a stronger conclusion
about the human influence on global climate stating that: ‘there is new and stronger
evidence that most of the warming over the last 50 years is attributable to human activ-
ities’ (Summary for Policy Makers in Reference no. 7).
102
CHAPTER 8
POTENTIAL IMPACTS OF CLIMATE CHANGE
8.1 Climate change is expected to have an impact on a wide range of ecological and
INTRODUCTION socio-economic areas including human health. It has been considered important
to investigate these impacts at the same time as climate change itself is assessed.
In its 1995 second assessment, IPCC provided an extensive report on potential
impacts of climate change (Reference no. 4).
Impact assessment is still in a preliminary stage as it is difficult to quantify
and most studies have been quite limited in scope. Analyses generally have used
simple assumptions about climate-change conditions and have considered only
limited aspects of the complex interactive stress factors. The overall impact on a
system depends on both sensitivity to the climate-condition changes and the
adaptability and compensating factors that the system itself possesses. In many
cases actual impacts will depend on regional climates for which the estimates of
change are far more uncertain than for global-mean conditions.
Some examples of potential impacts are presented here. These were chosen
to focus on areas in which operational meteorologists may be directly involved in
discussions with government agencies or citizens.
8.2 There are a number of ways in which climate change will affect terrestrial ecosystems.
TERRESTRIAL ECOSYSTEMS They are discussed briefly below and followed by a focused discussion of several
specific ecosystems.
Terrestrial ecosystems depend directly on temperature and precipitation clima-
tology as shown in Figure 1.16. Changes in temperature and precipitation, including
their extremes and seasonal or daily variations, will influence the distribution of
biomes in the world, including those in agriculture. Rising temperatures alone would
be expected to foster the poleward migration of biome species. For example, a
warming could be expected to improve options for agriculture in subarctic regions.
Changes in climate will influence other determinants of the ecosystem
condition such as disease, pest cycles and the incidence of fires. The increase in
atmospheric carbon dioxide concentrations is expected to increase the primary
productivity of plants, i.e. make them grow faster. This could change the balance
among plants competing for the same space.
The largest adverse impacts on the ecosystem are anticipated to be from
direct human activity itself. The clearing of land for agriculture and urbanization
and the segmentation of ecosystems will be important factors. More unfavourable
impacts are likely in tropical and subtropical developing countries compared to
developed countries because of population pressures and lack of resources for
adaptation and mitigation responses to climate change impacts.
8.2.1 Climate change will affect crops in a number of ways. These include the growth
AGRICULTURE (PLANT CROPS) process of crop plants and those of insects, weeds and diseases as discussed above.
Growth will be affected by changes in CO2 concentration as well as those in
temperature, moisture supply and severe weather. CO2 increases alone are
expected to increase the productivity of annual crops, particularly those that may
be limited by existing concentrations of carbon dioxide (i.e. most crops including
wheat, rice, barley, cassava and potato, and most trees). For these crops, increases
on the order of 30 per cent would be expected for a doubling in CO2 if there were
no other changes in conditions. The increases would be less for plants that have
a special CO2-concentrating mechanism (such crops as maize, millet, sugar cane,
sorghum, and many tropical grasses).
Temperature and moisture supply have a dominant influence on crop
growth. Each crop type has an optimal temperature range for growth. Moisture
supply is critical throughout the growth period of the crop. The moisture supply
103
depends on both precipitation and evaporation (evapotranspiration from the

plant). The diurnal and day-to-day variability for temperature and moisture
supply along with mean values are important climatic factors for plants because
of the negative impact of extremes.
Very preliminary impact assessments have been made for a selection of crops
using simple estimates of climate change. The estimates were derived from
climate-model simulations, historical data, or just an outright specification of a
temperature change. The impacts show extreme variability, often ranging from
increases to decreases of yields for a given crop. Some examples are shown for
areas in Africa, south Asia, Latin America, and western Europe in Tables 8.1 to 8.4.
One may find comparable summaries for all other major regions of the Earth in
chapter 13 in the 1995 IPCC Working Group II report (Reference no. 4). The
climate models referenced are briefly described earlier in Chapter 4.
It is clear from these tables that there is great uncertainty in the quantitative impact
assessments of climate change on agricultural crops. It is strongly suggested that crop
yields and productivity will vary a great deal from one region to another. Some areas will
see improvement in yields, others will see decreases. Agricultural patterns will change. It
is likely that overall impacts will be significant. The overall effects for a region will
depend on many factors in that region such as irrigation options, current agricultural
infrastructure, adaptations in farming practices and so forth. It will take serious planning
efforts to deal most constructively with the changes.
8.2.2 Overall regional impacts of climate change for forests are briefly summarized on p. 26 of
FORESTS Reference no. 4. As the projected temperature increases are smaller in tropical latitudes,
tropical forests will be affected less than those in other latitudes. However, climate
change in terms of the amount and seasonality of rainfall could have larger impacts.
Table 8.1 — Selected crop studies Even so, other human impacts will likely affect tropical forests more than does climate
for Africa and the Middle East for change. Temperate forests will be impacted by temperature, precipitation and CO2
climate-change scenarios from changes differently from region to region. However, the negative aspects of such
climate models, observations, and changes on temperate forests will be minimized by reforestation and forest
prescribed changes. [from page management programmes, since most temperate forests are located in developed
438, Reference no. 4]. countries. Boreal forests will be most affected by climate change as warming is expected
Study Scenario Geographic scope Crop(s) Yield impact Other comments

in per cent
Eid. 1994 GISS, GFDL, Egypt Wheat -75 to -18 w/CO2 effect; also temperature
UKMO Maize -65 to +6 and precipitation sensitivity;
adaptation would require heat-
resistant variety development.
Schulze et al., 1993 +2°C (1) South Africa Biomass decrease Mapped results, not
Maize increase summarized as average
change for entire region.
Muchena, 1994 GISS, GFDL, Zimbabwe Maize -40 to -10 w/CO2 effect; also
UKMO temperature and precipitation
sensitivity; adaptation (fertilizer
and irrigation) unable to fully
offset yield loss.
Downing, 1992 +2/+4°C, Zimbabwe Maize -17 to -5 Food availability estimated to
± 20 per cent Senegal Millet -70 to -63 decline in Zimbabwe; carrying
precipitation Kenya Maize decrease capacity fell 11 to 38 per cent in
Senegal; overall increase for all
crops in Kenya with zonal shifts.
Akong’a et al., 1988 Historical droughts, Kenya Maize, negative effects Considered
broader socio-
sensitivity livestock of drought economic impacts, small-holder
impacts, a policy implications.
Sivakumar, 1993 1945-1964 vs. Niger Growing reduced Crop variety development,
1965-1988 West Africa season 5-20 days timely climate information seen
as important adaptation strategies.
104
CHAPTER 8 — POTENTIAL IMPACTS OF CLIMATE CHANGE

in per cent
Rosenzweig GCMs Pakistan Wheat -61 to +67 UKMO, GFDL, GISS, and +2°C,
and Iglesias India Wheat -50 to +30 +4°C, and ± 20% precipitation
(eds.), 19941 Bangladesh Rice -6 to +8 range is over sites and GCM
Thailand Rice -17 to +6 scenarios with direct CO2 effect;
Philippines Rice -21 to +12 scenarios w/o CO2 and w/ adapta-
tion also were considered; CO2
effect important in offsetting
losses of climate-only effects;
adaptation unable to mitigate all
losses.
Qureshi and average of Bangladesh Rice +10 GCMs included UKMO, GFDLQ,
Hobbie, 1994 five GCMS India Wheat decrease CSIRO9, CCC and BMRC; GCM
Indonesia Rice -3 results scaled to represent 2010;
Soyabean -20 includes CO2 effect.
Maize -40
Pakistan Wheat -60 to -10
Philippines Rice decrease
Sri Lanka Rice -6
Soyabean -3 to +1
Coarse grain decrease
Coconut decrease
Parry et al., (eds), GISS Indonesia Rice approx. -4 Low estimates consider
1992 Soyabean -10 to increase adaptation; also estimated
Maize -65 to -25 overall loss of farmer income
ranging from $10 to $130 annu-
ally.
Malaysia Rice -22 to -12 Maize yield affected by
Maize -20 to -10 reduced radiation (increased
Oil palm increase clouds); variation in yield
Rubber -15 increases; range is across seasons.
Matthews et al., three GCMs Thailand sites Rice -5 to +8 Range across GISS, GFDL, and
1994a, 1994b India Rice -3 to +28 UKMO GCM scenarios and crop
Bangladesh -9 to +14 models; included direct CO2
Indonesia +6 to +23 effect; varietal adaptation was
Malaysia +2 to +27 shown to be capable of
Myanmar -14 to +22 ameliorating the detrimental
Philippines -14 to +14 effects of a temperature increase in
Thailand -12 to +9 currently high-temperature
environments.
1 Country studies were by Qureshi and Iglesias, 1994; Rao and Sinha, 1994; Karim et al., 1994; Tongyai, 1994; and Escaño and
Buendia, 1994, for Pakistan, India, Bangladesh, Thailand, and the Philippines, respectively.
Table 8.2 — Selected crop studies to be largest at high latitudes. Increased fire and pest outbreaks will negatively impact
for south and south-east Asia for the southern regions of boreal forests, whereas the increased temperature and moisture
climate change scenarios from supply in the northern regions of the boreal forests will enhance the forest, which is
climate models. [from page 439, expected to advance northward into the tundra.
Reference no. 4]. Forest fires are a matter of special interest. Forest fires occur commonly in
seasonally dry forest areas due to human and natural causes. They play an important
role in the ecosystem dynamics in tropical, temperate and boreal zones. Handling them
is an important part of forest management. In some cases of major fires there can be
extreme danger to human life and property locally, as well as region-wide health effects
from air pollution. Temperature increases due to climate change are expected to increase
drought conditions, which would lead to more favourable conditions for forest fires in
seasonally dry areas. This would be of particular concern to areas that do not have
integrated fire, pest and disease management.
8.2.3 Roughly 30 per cent of the earth’s land surface is desert or semi-desert; as shown by the
DESERTS, LAND DEGRADATION, dry climate areas (BS and BW classifications) in Figure 1.20. These are areas where the
AND DESERTIFICATION lack of moisture is a serious impediment to the growth of plants. Adjacent to these areas
and elsewhere are regions, estimated at 17 per cent of the total earth land surface, where
105

in per cent
Baethgen, 1992, GISS, Uruguay Barley -40 to -30 w/ and w/o CO2; with
1994 GFDL, Wheat -30 adaptation, losses were 15 to 35
UKMO1 per cent; results indicate
increased variability.
Baethgen and UKMO Argentina Wheat -10 to -5 w/CO2; high response to CO2,
Magrin, 1994 Uruguay high response to precipitation.
Siquera et al., GISS, Brazil Wheat -50 to -15 w/CO2; w/o adaptation;
1994; Siquera, GFDL Maize -25 to -2 adaptation scenarios did not fully
1992 UKMO1 Soybean -10 to +40 compensate for yield losses;
regional variation in response.
Liverman et al., GISS, Mexico Maize -61 to -6 w/CO2; adaptation only partly
1991, 1994 GFDL, mitigated losses.
UKMO1
Downing, 1992 +3°C. Norte Chico Wheat decrease The area is especially difficult to
25 per cent Chile Maize increase assess because of the large range
precip. Potatoes increase of climates within a small area.
Grapes decrease
Sala and Paruelo, GISS, Argentina Maize -36 to -17 w/ and w/o CO2; better adapted
1992, 1994 GFDL, varieties could mitigate most
UKMO1 losses.
1 These studies also considered yield sensitivity to +2 and +4°C and -20 and +20 per cent change in precipitation.
Table 8.3 — Selected crop studies ‘desertification’ ascribed to human activity is occurring. This process is an ecological
for Latin America for climate- degradation which causes economically-productive land to become less productive,
change scenarios from climate more desert-like and incapable of continuing to sustain an existing community of
models and prescribed changes. people. Currently about one sixth of the earth’s population lives in regions where such
[from page 444, Reference no. 4]. desertification is occurring.
Desertification involves a number of factors. Soil that is cultivated for agri-
culture can be eroded by water flow and wind. Salinization of soils can occur near
coastlines due to sea-level rises. Inland salinization can occur due to salt accumu-
lation related to erosion, seepage and wind deposition. Overall significant causes
of desertification can be traced to overcultivation, overstocking, fuel and wood
collection, salinization and urbanization.
It is difficult to define the effects of global climate change in these arid areas,
especially where direct effects due to human activity are already taking place.
Generally climate-change temperature increases would be expected to increase
stresses on plants. This would tend to make desert conditions more severe and to
accentuate desertification processes. The increase in CO2 would be expected to
reduce plant transpiration and to increase the water use efficiency of plants.
Climate model projections for changes in precipitation are very uncertain.
Impacts of precipitation changes depend critically on changes in distribution
throughout the year and in extreme events, aspects which are not reliably
handled by the models. Most model simulations to date do not suggest signifi-
cantly wetter conditions in arid regions.
In summary, local environmental change due to human activity (desertifica-
tion) is significant in many arid areas and is independent of global climate
change. It is not expected that climate change will offset the desertification
process, but rather that climate-change-related factors, such as increased drought
conditions resulting from rising temperatures, will increase the vulnerability of
land to desertification. Many arid and semi-desert areas are in developing coun-
tries where the negative impacts would be most severe.
8.3 Analysis of climate-change impacts on freshwater resources must include their

FRESHWATER RESOURCES effects on both water supply and water demand.
MANAGEMENT A rough analysis of water supply can be based on the runoff of surface water,
which depends on the difference between precipitation over a river catchment
106

in per cent
Oleson et al., 1993 * Northern Europe Cauliflower Increase Quality affected by temperature;
longer season.
Goudriaan and +3°C Northern Europe Maize (fodder) Increase Shift to grain production
Unsworth, 1990 possible.
Squire and +3°C Northern Europe Wheat Increase
Unsworth, 1988
Kettunen et al., 1988 GCMs Finland Potential yield +10 to +20 Range is across GISS and UKMO
GCMs.
Rötter and van +2°C Rhine area Cereals, sugar +10 to +30 Also +10 per cent winter
Diepen, 1994 (winter), beet, potato, precipitation; includes direct
+1.5°C grass effect of CO2; range is across
summer crop. agroclimatic zone, and soil
type; decreased evapotranspira-
tion (1 to 12 per cent), except
for grass.
U.K. Dept. of GCMs U.K. Grain, Increase or Increased pest damage; lower risk
Environment, +1, +2°C horticulture level increase of crop failure.
1991
Wheeler et al., 1993 * U.K. Lettuce Level Quality affected; more crops per
per season possible.
Semonov et al., * U.K./France Wheat Increase or Yield varies by region; UKMO
1993 decrease scenario negative; includes adap-
tation and CO2.
Delécolle et al., GCMs** France Wheat, Increase or Northward shift; w/adaptation,
1994 +2, +4°C maize level w/CO2; GISS, GFDL and UKMO
GCMs.
Iglesias and GCMs** Spain Maize -30 to -8 w/adaptation, w/CO2; irrigation
Minguez, 1993 efficiency loss; see also Minguez
and Iglesias, 1994.
Santer, 1985 +4°C Italy/Greece Biomass -5 to +36 Scenarios included -10 per cent
precipitation.
Bindi et al., +2, +4°C Italy Winter wheat Not estimated Crop growth duration decreases;
1993 and * adaptation (using slower devel-
oping varieties) possible.
* Climate scenarios included GISS, GFDL and UKMO and time-dependent scenarios, using GCM methodology, based on emis-
sion scenarios proposed by the IPCC in 1990. Composite scenarios for temperature and precipitation were based on seven
GCMs and scaled by the global-mean temperature changes associated with the IPCC 1990 emissions scenarios for the years
2010, 2030, and 2050 (Barrow, 1993).
** These studies also considered yield sensitivity to +2 and +4°C and -20 and +20 per cent change in precipitation.
Table 8.4 — Selected crop studies and evaporation over the same area. This approach would neglect the effects due
for Western Europe for climate- to groundwater recharge and salinization due to sea-level rise and so forth. As
change scenarios from climate discussed earlier, climate-change estimates for precipitation and evaporation from
models and prescribed changes. climate models have considerable uncertainty. It is useful to describe the water
[from page 445, Reference no. 4]. supply in terms of the amount available per person so that population impacts are
included but separated from climate-change impacts.
An impact analysis was made for the year 2050 using the results of three
climate-model simulations for climate change. Results for selected countries are
shown in Table 8.5. The values for the three quantities shown are for: current
(1990) fresh water availability in m3/yr/person; availability in the year 2050
assuming current water amounts but with the increased population expected; and
the range of water availability in the year 2050 based on three transient climate-
model scenarios. Comparison of the second and third quantities shows the
predicted impact of climate change on water availability.
As can be seen in Table 8.5, there is a wide range of estimates for the water
availability with climate-change conditions. This is due to the significant
107
variations in precipitation and evaporation among the three climate models.

Predictions for any given country range from increases to decreases in most cases.
Clearly there is great uncertainty in the results. However, overall the number of
countries with shortages in water supply is expected to increase by the year 2050.
This is due to population increases which are not offset by precipitation increases
(actually precipitation minus evaporation increases) even for the most favourable
of the three model predictions.
A value of 1000 m3/yr/person has been used as the baseline for minimum
water requirement. Based on this number, only two of the 21 countries listed in
Table 8.5 had a water shortage in 1990, eight of these countries would have a
water shortage in 2050 due to population increase alone, and anywhere from
seven to eleven would have a water shortage in 2050 based on both population
and climate change projections.
Freshwater demands are expected to increase with the warming associated
with climate change. However, this impact will be combined with population
growth, economic factors, and changes in agricultural, industrial and domestic
practices, acting at the same time. Irrigation for agriculture is the biggest user for
water extracted from the natural reservoirs (rivers, lakes and ground water).
Temperature increases are expected to increase irrigation requirements more than
precipitation enhancement would decrease such requirements. Impacts of climate
change on municipal uses for domestic and industrial purposes are not well
defined.
8.4 Observations have detected sea-level rises over the last 100 years of between eight
SEA-LEVEL RISE and 31 cm (cf. Figure 2.14). The overall global mean value is estimated to be
between 10 and 25 cm. This is related to the observed net melting of glaciers and
ice fields and the net warming of the surface air temperature. Temperature increases
in the ocean lead to change in sea level due to thermal expansion. Thermal expan-
sion and ice melting are estimated to be of comparable importance for the sea-level
changes. There are local variations in the rate of sea-level change due to geological
factors of ‘post-glacial rebound’ (rising of earth surface as the weight of an overlying
ice mass is removed) and other earth crust movements.
Table 8.5 — Water availability Present Present Scenario

(m3/yr/person) in 2050 for the Climate Climate Range
present climatic conditions and Country (1990) (2050) (2050)
for three transient climate model
scenarios (GFDL, UKMO, MPI) China 2,500 1,630 1,550–1,780
compared with the present. The Cyprus 1,280 820 620–850
range in simulated values for the France 4,110 3,620 2,510–2,970
three climate models is shown in Haiti 1,700 650 280–840
the right hand column. [from India 1,930 1,050 1,060–1,420
page 478, Reference no. 4]. Japan 3,210 3,060 2,940–3,470
Kenya 640 170 210–250
Madagascar 3,330 710 480–730
Mexico 4,270 2,100 1,740–2,010
Peru 1,860 880 690–1,020
Poland 1,470 1,250 980–1,860
Saudi Arabia 310 80 30–140
South Africa 1,320 540 150–500
Spain 3,310 3,090 1,820–2,200
Sri Lanka 2,500 1,520 1,440–4,900
Thailand 3,380 2,220 590–3,070
Togo 3,400 900 550–880
Turkey 3,070 1,240 700–1,910
Ukraine 4,050 3,480 2,830–3,990
United Kingdom 2,650 2,430 2,190–2,520
Vietnam 6,880 2,970 2,680–3,140
108
Climate-change warming is expected to cause a sea-level rise of between 20 and

100 cm by the year 2100 as ocean warming and ice melting occurs at a faster rate
than in the past 100 years. Figure 8.1 shows the wide range in the estimates which
have been made. This wide range results from great uncertainty about how ice melt
will proceed in polar regions. The rate of sea-level rise is projected to be anywhere
from the same to five times greater than that experienced in the last 100 years.
The rise in sea level will have important impacts on coastal zones and small
islands. Coastal erosion, flooding, saltwater intrusion and sedimentation changes
will be factors that impact on human settlements, agriculture, freshwater supply
and quality, fisheries, and human health. Millions of people will be affected in
countries such as Bangladesh, Benin, China, Egypt, India, Japan, The Netherlands
and Nigeria. In some, such as the Marshall Islands, entire countries will be
affected. Table 8.6 shows estimates for impacts for a 100 cm sea-level rise.
8.5 Changes in the frequency, intensity and location of storms may be an important
STORMS aspect of climate-change impacts over many parts of the world. Tropical cyclones,
i.e. typhoons and hurricanes, affect many areas in and adjacent to the Indian,
Pacific and Atlantic Oceans, primarily in the Northern Hemisphere. Damage
caused by flooding and high winds can be devastating for land areas. In the
middle latitudes, high winds and precipitation associated with extratropical
cyclones can have major impacts even at locations distant from the oceans. Some
comments on the effects due to climate change were made in Chapter 6 as part
of the discussion on changes in extreme events.
Both tropical cyclones and the severe weather aspects of extratropical cyclones are
of regional scale (mesoscale) and are difficult to represent in global-climate models.
Assessments of modification of these systems due to climate change up to now have
been based primarily on inferences from changes in large-scale conditions.
The ENSO cycle in the tropical Pacific Ocean area is known to affect both
tropical and extratropical cyclones. In particular the warm El Niño phase of ENSO
is associated with enhanced extratropical storminess in the western and southern
United States. Accordingly, it is important to identify the climate-change impact
on ENSO. Global climate models have been able to represent ENSO, but further
experimentation is needed to isolate climate change impacts.
8.6 It is anticipated that global climate change will have a wide-ranging and net
HUMAN HEALTH adverse impact on human health, including increased loss of life. This will be due
8.6.1 both to direct causes such as the increased severity of heat waves and to indirect
INTRODUCTION effects such as changes in local food productivity and in the range of diseases
transmitted by organisms in air or water. Direct health impacts will also result
from concurrent environmental changes such as in the concentration of toxic
and carcinogenic air pollutants. Figure 8.2 summarizes some of the major direct
and indirect impacts.
Many impacts will result from disturbances in ecological systems. Such changes
could bring disease conditions to a human population that was formerly outside the
range of such conditions. As an example, atmospheric warming could lead to
mosquitoes reaching higher levels in mountainous areas introducing malaria to
people living there. Populations in developing countries may require additional
resources to deal effectively with such changes in disease patterns.
100 96
Including changes in aerosol beyond 1990 HIGH
86
Sea level change (cm)
80 Constant 1990 aerosol
Figure 8.1 — Scenario IS92a sea- 60

55
MID
49
level rise from 1990 to 2100 for
40
high, medium and low ice-melt
parameter specifications. (See 20
23 LOW
20
Chapter 7 in Reference no. 3 for
more details.) [from page 296 in 0
2000 2020 2040 2060 2080 2100
109
People Capital value Land Wetland Adaptation/

affected at loss at loss at loss Protection Costs
no. of people % Million % % Million %
(1000s) Total US$1 GNP km2 Total km2 US$1 GNP
Antigua2, (Cambers, 1994) 38 50 - - 5 1.0 3 71 0.32

Argentina (Dennis et al.,1995a) - - >50007 >5 3400 0.1 1 100 >1 800 >0.02
Bangladesh (Huq et al.,1995:
Bangladesh Government, 1993) 71 000 60 - - 25 000 17.5 5 800 >1 00010 >0.06
Belize (Pernetta and Elder, 1993) 70 35 - - 1 900 8.4 - - -
Benin3 (Adam, 1995) 1350 25 118 12 230 0.2 85 >40010 >0.41
China (Bilan, 1993; Han et al., 1993) 72 000 7 - - 35 000 - - - -
Egypt (Delft Hydraulics et al., 1992) 4 700 9 59 000 204 5 800 1.0 - 13 10011 0.45
Guyana (Kahn and Sturm, 1993) 600 80 4 000 1 115 2 400 1.1 500 200 0.26
India (Pachauri, 1994) 7 1006 1 - - 5 800 0.4 - - -
Japan (Mimura et al., 1993) 15 400 15 849 000 72 2 300 0.6 - >156 000 >0.12
Kiribati2 (Woodroffe and McLean, 1992) 9 100 2 8 4 12.5 - 3 0.10
Malaysia (Midun and Lee. 1995) - - - - 7 000 2.1 6 000 - -
Marshall Islands2 (Holthus et al., 1992) 20 100 160 324 9 80 - >360 >7.04
Mauritius4 (Jogoo, 1994) 3 <1 - - 5 0.3 - - -
The Netherlands (Peerbolte et al., 1991) 10 000 67 186 000 69 2 165 5.9 642 12 300 0.05
Nigeria (French et al., 1995) 3 2006 4 17 0007 52 18 600 2.0 16 000 >1 400 >0.04
Poland (Pluijm et al., 1992) 240 1 22 000 24 1 700 0.5 36 1 400 0.02
Senegal (Dennis et al., 1995b) 1106 >1 >5007 >12 6 100 3.1 6 000 >1 000 >0.21
St. Kitts-Nevis2 (Cambers, 1994) - - - - 1 1.4 1 50 2.65
Tonga2 (Fifita et al., 1994) 30 47 - - 7 2.9 - - -
United States (Titus et al., 1991) - - - - 31 6008 0.3 17 000 >156 000 >0.03
Uruguay5 (Volonté and Nicholls, 1995) 136 <1 1 7007 26 96 0.1 23 >1 000 >0.12
Venezuela (Volonté and Arismendi, 1995) 566 <1 3307 1 5 700 0.6 5 600 >1 600 >0.03
Table 8.6 — Synthesized results 1 Costs have been adjusted to reflect 1990 US Dollars.
of country case studies. Results 2 Minimum estimates-incomplete national coverage.
are for existing development and 3 Precise year for financial values not given are assumed to be 1992 US$.
a one metre rise in sea level. 4 Results are linearly interpolated from results for a two metre sea-level rise scenario.
People affected, capital value at 5 See also review in Nicholls and Leatherman (1995a).
loss, land at loss, and wetland at 6 Minimum estimates—number reflects estimated people displaced.
loss assume no measures (i.e. no 7 Minimum estimates—capital value at loss does not include ports.
human response), whereas 8 Best estimate is that 20 000 kM2 of dry land are lost, but about 5.400 kM2 are
adaptation assumes protection converted to coastal wetlands.
except in areas with low 9 Adaptation only provides protection against a one-in-20 year event.
population density. All costs 10 Adaptation costs are linearly extrapolated from a 0.5-m sea-level rise scenario.
have been adjusted to 1990 US 11 Adaptation costs include 30-year development scenarios.
Dollars (adapted from Nicholls,
1995). [from page 308, A range of possible health impacts to climate change are summarized here.
Reference no. 4]. Research has yet to quantify such impacts and it may be difficult to isolate them
from overall trends in world health (except for stratospheric ozone reduction).
Nevertheless, it is important to be aware of the possibilities because of the poten-
tial impacts on whole communities or populations.
8.6.2 Global warming is expected to increase the frequency of extremely hot days as
POTENTIAL DIRECT EFFECTS discussed in Chapter 6. Recognizing that extremely hot spells are related to
increases in mortality rates, it is logical to conclude that mortality due to exces-
sively hot conditions will increase with global warming. Estimates of changes in
heat-related mortality rates for selected large cities using temperatures predicted
by global climate models show that by the year 2020 the rate will double and
nearly quadruple by the year 2050. These results are based on current mortality
rate increases due to hot weather. They neglect other factors that may operate in
a hot spell such as increased air pollution and mitigation programmes that might
be implemented with global warming.
110
Mediating process Health outcomes
DIRECT
Altered rates of heat- and cold-
Exposure to thermal extremes related illness and death
(especially heatwaves) (especially cardiovascular and
respiratory diseases)
Altered frequency and/or intensity Deaths, injuries and psychologi-

CLIMATE of other extreme weather events cal disorders; damage to public
CHANGE: (floods, storms, etc.) health infrastructure
TEMPERATURE,
PRECIPITATION
AND WEATHER INDIRECT
DISTURBANCES OF
ECOLOGICAL SYSTEM
Changes in geographic ranges
Effects on range and activity of and incidence of vector-borne
vectors and infective parasites diseases
Altered local ecology of water- Changed incidence of diarrheal

borne and food-borne infec- and certain other infectious dis-
tive agents eases
Altered food (especially crop) Regional malnutrition and

productivity due to changes in hunger, and consequent impair-
climate, weather events, and ment of child growth and devel-
associated pests and diseases opment
Sea-level rise with population Injuries, increased risks of vari-

displacement and damage to ous infectious diseases (due to
infrastructure (e.g., sanitation) migration, crowding, contamina-
tion of drinking water), psycho-
logical disorders
Levels and biological impacts Asthma and allergic disorders;

of air pollution, including other acute and chronic respira-
pollens and spores tory disorders and deaths
Social, economic and demo-

graphic dislocations due to Wide range of public health
adverse climate change impacts consequences (e.g., mental
on economy, infrastructure health, nutritional impairment,
and resource supply infectious diseases, civil strife)
Figure 8.2 — Ways in which
climate change can affect human NOTE: Populations with different levels of natural, technical and social
health. (from page 565 of resources would differ in their vulnerability to climate-induced health
Reference no. 4). impacts.
On the other hand, global warming is also expected to reduce the occurrence
of cold conditions in parts of the world. In such areas mortality due to cold condi-
tions, including accompanying respiratory diseases, is expected to decrease. On
balance for the world it is felt that the sensitivity of death rates to hotter summers
may be greater than that for the warmer winters. Thus, the overall temperature
impact is negative. Considerably more research, especially in developing coun-
tries, is needed to quantify the impacts of temperature on health.
111
Other weather extremes such as droughts, floods and high winds also cause
adverse health effects including deaths for humans. Understanding the impact of
these factors due to climate change will require defining the changes in such
weather extremes.
8.6.3 It is expected that climate change will have a significant impact on vector-borne
POTENTIAL INDIRECT EFFECTS diseases particularly in tropical and subtropical countries. ‘Vector-borne’ refers to
8.6.3.1 the processes by which the infective agent of the disease is transmitted by a living
Vector-borne diseases organism (the vector) such as a mosquito, tsetse fly, bug or tick. The climate
change may directly affect the living patterns of the vector which in turn will
affect the spread of the disease among humans as a secondary or indirect effect
Vectors such as mosquitoes are extremely sensitive to temperature. For
instance, the anopheline mosquito species, which transmits malaria, normally
does not survive if the mean winter temperature is below 16-18°C, or if nighttime
temperatures in summer are sufficiently low. Furthermore, the temperature at
which the mosquito can incubate the malaria parasite has a lower limit, e.g., 18°C
and 14°C for the Plasmodium falciparum and Plasmodium vivax parasites, respec-
tively. If climate change increases temperature, then the mosquito can become an
active vector for malaria at higher levels in mountainous country and at latitudes
more poleward of the tropical areas. This can expand the malaria area to include
populations which were not previously exposed and which would lack naturally-
acquired immunity.
A large number of vector-borne diseases are active in the tropical and
subtropical areas as listed in Table 8.7. The number of people likely to be exposed
is enormous. Alteration (expansion) of the risk areas due to climate change will
affect a large number of additional persons. The diseases considered to be most
likely to be involved are malaria, schistosomiasis, onchocerciasis, dengue fever
and yellow fever.
8.6.3.2 Alteration of water-borne and food-borne infectious diseases is another major

Water-borne and food-borne potential indirect impact of climate change. Diarrheal diseases such as cholera
diseases and dysentery are spread by untreated water and water systems infiltrated by
run-off water. Climate change would have an impact to the extent that it
changes or increases flooding situations and by providing warmer environ-
ments for bacterial development. The rise in sea level due to climate change
would be expected to increase coastal flooding and to degrade sewage disposal
systems.
8.6.3.3 Climate-change impacts such as increased drought, desertification, increased

Agricultural productivity and severe weather and increased coastal flooding could have serious impacts on food
food supplies supply, especially in developing countries. This could have negative impacts on
nutrition and health of the population.
8.6.3.4 Air pollution effects may be increased by climate change. This is in addition to
Air pollution the human-produced air pollution which is a factor in producing the climate
change in the first place. Climate changes may impact levels of pollen and other
biotic allergens from birch trees, grasses, oilseed rape crops and ragweed. These
effects would be further enhanced by pollutants, such as ozone, that are gener-
ated by fossil fuel combustion and the action of solar radiation.
8.6.4 A serious health hazard has arisen from reduction of ozone in the stratosphere
STRATOSPHERIC OZONE due to human-produced CFC and other chlorine- and bromine-containing gases.
DEPLETION AND INCREASED Ozone in the stratosphere helps to protect life at the Earth’s surface by absorbing
EARTH-SURFACE ULTRAVIOLET most of the harmful ultraviolet radiation in the sunlight. It has been estimated
RADIATION that the ozone concentration in the middle and high latitudes has, on average,
been reduced by 10 per cent in the past ten years. The increase of ultraviolet radi-
ation causes skin cancer, eye cataracts and damage to the local immune system in
the skin. The suppression of the immune system leads to increased susceptibility
to infectious diseases.
112
Number of Likelihood
Population people currently of altered
at risk infected or new Present distribution with
Disease Vector (million)1 cases per year distribution climate change
Malaria Mosquito 24002 300-500 million Tropics/Subtropics +++

Schistosomiasis Water snail 600 200 million Tropics/Subtropics ++
Lymphatic filariasis Mosquito 10943 117 million Tropics/Subtropics +
African Trypanosomiasis Tsetse fly 554 250 000-300 000 Tropical Africa +
(Sleeping sickness) cases per year
Dracunculiasis Crustacean 1005 100 000 per year South Asia/ ?
(Guinea worm) (Copepod) Arabian Peninsula/
Central-West Africa
Leishmaniasis Phlebotomine 350 12 million infected. Asia/Southern +
Sand fly 500 000 new Europe/Africa/
cases per year6 Americas
Onchocerciasis Black fly 123 17.5 million Africa/Latin America ++
(River blindness)
American Trypanosomiasis Triatomine bug 1007 18 million Central and +
(Chagas‘ disease) South America
Dengue Mosquito 1800 10-30 million per year All Tropical Countries ++
Yellow fever Mosquito 450 <5 000 cases per year Tropical South ++
America and Africa
Table 8.7 — Major tropical vector- + = likely. ++ = very likely. +++ = highly likely. ? = unknown.
borne diseases and the likelihood of
change of their distribution with 1 Top three entries are population-prorated projections (based on 1989 estimates).
climate change. [from page 572, 2 WHO, 1995b.
Reference no. 4]. 3 Michael and Bundy, 1995.
4 WHO, 1994a.
5 Ranque, personal communication.
6 Annual incidence of visceral leishmaniasis; annual incidence of cutaneous leishmaniasis is
between one and one and a half million cases/yr (PAHO, 1994).
7 WHO, 1995c.
The estimated increase in skin cancer is of major concern. Effects are more
notable at high latitudes than in tropical regions because the depletion of ozone
has been greater at high latitudes. As an example, it is estimated that if the current
latitude-dependent reduction in ozone is maintained for the next few decades the
incidence of the skin cancer, basal cell carcinoma, will increase by a factor of one
to two per cent at very low latitudes, three to five per cent for the 15-25° latitudes,
eight-12 per cent at the 35-45° latitudes, and 13-15 per cent at the 55-65° latitudes
(Madronich and de Gruijl, 1993).
113
CONCLUDING REMARKS
The broad scope of these lecture notes should make it clear that there is a wide
range of topics related to the study and understanding of climate change. It
requires collaboration among persons from many different professions to answer
the questions about the magnitude of climate change and the effects it is likely to
have on our way of life.
The lecture notes were designed to promote sufficient understanding of the
science and technology of climate change so that the student will be equipped to
critically evaluate reports concerning climate change and to explain material to
government representatives and lay persons. The material also gives background
information to those who wish to pursue further study. It should be emphasized
again that data and assessments of climate change are being continually updated.
Thus, it is necessary to seek out current reports and publications to keep up to date
on the topic.
The lecture notes have given some examples of the many types of impact
that climate change would have on the human community, and on plant and
animal communities. Some of the impacts may be considered to be favourable,
but a large number are unfavourable and would require actions to reduce their
negative effects. Climate change clearly implies many other kinds of changes. The
world community needs to be alert and provide assistance to those in the world
who lack the resources to respond to the negative impacts of climate change.
Some people still doubt that climate change has occurred despite statistical
evidence at the 95 per cent significance level. Despite their doubts relating to
climate change, environmental impacts of humans on the atmosphere, oceans,
and biosphere are recognized. The net result of these impacts has already been an
overall degradation of our environment. It is important for the meteorological
community to do its part to raise awareness of projected human effects on world
climate.
114
REFERENCES
Reference No. 1: Hartmann, D.L., 1994: Global Physical Climatology, Academic

Press, 411 pp.
Reference No. 2: Karl, T.R. (Editor), 1996: Long-Term Climate Monitoring by the
Global Climate Observing System. (International Meeting of Experts, Asheville,
north Carolina), Reprints from Climate Change, Vol. 31, pp. 132-648, Kluwer
Academic Publishers, Dordrecht, The Netherlands, 518 pp.
Reference No. 3: IPCC Working Group I, 1996: Climate Change 1995: The Science
of Climate Change, Houghton, J.T, L.G. Meira Filho, B.A. Callander, N. Harris,
A. Kattenberg and K. Maskell (Editors), prepared by Working Group I to the
Second Assessment Report of the IPCC, Cambridge University Press,
Cambridge, U.K., 572 pp.
Reference No. 4: IPCC Working Group II, 1996: Climate Change 1995: Impacts,
Adaptations and Mitigation of Climate Change: Scientific-Technical Analysis,
Watson, R.T., M.C. Zinyowera, and R.H. Moss (Editors), prepared by Working
Group II to the Second Assessment Report of the IPCC, Cambridge University
Press, Cambridge, U.K., 878 pp.
Reference No. 5: Peixoto, J.P. and A.H. Oort, 1992: Physics of Climate, American
Institute of Physics, New York, N.Y., 522 pp.
Reference No. 6: Trenberth, K.E. (Editor), 1992: Climate System Modeling,
Cambridge University Press, Cambridge, U.K., 788 pp.
Reference No. 7: IPCC, Working Group I, 2001: The Scientific Basis. Contribution
of Working Group I to the Third Assessment Report of the Intergovernmental
Panel on Climate Change. J.T. Houghton and D. Yihui D.J. Griggs, M. Noguer,
P.J. van der Linden, X. Dai, K. Maskell and C.A. Jonhson (Editors), Cambridge
University Press, Cambridge, U.K. and New York, NY, USA, 881 pp.
Adam, K.S., 1995: Vulnerability assessment and coastal management programme

in the Benin coastal zone. In: Preparing to Meet the Coastal Challenges of the 21st
Century, Vol. 2. Proceedings of the World Coast Conference, Noordwijk, 1-5
November 1993, CZM-Centre Publication No. 4, Ministry of Transport, Public
Works and Water Management, The Hague, The Netherlands, pp. 489-497.
Akong’a, J., T.E. Downing, N.T. Konijn, D.N. Mungai, H.R. Muturi, and H.L.
Potter, 1988: The effects of climatic variations on agriculture in central and
eastern Kenya. In: The Impact of Climatic Variations on Agriculture. Vol. 2,
Assessments in Semi-Arid Regions [Parry, M.L., T.R. Carter, and N.T.Konjin (eds.)].
Kluwer Academic Press, Dordrecht, The Netherlands, pp. 123-270.
Alexandersson, H., 1986: A homogeneity test applied to precipitation data. J.
Climatol., Vol. 6, pp. 661-675.
Andreae, M.O., 1995: Climate effects of changing atmospheric aerosol levels. In:
Future Climate of the World: A Modelling Perspective, World Survey and Climatology,
Vol. XVI., A. Henderson-Sellers (ed.)., Elsevier, Amsterdam, 608 pp.
Baethgen, W.E., 1994: Impact of climate change on barley in Uruguay: yield
changes and analysis of nitrogen management systems. In: Implications of
Climate Change for International Agriculture: Crop Modeling Study [Rosenzweig, C.
and A. Iglesias (eds.)]. U.S. Environmental Protection Agency, Uruguay chapter,
Washington D.C., pp. 1-13.
Bangladesh Government, 1993: Assessment of the Vulnerability of Coastal Areas to
Climate Change and Sea-Level Rise: A Pilot Study of Bangladesh. Bangladesh
Government, Dhaka, Bangladesh.
Bilan, D., 1993: The preliminary vulnerability assessment of the Chinese coastal zone
due to sea level rise. In: Vulnerability Assessment to Sea Level Rise and Coastal Zone
Management [McLean, R. and N. Mimura (eds.)]. Proceedings of the IPCC/WCC’93
115
Eastern Hemisphere workshop, Tsukuba, 3-6 August 1993, Department of

Environment, Sport and Territories, Canberra, Australia, pp. 177-188.
Bindi, M., M. Castellani, G. Maracchi, and F. Miglieta, 1993: The ontogenesis of
wheat under scenarios of increased air temperature in Italy: a simulation study.
European Journal of Agronomy, Vol. 2, pp. 261-280.
Braak, C., 1945: Invloed van de wind op regenwaarnemingen. Koninklijk Nederlands
Meteorologisch Instituut. No. 102. Mededelingen en Verhandelingen, No. 48
(102), pp. 7-74.
Broecker, W.S., 1987: The biggest chill. Natural History Magazine, October, pp. 74-
82.
Campbell, G.G. and T.H. Vonder Haar, 1980: “Climatology of radiation budget
measurements from satellites.” Atm. Sci. Paper, No. 323, Dept. Atmos. Sci.,
Colorado State University, Fort Collins, Colorado, 74 pp.
Cambers, G., 1994: Assessment of the vulnerability of coastal areas in Antigua and
Nevis to sea-level rise. In: Global Climate Change and the Rising Challenge of the
Sea [O’Callahan, J. (ed.)]. Proceedings of the third IPCC CZMS workshop,
Margarita Island, 9-13 March 1992, National Oceanic and Atmospheric
Administration, Silver Spring, MD, pp. 11-27.
Ciais, P., P.P. Tans, J.W.C. White, M. Trolier, R.J. Francey, J.A. Berry, D.R. Randall,
P.J. Sellers, J.G. Collatz and D.S. Shimel, 1995: Partitioning of ocean and land
uptake of CO2 as inferred by 13°C measurements from the NOAA Climate
Monitoring and Diagnostics Laboratory Global Air Sampling Network. J.
Geophys. Res., Vol. 100D, pp. 5051-5070.
Climate Prediction Center, 1997: Climate Diagnostics Bulletin, (ed. V. Kousky),
Climate Prediction Center/ NOAA, U.S. Dept. of Commerce, Washington, D.C.,
December Issue, 24 pp.
Cooke, W.F. and J.J.N. Wilson, 1996: A global black carbon aerosol model. J.
Geophys. Res., Vol. 101, No. D14, pp. 19395-19409.
Crowley, T.J., 1989: Paleoclimate perspectives on greenhouse warming. In Climate
and Geo-Sciences, A. Berger et al. (Eds.), Kluwer, Dordrecht, The Netherlands, pp.
179-207.
Dahlström, B., 1986: The improvement of point precipitation data on an operational
basis. Nordic Hydrological Programmeme, NHP-Report No. 17, 86 pp.
Delécolle, R., D. Ripoche, F. Ruget, and G. Gosse, 1994: Possible effects of increas-
ing CO2 concentration on wheat and maize crops in north and southeast
France. In: Implications of Climate Change for International Agriculture: Crop
Modeling Study [Rosenzweig, C. and A. Iglesias (eds.)]. U.S. Environmental
Protection Agency, France chapter, Washington, DC, pp. 1-16.
Delft Hydraulics, Resource Analysis, Ministry of Transport, Public Works and
Water Management and Coastal Research Institute, 1992: Vulnerability
Assessment to Accelerated Sea-Level Rise, Case Study Egypt. Delft Hydraulics,
Delft, The Netherlands.
Dennis, K.C., E.J. Schnack, F.H. Mouzo, and C.R. Orona, 1995a: Sea-level rise and
Argentina: Potential impacts and consequences. Journal of Coastal Research,
special issue Vol. 14, pp. 205-223.
Dennis, K.C., I. Niang-Diop, and R.J. Nicholls, 1995b: Sea-level rise and Senegal:
potential impacts and consequences. Journal of Coastal Research, special issue
Vol. 14, pp. 243-261.
Denton, G.H. and W. Karlén, 1973: Holocene climatic changes, their pattern, and
possible cause. Quat. Res., Vol. 3, pp.155-205.
Dignon, J. and S. Hameed, 1989: Global emissions of nitrogen and sulphur oxides
from 1860 to 1980. JAPCA, Vol. 39, pp. 180-186.
Dignon, J. and S. Hameed, 1992: Emission of nitrogen oxides and sulphur oxides
from the former Soviet Union. Ambio, Vol. 21, pp. 481-482.
Downing, T.E., 1992: Climate Change and Vulnerable Places: Global Food Security and
Country Studies in Zimbabwe, Kenya, Senegal, and Chile. Research Report No. 1,
Environmental Change Unit, University of Oxford, Oxford, UK, 54 pp.
116
REFERENCES
ECMWF, 1998: Internet Web Page <http://www.ecmwf.int/html/seasonal/info/

info.html>, European Centrefor Medium Range Weather Forecasts, Bracknell,
UK., plume diagram link.
Eid, H.M., 1994: Impact of climate change on simulated wheat and maize yields
in Egypt. In: Implications of Climate Change for International Agriculture: Crop
Protection Agency, Egypt chapter, Washington, DC, pp. 1-14.
Eischeid, J.K., C.B. Baker, T.R. Karl and H.F. Diaz, 1995: The quality control of
long-term climatological data using objective data analysis. J. Appl. Meteor.,
Vol. 34, pp. 2787-2795.
Escano, C.R. and L.V. Buendia, 1994: Climate impact assessment for agriculture in
the Philippines simulation of rice yield under climate change scenarios. In
Implications of Climate Change for International Agriculture Crop Modeling Study
[Rosenzweig, C. and A. Iglesias (eds.)]. U.S. Environmental Protection Agency,
Philippines chapter, Washington, DC, pp. 1-13.
Eswaran, H., E. Van den Berg, and P. Reich, 1993: Organic carbon in soils of the
world. Soil Sci. Soc. America J., Vol. 57, pp. 192-194.
Ferguson, M.L. and D.M. Pollack, 1971: Estimating snowpack accumulation for
runoff prediction. Proc., Canadian Hydrology Symp. No. 8, Runoff from Snow and
Ice, Quebec City, pp.7-27.
Fifita, P.N., N. Mimura, and N. Hori, 1994: Assessment of the vulnerability of the
Kingdom of Tonga to sea-level rise. In: Global Climate change and the Rising
Challenge of the Sea [O’Callahan, J. (ed.)]. Proceedings of the third IPCC CZMS
workshop, Margarita Island, 9-13 March 1992, National Oceanic and
Atmospheric Administration, Silver Spring, MD, pp. 119-139.
Fischer, H. and P. Lemke, 1994: On the required accuracy of atmospheric forcing
fields for driving dynamic-thermodynamic sea-ice models. In: The Polar Oceans
and Their Role in Shaping the Global Environment, O.M. Johannessen, R.D.
Muench, and J.E. Overland (eds.), Geophysical Monograph 85, American
Geophysical Union, Washington, pp. 373-381.
Folland, C.K., N. Rayner, P. Frich, T. Basnell, D. Parker, and B. Horton, 2000:
Uncertainties in climate data sets - a challenge for WMO, WMO Bull., 49, 59-68
French, G.T., L.F. Awosika, and C.E. Ibe, 1995: Sea-level rise in Nigeria: potential
impacts and consequences. Journal of Coastal Research, special issue Vol. 14, pp.
224-242.
Friedli, H., H.Lotscher, H. Oeschger, U. Seigenthaler, and B. Stauffer, 1986: Ice core
record of the 13C/12C record of atmospheric CO2 in the past two centuries.
Nature, Vol. 324, pp. 237-238.
Friis-Christensen, E. and K. Lassen, 1991: Length of the solar cycle: An indicator
of solar activity closely associated with climate. Science, Vol. 254, pp. 698-700.
Fröhlich, C. and J. Lean, 1998: In: Proceedings of the IAU Symposium 185, Kyoto,
August 1998, [F.I. Deubner, ed.], Kluwer Academic Publ., Dordrecht, The
Netherlands.
Gates, W.L., 1962: AMIP: The atmospheric model intercomparison project. Bull.
Amer. Met. Soc.,Vol. 73, pp. 1962-1970.
GEWEX, 1998: Internet Web Page <http://www.cais.com/gewex/datasets.html>,
Global Energy and Water Cycle Experiment (GEWEX), Washington, DC.
Goodison, B.E. and P.Y.T. Loule, 1986: Canadian methods for precipitation meas-
urement and correction. WMO/TD-No. 104. Instruments and observing
methods. Report No. 25. Workshop on the Correction of Precipitation
Measurements, Zürich, WMO, pp. 141-145.
Goody, R.M., 1964: Atmospheric Radiation: I. Theoretical Basis. Clarendon, Oxford,
UK., 436 pp.
Goudriaan, J. and M.H. Unsworth, 1990: Implications of increasing carbon
dioxide and climate change for agricultural productivity and water resources.
In: Impact of Carbon Dioxide Trace Gases and Climate Change on Global
Agriculture. ASA Special Publication No. 53, American Society of Agronomy,
Madison, WI, pp.111-130.
117
Groisman, P.Y., V.V. Koknaeva, T.A. Belokrylova, and T.R. Karl, 1991: Overcoming
biases of precipitation measurement: A history of the USSR experience. Bull.
Amer. Meteor. Soc., Vol. 72, pp. 1725-1733.
Han, M., N.Mimura, Y. Hosokawa, S. Machida, K. Yamada, L. Wu, and J. Li, 1993:
Vulnerability assessment of coastal zone to sea level rise: a case study on the
Tianjin coastal plain, North China, by using GIS and Landsat imagery. In:
Vulnerability Assessment to Sea Level Rise and Coastal Zone Management [McLean,
R. and N. Mimura (eds.)]. Proceedings of the IPCC/WCC’93 Eastern
Hemisphere workshop, Tsukuba, 3-6 August 1993, Department of
Environment, Sport and Territories, Canberra, Australia, pp. 189-195.
Harrison, E.F., P. Minnis, B.R. Barkstrom, V. Ramanathan, R.D. Cess and G.G.
Gibson, 1990: Seasonal variation of cloud radiative forcing derived from the
Earth Radiation Budget Experiment. J. Geophys. Res., Vol 95, pp. 18687-18703.
Harrison, M., T. Evans, M. Davey and A. Colman, 1997: A dynamical one-month
lead seasonal rainfall prediction for July to September 1997 for North Africa
from 20°N to the equator. Experimental Long-Lead Forecast Bulletin,Climate
Prediction Center, National Centers for Environmental Prediction,
NWS/NOAA, Washington, D.C., Vol. 6, June, pp. 24-28.
Hasselmann, K., 1988: Some problems in the numerical simulation of climate variability
using high-resolution coupled models. In Physically Based Modelling and Simulation
of Climate and Climate Change, NATO Advanced Study Institute (1986: Erice, Italy),
M.E. Schlesinger (Ed.), Kluwer, Dordrecht, The Netherlands, pp. 583-605.
Hasselmann, K., L. Bengtsson, U. Cubasch, G.C. Hegerl, H. Rodhe, E. Roeckner, H.
v. Storch, R. Ross and J. Waszkewitz, 1995: Detection of anthropogenic climate
change using a fingerprint method. In: Proceedings of Modern Dynamical
Meteorology, Symposium in honour of Aksel Wiin Nielsen, 1995, P. Ditlevsen (ed.),
ECMWF press, Shinfield Park, UK.
Heino, R., 1989: Changes of precipitation in Finland. Proc. Conference on Climate
and Water, Helsinki, Finland, Vation Painatuskeskus, pp. 111-120.
Holthus, P., M. Crawford, C. Makroro, and S. Sullivan, 1992: Vulnerability
Assessment for Accelerated Sea Level Rise Case Study: Majuro Atoll, Republic of the
Marshall Islands. SPREP Reports and Studies Series No. 60, South Pacific
Regional Environment Programme, Apia, Western Samoa.
Horton, B.H., 1995: The geographical distribution of changes in maximum and
minimum temperatures. Atmos. Res., Vol. 37, pp. 101-117.
Howard, J.N., D.L. Burch, and D. Williams, 1955: ‘Near-infrared transmission
through synthetic atmospheres.’ Geophys. Res. Papers No. 40, Geophys. Res.
Dir., Air Force Cambridge Research Center, Cambridge, Massachusetts, 244 pp.
Hoyt, D.V. and E.H. Schatten, 1997: The Role of the Sun in Climate Change, Oxford
University Press, New York, N.Y., 279 pp.
Huq, S., S.I. Ali, and A.A. Rahman, 1995: Sea-level rise and Bangladesh: a
preliminary analysis. Journal of Coastal Research, special issue, Vol. 14, pp. 44-53.
Hulme, M., 1991: An intercomparison of model and observed global precipitation
climatologies. Geophys. Res. Lett., Vol. 22, pp. 1345-1348.
Hulme, M., Z-C. Zhao, and T. Jiang, 1994: Recent and future climate change in
East Asia. Int. J. Climatology, Vol. 14, pp. 637-658.
Iglesias, A. and M.I. Minguez, 1994: Perspectives for future crop water require-
ments in Spain: The case of maize as a reference crop. In: Diachronic Climatic
Changes: Impacts on Water Resources [Angelakis, A. (ed.)]. Springer-Verlag, New
York, NY.
INPE, 1992: Deforestation in Brazilian Amazonia. Instituto Nacional de Pesquisas
Especiais, Sao Paulo, Brazil.
IPCC, 1990: Climate Change: The IPCC Scientific Assessment, [J.T. Houghton, G.J.
Jenkins and J.J. Ephraums (eds.)]. Cambridge University Press, Cambridge, UK.
365 pp.
IPCC, 1992: Climate Change, The IPCC Scientific Assessment, [J.T. Houghton, G.J.
Jenkins and J.J. Ephraums (eds.)], Cambridge University Press, Cambridge, UK,
198 pp.
118
REFERENCES
IPCC, 1992a: Climate Change, The IPCC 1990 and 1992 Assessments (IPCC First
Assessment Report, Overview and Policymaker Summaries and 1992 IPCC
Supplement), World Meteorological Organization, Geneva, Switzerland.
IPCC, 1994: [See IPCC (1995) following. Actual publication year was 1995.]
IPCC, 1995: Climate change 1994: Radiative Forcing of Climate Change and an
Evaluation of the IPCC IS92 Emission Scenarios, [J.T. Houghton, L.G. Meira Filho,
J. Bruce, Hoesung Lee, B.A.Callander, E.F. Haites, N. Harris and K. Maskell
(eds.)], Cambridge University Press, Cambridge, UK. 339 pp. [Note: 1995 is the
actual publication year]
Jaeger, L., 1976: Monatskarten des Niederschlags für die ganze Erde. Ber. Deutschen
Wetterdienstes, Nr. 139, 38 pp.
Jenne, R.L., 1975: Data sets for meteorological research. NCAR Technical Note,
NCAR-TN/1A-111, NCAR, Boulder, Colorado, 194 pp.
Ji, M., A. Kumar, and A. Leetmaa, 1997: Forecasts of tropical Pacific SST using a
comprehensive coupled ocean-atmosphere dynamical model, Experimental Long-Lead
Forecast Bulletin, Climate Prediction Center, National Centers for Environmental
Prediction, NWS/NOAA, Washington, D.C., Vol. 6, June, pp. 10-14.
Jogoo, V.K., 1994: Assessment of the vulnerability of Mauritius to sea-level rise.
Global Climate Change and the Rising Challenge of the Sea [O’Callahan, J. (ed.)].
Proceedings of the third IPCC CZMS workshop, Margarita Island, 9-13 March
1992, National Oceanic and Atmospheric Administration, Silver Spring, MD,
pp. 107-118.
Jones, P.D. , 1994: Hemispheric surface air temperature variations: a reanalysis
and an update to 1993. J. Climate, Vol. 7, pp. 1794-1802.
Jones, P.D. and K.R. Briffa, 1992: Global surface air temperature variations during
the Twentieth century: Part I. Spatial, temporal and seasonal details. Holocene,
Vol. 32, pp. 77-88.
Jones, P.D., C.K. Folland, B. Horton, T.J. Osborn, K.R. Briffa, and D.E. Parker, 2000:
Accounting for sampling density in grid-box surface temperature time series, J.
Geophys.l Res., submitted
Kahn, M. and M.F. Sturm, 1993: Case Study Report Guyana: Assessment of the
Vulnerability of Coastal Areas to Sea-Level Rise. CZM-Centre Publication No. 1,
Ministry of Transport, Public Works and Water Management, The Hague, The
Netherlands.
Karim, Z., M. Ahmed, S.G. Hussain, and Kh.B. Rashid, 1994: Impact of climate
change on production of modern rice in Bangladesh. In Implications of Climate
Change for International Agriculture Crop Modeling Study [Rosenzweig, C. and A.
Iglesias (eds.)]. U.S. Environmental Protection Agency, Bangladesh chapter,
Washington, DC, pp. 1-11.
Karl, T.R., Quayle, R.G., and Groisman, P.Y., 1993: Detecting climate variations
and change: New challenges for observing and data management systems, J.
Clim., Vol. 6, pp. 1481-1494.
Keeling, C.D., R.B. Bacastow, A.F. Carter, S.C. Piper, T.P. Whorf, M. Heimann, W.G.
Mook, and H. Roeloffzen, 1989: A three-dimensional model of atmospheric CO2
transport based on observed winds. 1. Analysis of observational data. Geophys.
Monogr. D.H. Peterson, ed.; American Geophysical Union, Vol. 55, pp. 165-236.
Kelly, P.M., and T.M.L. Wigley, 1992: Solar cycle length, greenhouse forcing, and
global climate, Science, Vol. 360, pp. 328-330.
Kettunen, L., J. Makula, V. Pohjonen, O. Rantanen, and U. Varjo, 1988: The effects
of climatic variation on agriculture in Finland. In: The Impact of Climatic
Variations on Agriculture. Vol. 1, Cool Temperate and Cold Regions [Parry, M.L.,
T.R. Carter, and N.T. Konijn (eds.)]. Kluwer Academic Press, Dordrecht, The
Netherlands, pp. 511-614.
Kiehl, J.T. and K.E. Trenberth, 1997: Earth’s annual global mean energy budget,
Bull. Amer. Meteor. Soc, Vol. 78, pp. 197-208.
Köppen, W. , 1931: Grundriss der Klimakunde, 2nd Ed., Walter de Gruyter and
Co.,Berlin, 131 pp.
LaMarche, V.C., 1974: Paleoclimatic inferences from long treering records. Science,
Vol. 183, pp. 1043-1048.
119
Lamb, H.H., 1969: Climatic fluctuations. In World Survey of Climatology, Vol. 2,

General Climatology, [H. Flohn, ed.], Elsevier Scientific Publishers, Dordrecht,
The Netherlands, pp. 173-249.
Langner, J. and H. Rodhe, 1991: A global three-dimensional model of the tropos-
pheric sulphur cycle. J. Atmos. Chem., Vol. 13, pp. 225-263.
Lemke, P., W.D. Hibbler, G. Flato, M. Harder, and M. Kreyscher, 1997: On the
improvement of sea ice models for climate simulations: Sea ice model inter-
comparison project. Annals of Glaciology, Vol. 25, pp. 183-187.
Le Treut, H., M. Forichon, O. Boucher and Z.X. Li, 1996: Aerosol and greenhouse
gases forcing: Cloud feedbacks associated to the climate response. In: Physical
Mechanisms and their Validation, H. Le Treut (ed.), NATO ASI Series Vol. I.34,
Springer-Verlag, Berlin, pp. 267-280.
List, R. J. (editor), 1951: Meteorological Table, 6th ed., Smithsonian Institute,
Washington, D.C., 527pp.
Liverman, D.M., 1991: Global warming and Climate change in Mexico. Global
Environmental Change, Vol. 1, 351-364.
Liverman, D., M. Dilley, K. O’Brien, and L. Menchaca, 1994: Possible impacts of
climate change on maize yields in Mexico. In: Implications of Climate Change for
International Agriculture: Crop Modeling Study [Rosenzweig, C. and A. Iglesias
(eds.)]. U.S. Environmental Protection Agency, Mexico chapter, Washington,
DC, pp. 1-14.
Lacis, A., J. Hansen, and M. Sato, 1992: Climate forcing by stratospheric aerosols.
Geophys. Res. Lett., Vol. 19, pp. 1607-1610.
Lorenz, E.N., 1969: The predictability of a flow which possesses many scales of
motion. Tellus, Vol. 21, pp. 289-307.
Lorenz, E.N., 1982: Atmospheric predictability experiments with a large numeri-
cal model. Tellus, Vol. 34, pp. 505-513.
Lorenz, E.N., 1990: Can chaos and intransivity lead to interannual variability?
Tellus, Vol. 42A, pp. 378-389.
Lutgens, F.K. and E.J. Tarbuck, 1995: The Atmosphere: An Introduction to
Meteorology, 6th ed. Prentice Hall, Englewood Cliffs, N.J., 430 pp.
Madronich, S. and F.R. de Gruijl, 1993: Skin cancer and UV radiation. Nature, Vol.
366, p. 23.
Manabe, S. and R.F. Strickler, 1964: Thermal equilibrium of the atmosphere with
a convective adjustment. J. Atmos. Sci., Vol. 21, pp. 361-385.
Maochange, C., H. von Storch and E. Zorita, 1995: Coastal sea-level and the large-
scale climate state: A downscaling exercise for the Japanese Islands. Tellus, Vol.
47A, pp. 132-144.
Marland, G., 1989: Fossil fuels CO2 emission: Three countries account for 50 per cent
in 1986. Carbon Dioxide Information Analysis Center Communications, Oak
Ridge, TN, Winter, pp. 1-2.
Matthews, R.B., M.J. Kropff, and D. Bachelet, 1994a: Climate change and rice
production in Asia. Entwicklung und Ländlicherraum, Vol. 1, pp. 16-19.
Matthews, R.B., M.J. Kropff, D. Bachelet, and H.H. van Laar, 1994b: The Impact of
Global Climate Change on Rice Production in Asia: a simulation study. Report No.
ERL-COR-821, U.S. Environmental Protection Agency, Environmental Research
Laboratory, Corvallis, OR.
Michael, E. and D.A.P. Bundy, 1995 [see final reference description following]
Michael, E. and D.A.P. Bundy, 1996: The global burden of lymphatic filariasis. In:
The Global Burden of Diseases: A comprehensive assessment of mortality and disabil-
ity from diseases, injuries, and risk factors in 1990 and projected to 2020 [Murray,
C.J.L. and A.D. Lopez (eds.)]. Harvard School of Public Health on behalf of the
World Health Organization and the World Bank, Harvard University Press,
Cambridge, MA, 990 pp.
Midun Z. and S.-C. Lee, 1995: Implications of a greenhouse-induced sea-level rise:
A national assessment for Malaysia. Journal of Coastal Research, Vol. 14, pp. 96-
115.
Mimura, N, M. Isobe, and Y. Hosokawa, 1993: Coastal zone. In: The Potential
Effects of Climate Change in Japan [Nishioka, S., H. Harasawa, H. Hashimoto, T.
120
REFERENCES
Ookita, K. Masuda, and T. Morita (eds.)]. Center for Global Environmental

Research, Environment Agency, Tokyo, Japan, pp. 57-69.
Mitchell, J.F.B., T.C. Johns, J.M. Gregory, and S.F.B. Tett, 1995: Climate response
to increasing levels of greenhouse gases and sulphate aerosols. Nature, Vol. 376,
pp. 501-504.
Mitchell, J.M., 1963: On the worldwide pattern of secular temperature change. In:
Changes of Climate, Arid Zone Research, UNESCO, Paris, pp. 161-181.
Mitchell, J.M., 1976: An overview of climatic variability and its causal mecha-
nisms. Quaternary Res., Vol. 6, pp. 481-493.
Miyakoda, K., G.D. Hembree, R.F. Strickler, and I. Shulman, 1972: Cumulative
results of extended forecast experiments. I. Model performance for winter
cases. Mon. Wea. Rev., Vol. 100, pp. 836-855.
Muchena, P., 1994: Implications of climate change for maize yields in Zimbabwe.
In: Implications of Climate Change for International Agriculture: Crop Modeling
Study [Rosenzweig, C. and A. Iglesias (eds.)]. U.S. Environmental Protection
Agency, Zimbabwe chapter, Washington, DC, pp. 1-9.
Nakić enović , N, J. Alcamo, G. Davis, B. de Vries, J. Fenhann, S. Gaffin, K.
Gregory, A. Grübler, T.Y. Jung, T. Kram, E.L. La Rovere, L. Michaelis, S. Mori, T.
Morita, W. Pepper, H. Pitcher, L. Price, K. Raihi, A. Roehrl, H-H. Rogner, A.
Sankoviski, M. Schlesinger, P. Shukla, S. Smith, R. Swart, S. van Rooijen, N.
Victor, Z. Dadi, 2000: IPCC Special Report on Emissions Scenarios, Campbridge
University Press, Cambridge, United Kingdom and New York, NY, USA, 599
pp., 2000
NASA, 1995: MTPE EOS Reference Handbook, (G.Asrar and R.Greenstone,eds.),
NASA/Goddard Space Flight Center, Greenbelt, MD, 277 pp.
Neftel, A., E. Moor, H. Oeschger, and B. Stauffer, 1985: Evidence from polar ice
cores for the increase in atmospheric CO2 in the past two centuries. Nature,
Vol. 315, pp. 45-47.
Neumann, C.J., G.W. Cry, E.L. Caso, and B.R. Jarvinen, 1981: Tropical cyclones of
the North Atlantic Ocean, 1871-1980. U.S. Dept. of Commerce, Natl. Climatic
Center, Asheville, N.C., U.S. Government Printing Office, Washington, D.C.,
174 pp.
Nicholls, R.J., 1995: Synthesis of vulnerability analysis studies. In: Preparing to
Meet the Coastal Challenges of the 21st Century, Vol. 1. Proceedings of the World
Coast Conference, Noordwijk, 1-5 November 1993, CZM-Centre Publication
No. 4, Ministry of Transport, Public Works and Water Management, The
Hague, The Netherlands, pp. 181-216.
Nicholls, R.J. and S.P. Leatherman (eds.), 1995a: The potential impact of acceler-
ated sea-level rise on developing countries. Journal of Coastal Research, special
issue Vol. 14, pp. 1-324.
Olesen, J.E., F. Friis, and K. Grevsen, 1993: Simulated effects of climate change on
vegetable crop production in Europe. In: The Effect of Climate Change on
Agricultural and Horticultural Potential in Europe [Kenny, G.J.,P.A. Harrison, and
M.L. Parry (eds.)]. Environmental Change Unit, University of Oxford, Oxford,
UK, pp. 177-200.
Oort, A.H. and H. Liu, 1993: Upper-air temperature trends over the globe, 1958-
1989, J. Climate, Vol. 8, pp. 401-408.
Pachauri, R.K., 1994: Climate Change in Asia: India. Asian Development Bank,
Manila, Philippines.
PAHO, 1994: Leishmaniasis in the Americas. Epidemiological Bulletin, Vol. 15(3),
pp. 8-13.
Parrilla, G, A. Vavin, H. Bryden, M. Garcia, and R. Millard, 1994: Rising tempera-
tures in the subtropical North Atlantic Ocean over the past 35 years. Nature,
Vol. 369, May 5, pp. 48- 51.
Parker, D.E., P.D. Jones, C.K. Folland, and A.C. Bevan, 1994: Interdecadal changes
of surface temperature since the late nineteenth century. J. Geophys. Res., Vol.
99, pp. 14373-14399.
121
Parry, M.L., M. Blantran de Rozari, A.L. Chong, and S. Panick (eds.), 1992: The
Potential Socio-Economic Effects of Climate Change in South-East Asia. United
Nations Environment Programmeme, Nairobi, Kenya.
Peerbolte, E.B., J.G. de Ronde, L.P.M. de Vrees, M. Mann, and G. Baarse, 1991:
Impact of Sea Level Rise on Society: A Case Study for the Netherlands. Delft
Hydraulics and Ministry of Transport, Public Works and Water Management,
Delft and The Hague, The Netherlands.
Pernetta, J.C. and D.L. Elder, 1993: Preliminary assessment of the vulnerability of
Belize to accelerated sea-level rise: difficulties in applying the seven step
approach and alternative uses of available data. In: Vulnerability Assessment to
Sea Level Rise and Coastal Zone Management, [McLean, RF. and N. Mimura
(eds.)]. Proceedings of the IPCC/WCC’93 Eastern Hemisphere workshop,
Tsukuba, 3-6 August 1993, Department of Environment, Sport and Territories,
Canberra, Australia, pp. 293-308.
Pluijm, M., G. Toms, R.B. Zeidler, A. van Urk, and R. Misdorp, 1992: Vulnerability
Assessment to Accelerated Sea Level Rise: Case Study Poland. Ministry of Transport,
Public Works and Water Management, The Hague, The Netherlands.
Potter, C.S., J.T. Randerson, C.B. Field, P.A. Matson, P.M. Vitousek, H.A. Mooney
and S.A. Klooster, 1993: Terrestrial ecosystem production: A process model
based on global satellite and surface data. Global Biogeochem. Cycles, Vol. 7, pp.
811-841.
Qureshi, A. and D. Hobbie, 1994: Climate Change in Asia: Thematic Overview. Asian
Development Bank, Manila, Philippines, 351 pp.
Qureshi, A. and A. Iglesias, 1994: Implications of global climate change for agri-
culture in Pakistan impacts on simulated wheat production. In Implications of
Climate Change for International Agriculture Crop Modeling Study [Rosenzweig, C.
and A. Iglesias (eds.)]. U.S. Environmental Protection Agency, Pakistan chapter,
Washington, DC, pp. 1-11.
Rao, D.G. and S.K. Sinha, 1994: Impact of climate change on simulated wheat
production in India. In Implications of Climate Change for International
Agriculture Crop Modeling Study [Rosenzweig, C. and A. Iglesias (eds.)]. U.S.
Environmental Protection Agency, India chapter, Washington, DC, pp. 1-10
Rasmusson, E.M. and T.H. Carpenter, 1982: Variations in tropical sea surface
temperature and surface wind fields associated with the Southern
Oscillation/El Niño. Mon. Wea. Rev.,Vol. 110, pp. 354-384.
Reynolds , R.W. and T.M. Smith, 1995: A high-resolution global sea surface
temperature climatology, J. Climate, Vol. 8, pp. 1571-1583.
Roeckner, E., T. Siebert, and J. Feichter, 1995: Climatic response to anthropogenic
sulphate forcing simulated with a general circulation model. Aerosol Forcing of
Climate, R.J. Charlson and J. Heintzenberg (eds.), John Wiley and Sons,
Chichester, UK, pp. 349-362.
Ropelewski, C.F. and M.S. Halpert, 1987: Global and regional scale precipitation
patterns associated with the El Niño/Southern Oscillation. Mon. Wea. Rev., Vol.
115, pp. 606-1626.
Rosenzweig, C. and A. Iglesias (eds.), 1994: Implications of Climate Changes for
International Agriculture: Crop Modeling Study. EPA230-B-94-003, U.S.
Environmental Protection Agency, Washington, DC, 312 pp.
Rötter, R. and C.A. van Diepen, 1994: Rhine Basin Study, Vol. 2, Climate Change
Impact on Crop Yield Potentials and Water Use. SC-DLO Report, 85.2,
Wageningen and Lelystad, The Netherlands, 145 pp.
Rotty, R.M. and G. Marland, 1986: Production of CO2 from fossil fuel burning by
fuel type, 1860- 1982. Carbon Dioxide Information Center, Oak Ridge National
Laboratory, Report NDP-006, Oak Ridge, TN, 20 pp.
Sala, O.E. and J.M. Paruelo, 1994: Impacts of global climate change on maize production
in Argentina. In: Implications of Climate Change for International Agriculture: Crop
Modeling Study [Rosenzweig, C. and A. Iglesias (eds.)]. U.S. Environmental Protection
Agency, Argentina chapter, Washington, DC, pp. 1-12.
122
REFERENCES
Santer, B., 1985: The use of general circulation models in climate impact analysis
— a preliminary study of the impacts of a CO2-induced climatic change on
western European agriculture. Climate Change, Vol. 7, pp. 71-93.
Santer, B.D., K.E. Taylor, T.M.L. Wigley, P.D. Jones, D.J. Karoly, J.F.B. Mitchell, A.H.
Oort, J.E. Penner, V. Ramaswamy, M.D. Schwarzkopf, R.J. Stouffer, and S. Tett,
1995: A search for human influences on the thermal structure of the atmos-
phere. PCMDI Report No. 27, Lawrence Livermore National Laboratory,
Livermore, CA, 26 pp.
Sato, M., J.E. Hansen, M.P. McCormick and J.B. Pollack, 1993: Stratospheric
aerosol optical depths, 1985-1990. J. Geophys. Res, Vol. 98, pp. 22987-22994.
Schimel, D.S. and E. Sulzman, 1995: Variability in the earth climate system:
Decadal and longer timescales. Reviews of Geophysics, Supplement July 1995,
pp. 873-882.
Schubert, S.D., C.-Y. Wu, J. Zero, J.-K. Schemm, C.-K. Park, and M. Saurez, 1992:
Monthly means of selected climate variables from 1985 to 1989, NASA Tech.
Memo 104565, Goddard Space Flight Center, Greenbelt, MD, 376 pp.
Schulze, R.E., G.A. Kiker, and R.P. Kunz, 1993: Global climate change and agri-
cultural productivity in southern Africa. Global Environmental Change, Vol 4(1),
pp. 329-349.
Semenov, M.A., J.R. Porter, and R. Delecolle, 1993: Simulation of the effects of
climate change on growth and development of wheat in the UK and France.
In: The Effect of Climate Change on Agricultural and Horticultural Potential in
Europe [Kenny, G.L., P.A. Harrison, and M.L. Parry (eds.)]. Environmental
change Unit, University of Oxford, Oxford, UK, pp. 121-136.
Sevruk, B., 1989: Inhomogeneities in precipitation time series. Proc., Fourth Int.
Meeting on Statistical Climatology, Rotorua, New Zealand Meteorological Service,
New Zealand, pp. 24-27.
Sevruk, B. and S. Klemm, 1989: Catalogue of national standard precipitation guages.
Instruments and Observing Methods, World Meteorological Organization,
Geneva, Switzerland, Report No. 39, 50 pp.
Shea, D.J., 1986: Climatological Atlas: 1950-1979. Surface Air Temperature,
Precipitation, Sea- Level Pressure, and Sea-Surface Temperature. NCAR Technical
Note, NCAR/TN-269+STR, Boulder, CO, 210 pp.
Shea, D.J., K.E. Trenberth and R.W. Reynolds, 1990: A global monthly sea surface
temperature climatology. NCAR Technical Note NCAR/TN-345+STR, Boulder,
CO, 167 pp.
Shackleton, N.J. and N.D. Opdyke, 1973: Oxygen isotope and paleomagnetic
stratigraphy of equatorial Pacific core V28-238: Oxygen isotope temperatures
and ice volumes on a 105 and 106 year scale. Quat. Res., Vol. 3, pp. 39-55.
Siegenthaler, U. and J.L. Sarmiento, 1993: Atmospheric carbon dioxide and the
ocean. Nature, Vol. 365, pp. 119-125.
Siqueira, O.E. de, J.R. Boucas Farias, and L.M. Aguiar Sans, 1994: Potential effects
of global climate change for Brazilian agriculture: applied simulation studies
for wheat, maize, and soybeans. In: Implications of Climate Change for
International Agriculture: Crop Modeling Study [Rosenzweig, C. and A. Iglesias
(eds.)]. U.S. Environmental Protection Agency, Brazil chapter, Washington,
DC, pp. 1-28.
Sivakumar, M.V.K., 1993: Global climate change and crop production in the
Sudano-Sahelian zone of West Africa. In: International Crop Science, vol. I. Crop
Science Society of America, Madison, WI.
Squire, G.R. and M.H. Unsworth, 1988: Effects of CO2 and Climatic Change on
Agriculture: 1988 Report to the UK Department of the Environment. University of
Nottingham, Nottingham, UK.
Stocker, T.F. and A. Schmittner, 1997: Influence of CO2 emission rates on the
stability of the thermohaline circulation. Nature, Vol. 388, pp. 862-865.
Stouffer, R.J., S. Manabe, and K. Ya Vinnikov, 1994: Model assessment of the role
of natural variability in recent global warming. Nature, Vol. 367, pp. 634-636.
Taylor, K. and J.E. Penner, 1994: Climate system response to aerosols and green-
house gases: a model study. Nature, Vol. 369, pp. 734-737.
123
Titus, J.G., R.A. Park, S.P. Leatherman, J.R. Weggel, M.S. Greene, P.W. Mausel, S.
Brown, C.Gaunt, M.Trehan, and G. Yohe, 1991: Greenhouse effect and sea
level rise: potential loss of land and the cost of holding back the sea. Coastal
Management, Vol. 19, pp. 171-204.
Tolmazin, D., 1985: Elements of Dynamic Oceanography. Allen and Unwin,
Winchester, MA, 181 pp.
Tongyai, C., 1994: Impact of climate change on simulated rice production in
Thailand. In Implications of Climate Change for International Agriculture Crop
Protection Agency, Thailand chapter, Washington, DC, pp. 1-13.
Trenberth, K.E. and D.J. Shea, 1987: On the evolution of the Southern Oscillation.
Mon. Wea. Rev., Vol. 115, pp. 3078-3096.
Trewartha, G. T. and L.H. Horn, 1980: An Introduction to Climate, 5th ed., McGraw-
Hill, New York, NY. 416 pp.
United Kingdom Department of the Environment, 1991: United Kingdom Climate
Change Impacts Review Group: The Potential Effects of Climate Change in the United
Kingdom. Climate Change Impacts Review Group, HMSO, London, UK, 124 pp.
Untersteiner, N., 1984: The cryosphere. In The Global Climate, [J.T. Houghton,
ed.], Cambridge University Press, Cambridge, UK, pp. 121-140.
U.S. Department of Commerce, 1963: History of Weather Bureau Precipitation
Measurements. Weather Bureau, Key to meteorological records documentation
No. 3.082, Washington, D.C., 19 pp.
Van der Hammen, T., T.A. Wijmstra, and W.M. Zagwijn, 1971: The floral record of
the late Cenozoic of Europe. In The Late Cenozoic Glacial Ages, K. Turekian, ed.
Yale University Press, New Haven, CT, pp. 391-424.
Volonté, C.R. and J. Arismendi, 1995: Sea-level rise and Venezuela: potential
impacts and responses. Journal of Coastal Research, special issue Vol. 14, pp.
285-302.
Volonté, C.R. and R.J. Nicholls, 1995: Sea-level rise and Uruguay potential
impacts and responses. Journal of Coastal Research, special issue Vol. 14, pp.
262-284.
Vose, R.S., R.L. Schmoyer, P.M. Steurer, T.C. Peterson, R. Heim, T.R. Karl, and J.
Eischeid, 1992: The Global Historical Climatology Network: Long-term monthly
temperature, precipitation, sea level pressure, and station pressure data. Report
ORNL/CDIAC-53, NDP-041 (Available from Carbon Dioxide Information
Analysis Center, Oak Ridge National Laboratory, Oak Ridge, Tennessee.)
Wallace, J.M. and P.V. Hobbs, 1977: Atmospheric Science: An Introductory Survey.
Academic, Press, New York, 467 pp.
Watson, A.J. and J.E. Lovelock, 1983: Biological homeostasis of the global envi-
ronment: The parable of Daisyworld. Tellus, Vol. 35B, pp. 284-289.
Watson, R.T., H. Rodhe, H. Oeschger, and U. Siegenthaler, 1990: Greenhouse gases and
aerosols. In Climate Change: The IPCC Scientific Assessment, J.T. Houghton, G.J. Jenkins,
and J.J. Ephraums, eds., Cambridge University Press, Cambridge, UK, pp. 1-40.
Wheeler, T.R., J.I.L. Morrison, P. Hadley, and R.H. Ellis, 1993: Whole-season exper-
iments on the effects of carbon dioxide and temperature on vegetable crops.
In: The Effect of Climate Change on Agricultural and Horticultural Potential in
Europe [Kenny, G.L., P.A. Harrison, and M.L. Parry (eds.)]. Environmental
Change Unit, University of Oxford, Oxford, UK, pp. 165-176.
Whittaker, R.H., 1975: Communities and Ecosystems. MacMillan, New York, NY,
385 pp.
Wigley, T.M.L. and S.C.B. Raper, 1987: Thermal expansion of sea water associated
with global warming, Nature, Vol. 330, pp.127-131.
Wigley, T.M.L. and S.C.B. Raper, 1992: Implications for climate and sea level of
revised IPCC emissions scenarios, Nature, Vol.357, 293-300.
WMO, 1998: World Climate News, World Meteorological Organization, Geneva,
Switzerland, No. 12 January, 12 pp.
Woodroffe, C.D. and R.F. McLean, 1992: Kiribati Vulnerability to Accelerated Sea-
Level Rise: A Preliminary Study. Department of the Arts, Sport, Environment and
Territories, Canberra, Australia.
124
REFERENCES
Wong, C.S., Y.-H. Chan, J.S. Page, G.E. Smith and R.D. Bellegay, 1993: Changes in
equatorial CO2 flux and new production estimated from CO2 and nutrient
levels in Pacific surface waters during the 1986/87 El Niño. Tellus, Vol. 45B, pp.
64-79.
WHO, 1994a: Progress Report Control of Tropical Diseases. CTD/MIP/94.4, unpub-
lished document.
WHO, 1995b: Action Plan for Malaria Control 1995-2000. Unpublished document.
WHO, 1995c: Chagas Disease: Important Advances in Elimination of Transmission in
Four Countries in Latin America. WHO Press Office Feature No. 183, Geneva,
Switzerland.
WMO, 1997: The First International Conference on Re-analysis, Silver Spring,
Maryland, U.S., 27- 31 October 1997. WMO/TD No. 876, World Meteorological
Organization, Geneva.
Wyrtki, K., 1982: The Southern Oscillation, ocean-atmosphere interaction, and El
Niño. Marine Technol. Sci. J., Vol. 16, pp. 3-10.
125
SUBJECT INDEX
A C
absorption. See also radiation California (ocean current), 15-16
solar, 5-6 carbon cycle
spectrum, 7 annual averages, 16
terrestrial, 7 biosphere interaction, 34
acid rain, 48 numerical modelling, 53, 98
acoustic wave measurements, 86 vegetation relationship, 19
aerosols. See also nitrate aerosols; sulfate aerosols carbon dioxide. See also greenhouse gases
human impact enhancement, 45-48 agricultural influence, 103
observational measurements, 81 annual global emissions, 43
agriculture, 103-107, 112 human impact enhancement, 42-44
air pollution, 110-112 model predictions, 65, 69, 87-88, 90-94
airflow. See winds observational measurements, 80
albedo, 6, 8 oceanic influence, 16
climate feedbacks, 21 sources and reservoirs, 42
cryosphere, 18 cement manufacturing, 42-43
human impact, 49-50 chaos theory, 11, 35-36, 69
land surfaces, 17 chlorofluorocarbons (CFCs), 22, 44-45, 112
meridional profiles, 9 climate
numerical modelling, 52 classifications, 24-26
various surfaces of, 18 definition, 3
vegetation, 34 feedbacks, 21-22
annual cycles, 28-29 geography, 11, 24
astronomical effects, 27-29 global variations, 11-12, 22-23
atmosphere local variations, 23, 25, 75
climate conditions, 10-11 observational data, 81
constituent circulation, 11 predictability, 69-71
forcing factors, 10 regional variations, 24-25
gas absorption spectra, 7 transient features, 10
internal processes, 35 variables, 3
land surface exchange, 17, 20 climate change
numerical modelling, 62-64 definition, 3
oceanic interaction, 53 detection and attribution, 98-99
Atmospheric General Circulation Model recent progress, 99-100
(AGCM), 58, 102 model results, 87-98
Atmospheric Model Intercomparison Project improvements, 97-98
(AMIP), 62, 66 mean conditions, 90-94
predictions, 89-97
recent changes, 87-90
scenarios, 91, 93-95
B variability, 94-95
back radiation, 8-9 numerical modelling, 51-68
Benguela (ocean current), 15-16 perspectives, 1
biomass burning, 45, 47-48. See also deforestation potential impacts, 103-113
biosphere, 19, 21 Climate Information and Prediction Services
climate feedbacks, 22 (CLIPS), 73
interactions, 34-35 climate models. See also numerical modelling
numerical modelling, 52 model comparisons, 89
black body radiation, 7, 17 model evaluation, 58-61
boreal forests, 104 sensitivity, 90
buoy observation measurements, 83 improvements, 66-68
uncertainties, 93
126
SUBJECT INDEX
Upwelling Diffusion and Energy Balance (UD/EB) diseases

model, 57, 89, 93-94, 100-101 agricultural, 103
validation by past climates, 61-62 human, 112
climate monitoring diurnal
clouds, 81 cycles, 28, 63
cryosphere, 82 temperature ranges, 88
deficiencies, 78 downscaling, 75-76
greenhouse gases, 80
improvement strategies, 86
introduction, 78
long-term climate modelling, 78-79 E
oceans, 81-82 Earth Observing System (EOS), 86
precipitation, 84-85 eccentricity (Earth's orbit), 29
principles, 78-79 El Niño - Southern Oscillation (ENSO)
solar radiation, 79 atmospheric interaction, 30-31
surface hydrology, 82 cyclone effects, 109
surface land cover, 82 forecasts, 71-73
surface temperature, 83 global variability, 22-23
climate prediction, 69-77.See also climate change, model results, 95
model results numerical modelling, 64
limitations, 70-71 observational indicators, 32
long-range forecasts, 74 precipitation anomalies, 31, 33
medium-range forecasts, 73-74 electrical field measurements, 86
regional forecasts, 75-76 emission, 5. See also radiation
short-term forecasts, 71-73 Empirical Orthogonal Functions (EOFs), 100
climate system energy budget
components, 3-4 radiative properties, 6-9
atmosphere, 10-11 vegetation, 34
biosphere, 19-21 ensemble forecasting modelling strategy, 70-71
characteristics, 10-21 equation of state, the, 52
cryosphere, 17-18 European Centre for Medium Range Weather
interactions, 30-35 Forecasts (ECMWF), 72, 85
land surface, 17 evaporation. See latent heat
observational data, 81 external forcings, 27-29
ocean, 14-16 extreme events, 95-97. See also severe weather
definition, 3 precipitation, 97
forcing factors, 3 temperature, 96-97
human impacts, 41-50 wind, 96
clouds
aerosol and radiative effects, 48-49
changes due to human impact, 41
numerical modelling, 63, 67, 97 F
observational data, 81 fast system, 74
cold start (climate models), 88-89 fingerprints (spatial patterns), 100
computers, advancement of, 56-57 finite-difference numerical modelling, 54-55
crops. See agriculture food-borne diseases, 112
cryosphere, 17-18 forecasting. See climate prediction; weather prediction
effects, 33-34 modelling
numerical modelling, 53. 65. 74 forecasts
observational data, 82 climate
currents (ocean), 15-16, 33 long-range, 74
medium-range,73-74
regional, 75-76
short-range (weather), 71-72
D ENSO, 71-73
data recovery, 84 forests, 104-105
deforestation, 17, 41, 43, 49. See also biomass fossil fuel combustion, 42-43, 45, 47
burning Fourier transform, 56
desertification, 105-106 fresh water resources, 106-108
127
G Intergovernmental Panel on Climate Change

Gaia model, 34 (IPCC), 87
General Circulation Models (GCM). See numerical current climate modelling, 58
modelling, GCMs emission scenarios, 92-93
geological effects, 30 human impact, 41-42
glaciers. See cryosphere model improvements, 97-98
global climate models, 75-76. See also climate ocean general circulation models, 64
models potential impacts, 103
Global Climate Observing System (GCOS), 78 precipitation uncertainty in models, 84
gravity waves, 35 report summary, 94-95
great ocean conveyor belt, 33 International Satellite Cloud Climatology
greenhouse effect, 7-8 Project (ISCCP), 81
comparison with past climates, 37 Intertropical Convergence Zone (ITCZ), 25, 29
energy balance, 8 intransitive climates, 70
human impacts, 41-45
greenhouse gases. See also under specific type
absorptivity, 7
concentrations, 46-47 K
constituents, 7, 42-43 Kelvin waves, 35
enhancement due to human impacts,41-45, 50, 88 Köppen, W., 24
future scenarios, 91-93, 101 Kuroshio, 15-16
interaction with biosphere, 34-35
lifetimes, 42, 46-47
numerical modelling, 52-53, 98
observational data, 80-81 L
radiative forcing, 41, 44-47 La Niña, 31
sources, 41, 47 land degradation, 105–106
Gulf Stream, 15-16 land surface, 17
exchange with atmosphere, 17
human effects, 17
numerical modelling, 52, 64-65, 67
H observational data, 82
Hadley Circulation, 31 radiative properties, 49
halocarbons, 44-47. See also chlorofluorocarbons; temperature anomalies, 38
hydrofluorocarbons topographical effects, 17
health, human, 109-113 latent heat, 21, 53, 65
heat capacity latitude
cryosphere, 18 energy transfer, 22
land, 17 radiation variations, 9, 29, 63
ocean, 14, 16 Little ice age, 36-37 . See also past climates
heat energy transfer, 21-22 local variability in climate, 23, 75
heat island effect, 83 long-range forecasts
heating (absorption), 6 climate, 74
human impacts, 41-50 weather, 71-73
biosphere, 21, 103 Lorenz, E. N., 35-36, 69
measurements, 41
radiative forcing, 41
summary of radiative energy transfers, 50
hurricanes. See tropical cyclones M
hydrofluorocarbons (HFCs), 44, 46-47. See also Madden-Julian Oscillation, 35, 64
chlorofluorocarbons marine biochemical processes
hydrologic cycle, 10 numerical modelling, 53
mass extinction coefficient, 45, 47
medium-range climate forecasts, 73-74
meteors, 29
I methane, 42-44. See also greenhouse gases
ice cover. See cryosphere Milankovitch, M., 29
index cycles, 35 modelling. See numerical modelling
infrared radiation. See radiation, long-wave monsoons, 23-24, 95
128
SUBJECT INDEX
Montreal Protocol (1987), 44, 46-47 density, 15

mortality rate, 110-112 energy transport, 15
mosquitoes, 112 forcings, 15
multiple attractors, 70 heat capacity, 14, 17, 30
interaction with atmosphere, 53
numerical modelling, 57-58, 6, 73
observational data, 64, 78, 81
N role with greenhouse gases, 16
National Aeronautics and Space Administration salinity, 15-16, 34
(NASA), 86 surface currents, 16
National Center for Atmospheric Research thermohaline currents, 15, 33, 64, 73
(NCAR), 85 time scales, 30-32
National Centers for Environmental Prediction Ocean General Circulation Model (OGCM), 58
(NCEP), 85 optical depth, 45, 47, 49
natural variability, 27-39, 98 orbital parameters, 29
NINO Index, 72. See also El Niño - Southern oscillations (atmospheric), 35
Oscillation (ENSO) ozone, 42
nitrate aerosols, 45, 47. See also aerosols depletion, 22, 44, 112-113
nitrous oxide, 42-44. See also greenhouse gases halocarbon relationship, 44
nonlinear effects (atmospheric), 35 observational data, 80
North Atlantic Oscillation, 40
numerical instability, 56
numerical modelling. See also climate models
approximations, 51, 54-55 P
climate change, 51-68 Paleoclimate Modelling Intercomparison
climate monitoring, 85 Project (PMIP), 62
computer advancement, 56-57 paleoclimates. See past climates
finite-difference method, 54-55 parallel-processor vector machines, 56
flux corrections, 58, 60, 97 past climates, 21, 34, 36-38, 52
GCMs, 63 long-range forecasts, 74
governing equations, 52-53 numerical modelling, 61-62
initial conditions, 68, 88 Peru (ocean current), 15-16
mathematics, 54-56 pests, 103, 105
model evaluation, 58-62, 74 polar ice caps, 34
model variability, 60-61, 64 precipitation. See also rainfall
numerical instability, 56 climatology, 11-12
parameterization, 54, 64, 66-67 ENSO anomalies, 31, 33
physical and chemical interactions, 53 extreme events, 97
sensitivity improvements, 66-67 fresh water resources, 107-108
spatial resolution, 51, 54-55, 67 land surface anomalies, 39
spectral method, 54-56 numerical modelling, 60-62
surface parameters, 52 observational data, 84-85
transport, 56 observed climate variability, 27, 36, 39
use of observational data, 51-52 regional forecasts, 75
variables, 58-60 seasonal forecasts, 71-73, 92
weather prediction, 56, 69 predictability, 69-71
Project for Intercomparison of Land Surface
Parameterization Schemes (PILPS), 64–66
O
obliquity (tilt of Earth's axis), 29 Q
observational measurements Quasi-Biennial Oscillation (QBO), 35
climate forcing factors, 79-84
new techniques, 86
number increase, 86
principles for long-term climate modelling, 78-79 R
observed climate variability, 36-38 radar measurements, 84
ocean, 14-16 radiation, 5-9
decadal variability, 40 absorption spectra, 7
129
albedo, 6, 8-9 relation to ENSO, 22-23, 30-31

energy balance, 8-9 sea surface temperature (SST)
forms, 5 annual anomalies, 38
horizontal variations, 5 climatology, 14–15
human effects, 5 decadal variability, 40
latitude variations, 9 forecasts, 71-73
long-wave, 5, 7 model predictions, 95
meridional profile, 9 observational measurements, 83
net forcing, 9 relation to ENSO, 22-23, 30-31
principles, 5 thermal expansion (sea level rise), 109
short-wave, 5 seasonal forecasts, 71-73
vertical variations, 6 sensible heat, 22, 53, 65
zenith angle, 6 severe weather, 39-40, 109
radiation energy transfer, 21, 48, 63 short-term climate forecasts, 71-73
radiative forcing, 41, 44-47, 50 single-processor machines, 56-57
direct, 48-49 skin cancer, 113
indirect, 48 slow system, 74
numerical modelling, 63 snow, seasonal extent, 19
radiative properties soil moisture, 65, 82, 92. See also deserts and
clouds, 48-49 desertification; land surface
land surface, 49 solar beam, 6
rain gauge measurements, 84 solar luminosity, 34
rainfall. See also precipitation solar radiation, 6 . See also radiation, short-wave
local variability, 25 aerosol effects, 48
severe events, 39 annual cycles, 28-29
reanalysis (climate monitoring), 84-85 black body curves, 7
recent climate change, 87-89 diurnal cycles, 28
regional climate models, 76 energy balance, 8
regional climate predictability, 75 land surface absorption, 17
rehabilitation (data recovery), 84 meridional profile, 9
remote sensing, 82 . See also satellite measurements numerical modelling, 63
sunspots, 27
top of atmosphere, 6
solar-cycle length, 99
S spatial distribution maps, 84
salinity, 15, 34 spatial patterns, 100
satellite measurements spectral numerical modelling, 54–56
cryosphere, 82 SST. See sea surface temperature
International Satellite Cloud Climatology Project statistical downscaling. See downscaling
(ISCCP), 81 storms, 109. See also severe weather; tropical
land surface, 82 cyclones
new systems for climate monitoring, 86 stratosphere
ozone, 80 ozone depletion, 22
precipitation, 84 volcanic effects, 30
sea level, 82 sulfate aerosols, 30, 45, 47-48. See also aerosols
solar radiation, 79 sunspots, 27-28, 80, 99
surface temperature, 83 surface observation stations, 83
water vapour, 80 surface temperature. See temperature, surface
sea ice. See also cryosphere
global inventories, 19
models, 65,67
winter extent over hemispheres, 19 T
Sea Ice Model Intercomparison Project tectonics, 30
(ACSYS), 65 teleconnections, 23, 31, 33
sea level, 34, 40 temperate forests, 104
observational data, 82 temperature
potential impacts from climate change, 109-110 climate feedback, 21
sea level pressure deep ocean, 81-82
numerical modelling, 62 diurnal ranges, 88, 96
130
SUBJECT INDEX
electrical field measurements, 86 V

extreme events, 96 vector-borne diseases, 112
surface vegetation, 19-20
climatology, 11 exchange with atmosphere, 20
global mean changes, 74, 88-95, 99-100 potential impacts from climate change, 103-104
local variability, 25 types, 20
meteor effects, 29 Vienna Convention to Protect the Ozone Layer, 44
numerical modelling, 60-63 volcanoes, 23, 30, 48-49
observational data, 83-84
observed climate variability, 28-29, 36-38
relation to sunspots, 28
seasonal forecasts, 71-73 W
spatial variations, 10 Walker Circulation, 31
temporal variations, 27 water vapour. See also greenhouse gases
vertical changes, 100, 102 numerical modelling, 52-53, 67
terrestrial ecosystems, 20, 103-107 observational data, 80
terrestrial radiation, 7 . See also radiation, long- relation to greenhouse effect, 42
wave water-borne diseases, 112
aerosol effects, 48 weather
black body curves, 7 definition, 3
energy balance, 8 long-range forecasts, 71-73
meridional profile, 9 regimes, 35
numerical modelling, 63 weather prediction modelling, 56, 69
thermal expansion (sea level), 108 winds
Thermal maximum of the 1940's, 36-37. See also climatology, 11-12
past climates numerical modelling, 62
topography, 17, 24-25 WMO World Weather Watch (WWW), 78
transform values, 54-55 World Climate Research Programme
transitive climates, 70 (WCRP), 86-87
tropical cyclones, 96, 109. See also severe weather; World Meteorological Organization
storms (WMO), 1, 73
Atlantic, 40
tropical forests, 104
Y
Younger Dryas cold interval, 36-37. See also past
climates
U
ultraviolet radiation, 112-113
Upwelling Diffusion and Energy Balance Z
(UD/EB) models, 57, 89, 93-94 zenith angle, 6, 28
131

Wmo 100

Uploaded by

Copyright:

Available Formats

Wmo 100

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Wmo 100

Uploaded by

Copyright:

Available Formats

WORLD METEOROLOGICAL ORGANIZATION

ISBN No. 92-63-10926-5

The designations employed and the presentation of material in this publication

CHAPTER 1 UNDERSTANDING THE CLIMATE SYSTEM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

CHAPTER 2 NATURAL TEMPORAL VARIABILITY IN THE CLIMATE SYSTEM . . . . . . . . . . . . . . . . . . 27

CHAPTER 3 HUMAN IMPACTS ON THE CLIMATE SYSTEM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

CHAPTER 4 MODELLING CLIMATE CHANGE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

CHAPTER 5 CLIMATE PREDICTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . ............... 69

CHAPTER 6 OBSERVATIONS FOR LONG-TERM CLIMATE MONITORING . . . . . . . . . . . . . . . . . . . . 78

6.3 Status of selected observations critical for climate change . . . . . . . . . 79

CHAPTER 8 POTENTIAL IMPACTS OF CLIMATE CHANGE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103

CONCLUDING REMARKS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114

UNDERSTANDING THE CLIMATE SYSTEM

Changes in the Atmosphere: Changes in the

Changes in the Ocean: Changes in/on the Land Surface:

Figure 1.2 — Daily total of the

Terrestrial (long-wave) radiation is both emitted and absorbed by material

Figure 1.3 — (a) Black body

Reflected Solar Incoming 235 Outgoing

of the growing season whereas warming of summers in the Sahara, where it is

Figure 1.6 — Global map of (a)

Figure 1.7 — Global distribution

Figure 1.8 — Global

Figure 1.9 — Global distributions

Figure 1.10 — Monthly mean

winter hemisphere is not as rapid as that experienced for land surface

Figure 1.11 — Contours of

Figure 1.12 — A map of the

dynamic and coupled to the Dissolved Intermediate and

radiative energy transfers at the ocean–atmosphere interface. The ocean affects

1.4.4 Land surface is an important interactive component of the climate system. It

Table 1.1 — Albedos for various

Snow and ice

Table 1.2 — Estimated global

Sea ice Max. 18 × 106 2 × 104

Figure 1.14 — Maximum extent

Figure 1.15 — Important Atmosphere

Climate conditions of the atmosphere have a direct effect on the type of

Figure 1.16 — An example of a

The interrelationship between temperature and radiation provides a negative

1.6.3 The El Niño phenomenon in its original definition referred to warmer-than-

Figure 1.17 — Anomalous sea-

Observational data reveals ‘teleconnections’ (significant correlations)

1.6.4 The monsoon refers to quasi-stationary circulation patterns and associated

Figure 1.19 — Asia’s monsoon

NATURAL TEMPORAL VARIABILITY IN THE CLIMATE SYSTEM

Figure 2.1 — Idealized,

Figure 2.2 — Annual mean

and 12 years. Although the presence of a sunspot itself (a relatively darkened

Figure 2.3 — Schematic diagram

Figure 2.4 — Two sets of

The upper set of three panels

Figure 2.5 — Schematic

Figure 2.6 — Schematic diagram

surface which helps to maintain the cool temperatures needed to sustain or

Figure 2.7 — The variations of X

Figure 2.8 — Rough schematic

Figure 2.10 — (a) to (c): (a) Northern Hemisphere

relative to 1961 to 1990