ELEMENTS of

METEOROLOGY and
AIR POLLUTION
Class Notes

Aron Jazcilevich
Centro de Ciencias de la Atmosfera, UNAM

August 16, 2022

Brief Contents

Preface xi

Foreword xiii
1 Urban Population and Air Pollution 1
2 The Atmosphere 31
3 Interactions between solar radiation and the atmosphere 57
4 Interactions between terrestrial radiation and the atmosphere 107
5 The Atmospheric Boundary Layer 133
6 Buoyancy and viscosity in the Atmospheric Boundary Layer 173
7 Transport Processes and Atmospheric Turbulence 197
A Dimensional analysis 231
B State variables, temperature, heat, sensible, and latent heat flux 233
Contents

Preface xi

Foreword xiii

1 Urban Population and Air Pollution 1

1.1 Urban Air Pollutants 2
1.2 Urban Population and Air Pollution Trends 6
1.3 Health effects of air pollution 10
1.4 Natural and anthropogenic drivers of air pollution 15
1.5 Society and Pollution: environmental equality. 16
1.6 “London” and “Los Angeles” smog 20
1.7 The Harvard Six Cities Study 22
1.8 Socioeconomic costs of Air Pollution 24
Exercises 26
Answers to Exercises 27
Chapter Bibliography 28

2 The Atmosphere 31
2.1 Thermal Vertical structure of the atmosphere 32
2.2 Concentration units of chemical species 34
2.2.1 Mixing ratios 34
2.2.2 Mass per volume 36
2.3 Constant chemical constituents of the atmosphere 37
2.3.1 Vertical chemical structure of the atmosphere 39
2.4 Varying chemical constituents of the atmosphere 41
2.4.1 Residence time, sinks and major sources of trace gases 42
2.5 The two facets of ozone 46
2.5.1 Ozone in the stratosphere: a shield against harmful solar radiation 46

–v–
vi CONTENTS

2.5.2 Ozone in the troposphere: harmful photochemical urban pollution 47

2.6 The nonlinearity of photochemistry 49
Exercises 52
Answers to Exercises 55
Chapter Bibliography 56

3 Interactions between solar radiation and the atmosphere 57

3.1 Electromagnetic radiation, quanta and photons 58
3.2 Radiometry concepts 63
3.3 Radiation balance and fundamental laws 71
3.4 The Sun’s electromagnetic field spectrum and its physical effects 76
3.5 Radiation Phenomena in the Atmosphere 79
3.5.1 Radiative Transfer: Beer-Lambert-Bouguer Law. 80
3.5.2 Scattering and Transmission. 87
3.5.3 Absorption and Translation 92
3.6 Attenuation (extinction) of solar radiation by atmospheric gases 96
3.6.1 Main absorptivity effects by atmospheric gases 96
3.6.2 Line broadening 99
3.7 Attenuation (extinction) of solar radiation by atmospheric particles 99
3.7.1 Extinction coefficients of gases and particles 100
3.7.2 Attenuation in typical particle polluted atmospheres 101
3.8 Epilogue 103
Exercises 103
Answers to Exercises 105
Chapter Bibliography 106

4 Interactions between terrestrial radiation and the atmosphere 107

4.1 Earth’s energy budget 108
4.2 Energy Fluxes Near the Surface 111
4.2.1 Energy Balance During a Diurnal Cycle 114
4.2.2 Energy Balance of Surfaces 116
4.2.3 Energy Balance of Canopies 117
4.3 Cooling and Warming in the Surface Layer 120
 CONTENTS vii

4.4 Empirical schemes to establish the surface radiation budget 123

4.5 A simple Model of the Greenhouse Effect 125
4.6 The first description of Greenhouse Effect 127
Exercises 129
Answers to Exercises 131
Chapter Bibliography 132

5 The Atmospheric Boundary Layer 133

5.1 A definition of the Atmospheric Boundary Layer 134
5.2 Convective and Stable Boundary Layer 136
5.3 Dry Atmospheric Boundary Layer 139
5.4 Moist Unsaturated Atmospheric Boundary Layer 146
5.5 Saturated Atmospheric Boundary Layer 151
5.6 Local and Nonlocal Static Atmospheric Stability 151
5.6.1 Local static stability 153
5.6.2 Nonlocal static stability 155
5.7 Pasquill Stability Classes 157
5.8 Pollution, Tropospheric Layers and Inversions 159
5.8.1 Vertical Temp and humidity profiles 161
5.9 A Simplified Pollution Box Model 163
Exercises 169
Answers to Exercises 171
Chapter Bibliography 172

6 Buoyancy and viscosity in the Atmospheric Boundary Layer 173

6.1 A viscous atmosphere 175
6.2 Interaction between buoyancy and mean wind 176
6.3 Wind velocity profiles in near-neutral boundary layers 177
6.3.1 Power-law profile 178
6.3.2 Logarithmic-law profile 179
6.4 Wind profile for non-neutral atmospheres: Monin-Obukhov 183
6.5 The Obukhov length and the gradient Richardson number 185
6.6 Prediction of the mixing height 189
viii CONTENTS

6.6.1 Convective mixing height zic 191

6.6.2 Mechanical mixing height zim 191
Exercises 193
Answers to Exercises 195
Chapter Bibliography 196

7 Transport Processes and Atmospheric Turbulence 197

7.1 Mass conservation, the continuum assumption and incompressibility 198
7.1.1 Continuity Equation. 200
7.1.2 Advection 201
7.1.3 Advection and turbulence 202
7.1.4 Solving the turbulence closure problem 207
7.2 The Advection Diffusion Equation (ADE) 208
7.2.1 Behaviour in time of the advection term 208
7.2.2 Behaviour in time of the diffusion term 209
7.2.3 Behavior in time of the advection and diffusion terms 210
7.3 Solution of the Advection-Diffusion Equation 211
7.3.1 Instantaneous source 211
7.3.2 Continuous source without reflection 213
7.4 The nature of the solution of the Turbulent Advection Diffusion Equation 214
7.5 Pointwise source Gaussian transport 215
7.5.1 The Pasquill-Gifford curves 217
7.5.2 Continuous source with reflection 220
7.6 Continuous source with deposition and settling: The Ermak solution 223
7.7 Effective stack height 224
7.7.1 Briggs calculations 224
7.7.2 Other effective height formulas 227
Chapter Bibliography 230

Appendices

A Dimensional analysis 231

A.0.1 The dimensional matrix 231
 CONTENTS ix

B State variables, temperature, heat, sensible, and latent heat flux 233
Preface

Here is preface.

– xi –
Foreword

Here is a foreword.

– xiii –
Chapter 1

Urban Population and Air

Pollution

1.1 Urban Air Pollutants 2

1.2 Urban Population and Air Pollution Trends 6

1.3 Health effects of air pollution 10

1.4 Natural and anthropogenic drivers of air pollution 15

1.5 Society and Pollution: environmental equality. 16

1.6 “London” and “Los Angeles” smog 20

1.7 The Harvard Six Cities Study 22

1.8 Socioeconomic costs of Air Pollution 24

Exercises 26

Answers to Exercises 27

Chapter Bibliography 28

At the same time that world urbanization seems unstoppable, humanity is facing unprece-
dented environmental threats. While at local level air pollution directly affects human
health and its surrounding environment, at the global level accelerated climate change
due to man made greenhouse effect poses environmental challenges not confronted in
modern human history.

–1–
2 Chapter 1. URBAN POPULATION AND AIR POLLUTION

Will the city of the future, envisioned in Fig. 1.1, provide an opportunity to remedy air
pollution, decrease urban energy demands and adapt to already inevitable environmental
crisis such as extreme weather?

Figure 1.1. A vision of a future city. While population density is increased low-
ering environmental imprint, green spaces are placed to enhance active mobility
through public places. Electric public transportation plays a central role. From:
https://www.hok.com/projects/

Knowledge of the underlying atmospheric physical and chemical phenomena, will

provide guidelines necessary to be followed by other disciplines to accomplish well de-
fined and measurable tasks to confront impending environmental challenges. This book
contributes towards this end.

1.1 Urban Air Pollutants

According to the Health Effects Institute ( HEI ), and the Institute for Health Metrics
and Evaluation (IHME),[1], air pollution was the 4th leading risk factor for early death
worldwide in 2019, surpassed only by high blood pressure, tobacco use, and poor diet.
Air pollution contributed in that year to 6.67 million deaths worldwide.
 1.1 URBAN AIR POLLUTANTS 3

The Environmental Protection Agency (EPA) of the USA defines an air pollutant as any
substance in air that could, in high enough concentration, harm human health and the environment
and cause property damage. Air pollutants can include almost any natural or artificial composition
of matter capable of being airborne-solid particles, liquid droplets, gases, or a combination thereof.
Air pollutants are often grouped in categories for ease in classification; some of the categories
are sulfur compounds, volatile organic compounds, particulate matter, nitrogen compounds, and
radioactive compounds,[2]. We will assume this definition and study this phenomenon
from its onset in the form of emissions, their transport, transformations and impact in the
urban atmosphere.
Among the most important air contaminants present in the urban atmosphere are the
criteria pollutants which are: carbon monoxide (CO), lead (Pb), nitrogen dioxide(NO2 ),
ozone (O3 ), particulate matter (PM), and sulfur dioxide (SO2 ). These are measured and
reported in many cities of the world. Some of these pollutants such as O3 are called sec-
ondary, since they are not emitted directly, but are formed by complex chemical reactions
in the atmosphere.
Particulate matter, or PM, is further classified according to their particle size. This is
defined as the equivalent aerodynamic diameter, or EAD,

ρp
r
D = Dg k , (1.1)
ρ0
where Dg is the geometric diameter, ρp is the density of the aerosol, ρ0 is a reference
density very close to that of water (1 g cm−3 ) and k is the shape factor who is equal to 1
in the case of a perfect sphere. Note that if the density of a spherical aerosol is high, its
equivalent aerodynamic diameter may be larger than its geometrical diameter. In general
ρp is lower than the bulk density of corresponding substances, since the aerosol particulate
is full of pores, cracks, voids, etc. Some photographs showing PM are shown in Fig. 1.2.
4 Chapter 1. URBAN POPULATION AND AIR POLLUTION

Example 1.1
Example: Equivalent Aerodynamic Diameter volume criteria.

The concept of Equivalent Aerodynamic Diameter (EAD) assigns a number to a

complex structure using either its volume, mass, surface area, sedimentation rate,
etc. In the field of air pollution the volume of a particle is a significant quantity,
since this is what many instruments measure instantly via laser Mie technology, as
discussed in Section 3.5. We then illustrate the concept of EAD using the volume
criteria. In the process, an idea of the role played by the shape factor as a reference
to a spherical particle will be understood. This material is based on https://atasci-
entific.com.au/wp-content/uploads/2017/02/AN020710-Basic.... .
a Find the EAD of a cylinder with a diameter D =20 µm and height h =100 µm. In
other words, we want to find the diameter of a sphere which has an equivalent
volume to the cylinder, made with same material density . Since the volume of
the cylinder is,
D2
π h = π102 100 = 10000π [µm3 ], (1.2)
2
and the sought volume of equivalent sphere is

4
V = πX 3 , (1.3)
3
then, s
3 3V
X= , (1.4)
4π
where X is the sphere radius. Then,
r
30000π
= 19.5 [µm3 ] .
3
X= (1.5)
4π
The EAD of the cylinder is then 30 µm3 .
b The height of the cylinder is now 200µm3 , and its base diameter is 20µm3 . Obtain
corresponding EAD.
c Increase aspect ratio and observe a pattern.

In practice, the widely used denominations PM10 , PM2.5 and PM0.5 , refer to particulate
matter with EAD, D, less than 10, 2.5 or 0.5 µm, respectively. Particles in these ranges are
responsible for health effects as they are inhaled into to the human respiratory system.
This will be discussed in Section 1.3.
 1.1 URBAN AIR POLLUTANTS 5

Figure 1.2. Particle photographs obtained using a Scanning Electron Microscope.

Amplification and voltage is shown in each picture. In (a) PM10 spongelike organic
particle, (b) PM2.5 metallic particle, (c) PM2.5 organic particle, and in (d)PM10 aggre-
gate of salts, quartz and mica. Credit: Javier Miranda, Instituto de Física, UNAM.
6 Chapter 1. URBAN POPULATION AND AIR POLLUTION

There are also a large number of compounds which have been determined to be haz-
ardous which are called Toxic Air Pollutants or Hazardous Air Pollutants. Toxic air
pollutants are known or suspected to cause cancer or other serious health problems,
such as reproductive difficulties or birth defects, or other adverse environmental effects.
Among them we have Ammonia (NH3 ), Formaldehide (HCHO ), Benzene, Polycyclic
Aromatic Hydrocarbons (PAH),to mention a few. A complete list can be consulted in
https://www.epa.gov/haps/initial-list-hazardous-air-pollutants-modifications.

1.2 Urban Population and Air Pollution Trends

The year 2007 was the first time in human history that the majority of the world’s popu-
lation lived in urban areas, according to the United Nations. While the urban population
has been rising, the rural population is leveling off and expected to decay, as shown in
Fig. 1.3. This situation is manifested in the growing number of large cities in the world.

Figure 1.3. Urban and rural population from 1500 projected to 2050. Total urban
and rural population, given as estimates to 2016, and UN Projections to 2050. (From
World Urbanization Prospects 2018, United Nations Department of Economic and Social
Affairs/Population Division.)

Figure 1.4 shows a list of megacities (cities with a population of more than 10 million
habitants) for years 1950, 1975, 2003 and 2015. In 1950 there were only two megacities,
 1.2 URBAN POPULATION AND AIR POLLUTION TRENDS 7

both in industrialized societies. In 2003 the expanded list include cities in countries with
less economic economical resources. In total, the world urban population was 3 billion in
2003, and this quantity grew to 4 billion in 2017. All of these people is affected by the air
of the city they live in. In fact, according to the World Health Organization (WHO), air
pollution accounted for 6.7 million premature deaths in 2019,[8]. By premature deaths is
meant deaths that occurs before the average age death in a certain population. This statistic places
air pollution as the main cause of premature deaths above water and lead pollution.
Positive steps have been taken in some cities to improve the air quality. Despite an in-
crease of vehicles and population in the Los Angeles Basin, ozone air quality has improved
substantially over the last 30 years. As shown in Fig. 1.5, during the 1970s, maximum 1-
hour concentrations were around 0.50 parts per million or ppm. These concentration
units and others will be explained in next chapter. In 2008, the maximum measured con-
centrations are less than one-third of that. The 2007 peak 8-hour indicator value of 0.1
ppm was 42 percent lower than the 1988 value. The 2008 three-year average of the max-
imum 8-hour concentration was over 41 percent lower than 1990. The trend for 1-hour
ozone is similar to that for 8 hour and the number of days above the standards has also
declined dramatically from 200 to about 16.

Figure 1.5. Ozone trends in Los Angeles basin. From ARB Almanac 2018 – Chapter 4:
Air Basin Trends and Forecasts – Criteria Pollutants.
 8
T ABLE 7. POPULATION OF URBAN AGGLOMERATIONS WITH 10 MILLION INHABITANTS OR MORE , 1950, 1975, 2003 AND 2015
(millions)
1950 1975 2003 2015
Urban agglomeration Population Urban agglomeration Population Urban agglomeration Population Urban agglomeration Population

1 New York, USA a 12.3 1 Tokyo, Japan 26.6 1 Tokyo, Japan 35.0 1 Tokyo, Japan 36.2
Chapter 1.

2 Tokyo, Japan 11.3 2 New York, USA a 15.9 2 Mexico City, Mexico 18.7 2 Mumbai (Bombay), India 22.6
a
3 Shanghai, China 11.4 3 New York, USA 18.3 3 Delhi, India 20.9
4 Mexico City, Mexico 10.7 4 São Paulo, Brazil 17.9 4 Mexico City, Mexico 20.6
5 Mumbai (Bombay), India 17.4 5 São Paulo, Brazil 20.0
a
6 Delhi, India 14.1 6 New York, USA 19.7
7 Calcutta, India 13.8 7 Dhaka, Bangladesh 17.9
8 Buenos Aires, Argentina 13.0 8 Jakarta, Indonesia 17.5
9 Shanghai, China 12.8 9 Lagos, Nigeria 17.0
10 Jakarta, Indonesia 12.3 10 Calcutta, India 16.8
b
11 Los Angeles, USA 12.0 11 Karachi, Pakistan 16.2
12 Dhaka, Bangladesh 11.6 12 Buenos Aires, Argentina 14.6
13 Osaka-Kobe, Japan 11.2 13 Cairo, Egypt 13.1
14 Rio de Janeiro, Brazil 11.2 14 Los Angeles, USA b 12.9
15 Karachi, Pakistan 11.1 15 Shanghai, China 12.7
16 Beijing, China 10.8 16 Metro Manila, Philippines 12.6
17 Cairo, Egypt 10.8 17 Rio de Janeiro, Brazil 12.4
URBAN POPULATION AND AIR POLLUTION

18 Moscow, Russian Federation 10.5 18 Osaka-Kobe, Japan 11.4

19 Metro Manila, Philippines 10.4 19 Istanbul, Turkey 11.3
20 Lagos, Nigeria 10.1 20 Beijing, China 11.1
21 Moscow, Russian Federation 10.9
22 Paris, France 10.0

NOTES: a Refers to the New York-Newark urbanized areas.

b
Refers to the Los-Angeles-Long Beach-Santa Ana urbanized area.

Figure 1.4. Urban agglomeration with 10 million inhabitants or more, 1950,

1975,2003 and 2015. From World Urbanization Prospects: The 2003 Revision, United
Nations Department of Economic and Social Affairs/Population Division.
 1.2 URBAN POPULATION AND AIR POLLUTION TRENDS 9

In Mexico City, improvements in air quality have been also achieved. For example,
as shown in Fig. 1.6, ozone yearly average concentrations have decreased more than 30%,

Figure 1.6. Pollution trends in Mexico City. From Informe Ejecutivo Febrero 2018,
Secretarìa del Medio Ambiente, GDF.
10 Chapter 1. URBAN POPULATION AND AIR POLLUTION

and PM10 about 40%. According to Rojas-Bracho, et al, in [10] and Evans et al. in [11], by
reducing citywide ambient PM2.5 concentrations from close to 40 µgm−3 in 1990 to close
to 20 µgm−3 in 2014, and simultaneously reducing citywide ambient seasonal hourly peak
ozone concentrations from over 160 ppb in 1990 to 85 ppb in 2015, Mexico City has been
able to reduce the number of deaths attributable to air pollution over this 25-year period
by an estimated 22.5 thousand. Roughly 80 % of the benefits are due to improvements
in PM2.5 .The methodology used to arrive to these statistics will be briefly discussed in
Section 1.8.
Many other cities in the world are highly contaminated but poor measuring and un-
systematic air quality report systems, keep governments and population in the dark. This
hampers a proper assessment of the situation necessary to implement adequate effective
air pollution strategies and resource deployment. According to [4], in-situ PM2.5 mea-
surements are missing for more than 50% of urban population. Employing remote sensing
satellite techniques, they obtained a 21-year time series of ground level PM2.5 concentra-
tions for 4321 urban populations with more than 100,000 habitants, comprising 2.9 billion
people. As shown in Fig. 1.7, sustained PM2.5 year average concentration increases of
48% in South Asian cities were observed, in contrast with decreases of 40% in Chinese
cities. Notwithstanding these trends, many of China’s cities continue to be among the
most polluted cities in the world. Also note important PM2.5 reductions in Eastern US,
Mexico, Western Europe, and East Africa. Reductions were also observed in Colombian
cities and in the Brazilian coast. The case of Lima, Peru, is a counterexample of these
positive trends in South America.

1.3 Health effects of air pollution

Poor air quality principally affects the body’s respiratory and cardiovascular systems. This
is explained because much of the respiratory system consists of open membrane in direct
contact with inspired air. On average an adult has an intake of 20 liter min−1 , and a
human inhales 400 million liters of air in a lifetime. In the lungs, a vital process takes
place: in the alveoli oxygen enters the blood torrent to be transported to the cells.
Carbon dioxide, CO2 , a waste product of cells metabolism, is taken out. This gas
exchange is accomplished by an adjusting control system that comprises the central ner-
vous system, the circulatory system, and the musculature of the diaphragm and the chest.
Therefore pollutants present in the air affects all respiratory tissues that come in contact
with, and may enter the circulatory system.
 1.3 HEALTH EFFECTS OF AIR POLLUTION 11

Figure 1.7. Most PM2.5 polluted cities in 1998 and 2018. Chinese cities were re-
placed mostly by increasingly polluted South Asian cities. Lima, Peru, is in contrast
with improvements in other in South American cities. Reductions in North America
and East Africa are also observed. From [4]

The effects on the respiratory and circulatory systems may be short-term or acute or
long-term or chronic. Examples of short-term effects include irritation to the eyes, nose
and throat, and upper respiratory infections such as bronchitis and pneumonia. Other
symptoms can include headaches, nausea, and allergic reactions. Short-term air pollution
can aggravate the medical conditions of individuals with asthma and emphysema. As an
example of an extreme event where acute effects were experienced by a population, in the
London smog episode of 1952 four thousand people died in a few days due to the high
concentrations of pollution.
12 Chapter 1. URBAN POPULATION AND AIR POLLUTION

Long-term health effects can include chronic respiratory disease, lung cancer, heart
disease, and even damage to the brain, nerves, liver, or kidneys. For example, in a study
carried out in Mexico City inflammatory lung responses are dependent on the geographi-
cal location in the city of PM size, [16]. Furthermore, in 2012 the National Cancer Institute
in USA determined that diesel exhaust exposure causes lung cancer in humans (Group
1)and represents a public health burden, [18]. Effects on the brain due to accumulation
of magnetite has been reported in Maher et al in [17].
As shown in Fig. 1.8 the health effects of air pollution can be seen as a pyramid. The
mildest but more common effects are at the bottom of the pyramid, and the least common
but more severe are at the top. The pyramid demonstrates that as severity decreases the
number of people affected increases.

Figure 1.8. Pyramid of air pollution health effects.

The items contained in the pyramid are endpoints associated with increased air pol-
lution. They are described as follows:
Mortality: All non-accidental mortality causes.
Hospital Admissions: Cardiovascular and Respiratory Hospital Admissions.
Emergency Room Visits: Visit to an emergency department.
Respiratory Symptom Days: Exacerbation of asthma symptoms in individuals with
diagnosed asthma.
 1.3 HEALTH EFFECTS OF AIR POLLUTION 13

Restricted Activity Days: Days spent in bed, missed work days, and days when activi-
ties are partially restricted due to illness. Respiratory-related symptoms such as chest
discomfort, coughing and wheezing.
A short list of more specific health effects of most common air pollutants in modern
cities is shown below.
Ozone – can damage the alveoli, the individual air sacs in the lung where oxygen and
carbon dioxide are exchanged, since at the molecular level ozone attacks substances
in the lung containing C=C bonds. It diminishes lung capacity, produces transient
irritation, coughing, and shortness of breath. Existing respiratory illnesses are aggra-
vated. Epidemiological studies carried in 6 in American cities by Dockery et al. in [7]
that was carried in 95 American urban centers, shows that for periods of high ozone
episodes daily cardiovascular mortality increased by 0.5 percent per 10 ppb increase.
Nitrogen dioxide, NO2 – Besides being a precursor of ozone, causes respiratory prob-
lems such as asthma, emphysema and bronchitis, aggravates existing heart disease
and damages lung tissue.
Sulfur dioxide, SO2 – Besides being a precursor of acid rain, it is a respiratory irritant
and aggravates lung and heart problems.
Particulate matter – Depending of the particle size it can get deep into your lungs (alve-
oli) or enter your blood stream, where, depending on the toxicity of the particle con-
stituents, it can cause cancer. Also it increases respiratory symptoms, such as irritation
of the airways, coughing, or difficulty breathing.
In Fig.1.9 is shown where particles tend to be deposed by human respiratory system,
according to their equivalent aerodynamic diameter. PM10 tend to accumulate in the
upper tract, while PM2.5 is deposed further inward. Fines particles, such as PM0.1 ,
reach the alveolar region.
14 Chapter 1. URBAN POPULATION AND AIR POLLUTION

Figure 1.9. Particle deposition in human lungs depending on their equivalent aero-
dynamic diameter.

Exposure to PM aggravates asthma, develops chronic bronchitis, irregular heartbeat,

nonfatal heart attacks and premature death in people with heart or lung disease. A
metaanalysis that evaluated over a dozen cohort studies summary coefficients showed
that a 1 µ g m−3 increase in annual average PM2.5 concentrations is associated with a
0.6% increment in all-cause mortality, and an 1.1% increment in cardiovascular mor-
tality, [9].

Since the issue of PM2.5 concentrations plays a central role in public health, we devote
to it Section 1.7. There, the Six Harvard Cities Study is described. Other landmark studies
on this subject are by the American Cancer Society Study, [14], and the Adventist Health
Study of Smog, [15].
Although everyone is at risk from the health effects of air pollution, certain sub-
populations are more susceptible. Individual reactions to air contaminants depend on
several factors such as the type of pollutant, the degree of exposure and how much of the
pollutant is present.
Age and health are also important factors. The elderly and people suffering from
cardio-respiratory problems, such as asthma, appear to be the most susceptible groups.
Children and newborns are also sensitive to the health effects of air pollution since they
take in more air than adults for their body weight and consequently, a higher level of
 1.4 NATURAL AND ANTHROPOGENIC DRIVERS OF AIR POLLUTION 15

pollutants. The mortality increases mentioned above are concentrated in children and the
elderly population. Financial income is also very important, leaving the poor at risk, since
healthy diet and access to medical services is limited. People who exercise outdoors on
hot and smoggy days are also at greater risk due to their increased exposure and intake
of pollutants in the air.

1.4 Natural and anthropogenic drivers of air pollution

Air pollution in a city is driven by several factors. Some factors are largely driven by
nature while others are driven by human influence.
The most important natural factor that influences air pollution is the local meteorology
that defines the temperature, humidity, intensity and direction of wind flows in a city.
The regional meteorology and orography will interact with the local meteorology and
together will determine wind flows and therefore the pollutant’s transport, [24],[23],[22].
Therefore a city surrounded by mountains will experience complex air flows that will
transport and affect residence time of pollutants in a different way than would a city
located on a plain.
The meteorology will also influence the height that pollutants reach, or mixing height,
and therefore the size of the volume in which they can disperse and distribute defining
the pollutant concentrations. For example, when a high pressure system is located over
a city, stabler atmospheric conditions will dominate, mixing height will tend to remain
closer to the surface and stagnant winds will hinder the ventilation of pollutants. On the
other hand, if a low pressure system is present, largely unstable conditions will prevail,
mixing heights will tend to be higher and pollutants will be diluted and transported away
by the wind.
Another factor that influences pollution is the height at which a city is located. It
determines the amount of sunlight or actinic flux that reaches its surface. Both sunlight
and temperature are important parameters that influence the level of photochemical pro-
cesses and therefore pollution. Cities like Mexico City and Denver, are examples where
their height exacerbates photochemical pollution.
Also, the presence of a large water body creates wind currents termed land-water
breeze. In some cases these breezes promote beneficial transport and ventilation of pol-
lutants and may cool the city temperature, [20], [19].
Human factors that drive pollution and that therefore are within our control through
research, planning, organization and investment are:
16 Chapter 1. URBAN POPULATION AND AIR POLLUTION

Population growth. As population increases so is the energy use, increasing emis-

sions. This situation can be ameliorated by using energy efficiently.
Land use. The emergence of residence, commercial and industrial spaces is related
with population growth and urbanization. It can severely change natural heat patterns
by increasing the so called Urban Heat Island (UHI) effect due to the heat capacity
of building materials and other factors. In Mexico City urbanization and dissecation
of large lake areas are responsible for an increase of 2.2 C in mean yearly temperature
in the 20th century, [20]. Also, the disappearance of these water bodies has decreased
the formation of land-water breezes formed by the temperature contrast between land
and water masses that could provide local ventilation, lowering the concentrations of
atmospheric pollutants.
Edification. The presence of large buildings besides enhancing the UHI influences
wind flows either by presenting a drag force to wind currents or by creating Venturi
or canyon effects between buildings.
Industrial, commercial and transportation emissions. These activities affect the air
pollution levels in a city by injecting into the urban atmosphere compounds that
directly or through chemical reactions affect human health and welfare or provoke
harmful environmental effects.
Interaction between urban and surrounding land use emissions . Forest and agricul-
tural fire emissions directly affect contiguous and even far away cities. For example,
New Delhi in India, has reported extreme pollution episodes due to seasonal agri-
cultural fires, [21]. Perhaps a more subtle interaction between an urban center and
surrounding forest is described in [22], where is shown that highly reactive hydro-
carbons emissions from vegetation are combined with anthropogenic nitrogen oxides
increasing surface ozone formation, aggravating air pollution episodes.
The interconnection of the meteorological and anthropogenic drivers mentioned here are
among the main topics in this text.

1.5 Society and Pollution: environmental equality.

In the field of environmental economy a model relating economic growth and pollution
levels is called the Environmental Kusnets Curve (EKC). It states that as a society increases
their GDP, there is an initial deterioration of the environment until a minimum is reached
followed by a recovery. It is argued that the inflection point is when annual income per
capita reaches about $14,000 USD of 2022. EKC seems to describe the situation for some
atmospheric pollutants such as sulphur dioxide (SO2 ) but not for other pollutants such
 1.5 SOCIETY AND POLLUTION: ENVIRONMENTAL EQUALITY. 17

as Carbon Dioxide (CO2 ). This and other inconsistencies raise questions about its validity
on different grounds.
Nevertheless, the EKC virtues and faults provides grounds for a brief discussion on
the relation of society and pollution. Atmospheric pollution is strongly related with the
available information and sensitivity of the population to this problem. Both of these are
accentuated as larger portions of the society becomes more affluent and in a democratic
society, more participatory.
Also as the economy develops, more resources are available to attack the problem and
employment shifts from heavy industry to less polluting ”high technology” industries
and services. Through knowledge, know-how, norms and legal battles many cities have
improved their air quality, at least when an informed society has recognized the dangers
involved to their health. That is probably why EKC works for an irritant and noxious
pollutant like SO2 or visually observable PM. On this regard, cities of the world with
higher income tend to have less PM2.5 year average concentrations, than lower income
cities. This is shown in Fig. 1.10.
18 Chapter 1. URBAN POPULATION AND AIR POLLUTION

Figure 1.10. Scatterplot comparing the annual-median of available World Health

Organization (WHO) PM2.5 monitors with corresponding remote sensing (RS) esti-
mates of annual average PM2.5 estimate. A total of 5626 annual average estimates
from 1998 to 2018 were available for 1007 cities with 1.2 billion inhabitants. Dashed
line indicates best fit regression relationship (R2 =0.91), slope 0.91. Marker propor-
tional to logarithm of city population. From,[4].

Since CO2 does not directly affect health, not only their emissions in many cities have
not decreased, but they are on the rise. Is this the reason why EKC fails here? Does the
recognized effect on world climate of CO2 will be enough to bring a change? Or simply
an economic need for more efficient use of energy will bring along a decrease in CO2
emissions?
Another aspect of EKC is whether it is applicable to a city, region, country or globally.
For example, if a city or country closes highly pollutant industries but transfers them to
another country with less stringent norms, EKC may be fulfilled locally but not globally,
even in a worldwide rising economy. See discussion in [3].
A more substantial criticism to EKC is weather national GDP is a valid parameter,
or it it has to be used in conjunction with other criteria to measure societal welfare. For
example the Organization for Economic Co-operation and Development (OECD), pro-
poses indexes such as housing, income, education, health and civic engagement, among
 1.5 SOCIETY AND POLLUTION: ENVIRONMENTAL EQUALITY. 19

other criteria. Many times, poor communities are in the receiving end of pollution. These
type of parameters that promote equality in a society, must be stressed when proposing
national economic and development policies to improve quality of life, which results in
environmental well being.
Clear examples of inequality have been reported in [4]. They show that in the U.S.,
nearly in all major emission categories across states, urban and rural areas, income levels,
and exposure levels—contribute to the systemic PM2.5 exposure disparity as experienced
by racial-ethnic minorities. This is shown in Fig. 1.11.

Figure 1.11. Source contributions to racial-ethnic disparity in PM2.5 exposure. (A to

E) Individual source type (n = 5434 source types) contributions to exposure (y axis)
and % exposure disparity (x axis, truncated at 200%, positive values are shaded
red, negative values are shaded blue), with dashed lines denoting percent expo-
sure caused by sources with positive exposure disparity. (F to J) Sources in (A) to (E)
grouped into source sectors (n = 14 groups) and ranked vertically according to abso-
lute exposure disparity, proportional to the area of each rectangle. As shown in (B),
people of color, or POC, experience greater-than-average exposures from source
types causing 75% of overall exposure. From: [4]
20 Chapter 1. URBAN POPULATION AND AIR POLLUTION

We learn from this discussion that atmospheric pollution is also driven by human con-
duct, economics and politics. Resulting environmental policies can be objectively gauged
using the knowledge provided in this work.

1.6 “London” and “Los Angeles” smog

In December of 1952 there were 4000 excess deaths in London. By excess deaths is meant
deaths beyond what is expected for a location and time of year based on previous statistics. Al-
though the specific causes for these deaths were not identified at the time, it is assumed
that high concentrations of sulphur dioxide (SO2 ) and particles present in the London’s
fog in combination with a stable atmosphere during this episode were responsible for
this situation. Fig. 1.12 shows how London looked during day hours while this episode
lasted.
Fig. 1.13 shows the SO2 and smoke concentrations as well as the excess death rate
during this episode. The World Health Organization (WHO) exposure mean level limits
for SO2 are 20 µgm−3 or 75 ppb in 24 hours and no more than 500 µgm−3 or 187.5 parts per
billion or ppb in 10 minutes. For suspended particles (smoke) it is 90 mgm−3 . Adverse
health effects will appear in the population if these limits are exceeded. Note that London
concentration levels were several times above the recommended levels.
Another city famous for smog episodes is Los Angeles, California. In the early 1940’s a
phenomenon attracted the attention of its inhabitants when eye watering and plant killing
pollutants were noticed. As opposed to the London smog case, these episodes took place
during warm and sunny days. Eventually it was discovered that this was a new form of
air pollution,[5].
Subsequent studies showed that the damage to the plants could be reproduced in the
laboratory by performing the following chemical experiment:

VOC0 s + NOx + sunlight → O3 + other products (1.6)

where by VOC’s is meant volatile organics compounds such as alkenes in gas phase, NOx
are nitrogen oxides (NOx = NO+NO2 ), and O3 is ozone. The reactive compounds to the
left of the formula are found in car exhausts and other combustion processes, pointing to
the sources of the Los Angeles smog,[6].
The fact that sunlight is necessary to carry out the reaction means that the process
is photochemical. The Los Angeles smog gave a tremendous importance to the area
of photochemistry. New compounds in the ”other products part ” of formula 1.6 such
as peroxyacetil nitrates (PAN’s) not found before in nature, were found in Los Angeles
 1.6 “LONDON” AND “LOS ANGELES” SMOG 21

Figure 1.12. Daytime scene in London during the smog episode of December 1952 .

Figure 1.13. Excess deaths, smoke and SO2 concentrations on December 1952 in London.
22 Chapter 1. URBAN POPULATION AND AIR POLLUTION

smog. Many cities in the world suffer from the same type of smog: in Mexico City O3
concentrations above 400 ppm have been recorded in the late 1980’s.

1.7 The Harvard Six Cities Study

The results of the Harvard Six Cities Study reported by Dockery and colleagues in 1993,
[7], played a critical role in the establishment of the U.S. ambient air quality objective
for PM2.5 of 15 µg m−3 in 1997. Such is its importance that it was audited and revised
confirming its findings in [13]. It is a landmark study that had world level implications
since it established an association between urban air pollution and mortality. This work
was a prospective cohort study which is a type of medical research method used to
investigate the causes of disease and to establish links between risk factors (PM2.5 urban
pollution) and health outcomes (excess mortality) which in this case were mainly due to
pulmonary and cardiopulmonary diseases. This study used a control for tobacco smoking
and other health risks by taking into account individual risk factors.
The word cohort means a group of people. In this case the groups of people were
statistically representative samples of inhabitants of six American cities: Watertown, Mas-
sachusetts; Harriman, Tennessee; St. Louis Missouri; Steubenville, Ohio; Portage and
Pardeeville, Wisconsin ; and Topeka, Kansas. The sample was restricted to the 8,111 white
subjects who were 25 through 74 years of age at the time of enrollment in the study, had
undergone spirometric testing and had completed a standardized questionnaire. Spirom-
etry measures lung function, specifically the amount (volume) and/or speed (flow) of air
that can be inhaled and exhaled to assess breathing patterns that identify conditions such
as asthma, pulmonary fibrosis and Chronic Obstructive Pulmonary Disease (COPD). The
time span of the study was from mid 70’s to late 80’s depending on city. The question-
naire included queries about age, sex, weight, height, education level, complete smoking
history, occupational exposures and medical history.
The main results are shown in Fig. 1.14. After adjusting for smoking and other risk
factors, robust association between mortality fine particles (PM2.5 ) and air pollution levels
was established: the higher fine PM concentration is in a city, the higher the mortality
rate. Steubenville clearly leads in particle concentration and so does in mortality rates,
while the least polluted Portage, has lowest mortality rates. All cities line up in between
as a function of fine PM pollution, including sulfates. It should be noted that smoking
was above all the leading factor for mortality in this study.
 1.7 THE HARVARD SIX CITIES STUDY 23

Figure 1.14. In (a), annual average concentrations of total particles, fine particles,
and sulfate particles in the six cities. In (b), estimated adjusted mortality-rate ratios
and pollution levels in the six cities:P denotes Portage, T Topeka, W Watertown, L St.
Louis, T Tennessee and S Steubenville. From Harvard Six Cities Study, [3].
24 Chapter 1. URBAN POPULATION AND AIR POLLUTION

1.8 Socioeconomic costs of Air Pollution

A procedure to estimate social costs due to health effects by exposure to air pollution is
presented. Specialized models have been developed for this task, such as the Environ-
mental Benefits and Analysis Program or BenMap, by the EPA. The material presented
in this subsection is based on this model.
The Health Effect, ∆y, is defined as the estimate of how much changes in air pollution
levels influence a type of sickness or death rate of a given exposed population. To obtain this
number, the Health Impact Function, (HIF), is evaluated. It has four components: the
affected population, the change in air quality, the baseline incidence rate, and the effect
estimate drawn from epidemiological studies (references). The most used HIF formula
is:  
∆y = y0 · eβ·∆x − 1 · P , (1.7)

where y0 is the baseline incidence rate for the health impact being quantified (for example
changes in mortality would use the baseline, or background, mortality rate for the given
population ); P is the population affected by the change in air quality; ∆x is the change
in air quality; and β is the effect coefficient obtained from the epidemiological study. The
adverse effects that the HIF quantify are included in the pyramid of Fig. 1.8: mortality or
premature deaths, hospital admissions, emergency room visits, asthma symptom days,
restricted activity days and acute respiratory symptoms.
Once the health effect of a control policy is obtained, its economic value, Ev , can be
obtained as follows,
Ev = ∆y ∗ Ve (1.8)

where Ve is an estimate of the economic value per case. There are different ways of
calculating the value of the health effect. For example, the value of an avoided premature
mortality is generally calculated using the Value of Statistical Life ,VSL. The value of a
statistical life is the monetary value that people are willing to pay to slightly reduce the
risk of premature death. For other health effects, the medical costs of the illness may be
the only valuation data available.
 1.8 SOCIOECONOMIC COSTS OF AIR POLLUTION 25

Example 1.2
Example: Obtaining health effect due to reduction of particulate matter, PM2.5 .

a Using Eqs. 1.7 and 1.8 together with the following typical values, estimate the
cost benefit of reducing in a locality by 10 µgm−3 ambient PM2.5 concentrations.
The locality has 500,000 inhabitants, a lower chronic respiratory mortality rate
of 47.1 in 100,000 inhabitants. From the Harvard Six Cities study we suppose a
0.6% mortality increment per 1 µg m−3 . We assign a statistical economical value
per case of 1 million USD. Then,

47.1  
∆y = · e−0.6·10 − 1 · 500, 000 = −235 . (1.9)
100, 000
Therefore, by lowering ambient PM2.5 concentrations by 10 µgm−3 , the health
impact in that locality is of 235 less mortality cases. This is only considering
lower chronic respiratory ailments.
b Same as above, but now consider a 10 µgm−3 increment in ambient PM2.5 con-
centrations.
c Consider same locality, but obtain benefit by lowering 10 ppb of ozone ambient
concentrations.
26 Chapter 1. URBAN POPULATION AND AIR POLLUTION

Exercises

1.1 Health Impact Estimates

Why health impact estimates Hee of Section 1.3 vary from country to country?
(Answer on Page 27)

1.2 Variation in Health Impact Estimates

Why health impact estimates Hee of Section 1.3 vary between PM10 and PM2.5 ?
(Answer on Page 27)

1.3 Which contaminants are most harmful?

Based on the Hee of Section 1.3, which contaminant is more harmful to human health?
(Answer on Page 27)

1.4 Health impacts for 1952 London Smog

Obtain He for the London smog episode. For Ep use 8 million, the approximate population
in London in the 1950’s.
(Answer on Page 27)

1.5 Obtain EV for He

Obtain Ev for the He obtained above. For Ve use $400,000 USD, an approximate value
used for cardiopulmonary mortality.
(Answer on Page 27)
 EXERCISES 27

Answers to Exercises

Answer to Exercise 1.1 (page 26)

The variation is accounted for by the north-south axis of the country’s location as well as
its height above sea level.

Answer to Exercise 1.2 (page 26)

Health impacts differ according to the PM level measured by the higher fine PM concen-
tration is in a city, the higher the mortality rate.

Answer to Exercise 1.3 (page 26)

Among the most important air contaminants present in the urban atmosphere are the
criteria pollutants which are carbon monoxide (CO), lead (Pb), nitrogen dioxide(NO2 ),
ozone (O3 ), particulate matter (PM), and sulfur dioxide (SO2 ).

Answer to Exercise 1.4 (page 26)

Hint: He = Aq ∗ Hee ∗ Ep ∗ Hbi

Answer to Exercise 1.5 (page 26)

Ev for the He above is approximately 42 according to Douglas Adams.
28 Chapter Bibliography

Chapter Bibliography

[1] Health Effects Institute and Institute for Health Metrics and Evaluation. State of Global Air
2020. https://www.stateofglobalair.org/, consulted: august, 11, 2022. HEI, IHME, 2022.
[2] USA Environmental Protection Agency. ROE Glossary. EPA, 2022.
[3] David I. Stern. The environmental kuznets curve�. In Reference Module in Earth Systems and
Environmental Sciences. Elsevier, 2018.
[4] Christopher W. Tessum, David A. Paolella, Sarah E. Chambliss, Joshua S. Apte, Jason D. Hill,
and Julian D. Marshall. Pm_2.5 polluters disproportionately and systemically
affect people of color in the united states. Science Advances, 7(18), 2021. DOI:eabf4491.
[5] John T. Middleton, James Blair Kendrick, and H. W. Schwalm. Injury to herbaceous plants by
smog or air-pollution. Plant disease reporter, 34:245–252, 1950.
[6] A. J. Haagen-Smit and M. M. Fox. Photochemical ozone formation with hydrocarbons and
automobile exhaust. Air Repair, 4(3):105–136, 1954.
[7] Douglas W. Dockery, C. Arden Pope, Xiping Xu, John D. Spengler, James H. Ware, Martha E.
Fay, Benjamin G. Ferris, and Frank E. Speizer. An association between air pollution and
mortality in six u.s. cities. New England Journal of Medicine, 329(24):1753–1759, 1993.
PMID: 8179653.
[8] World Health Organization. WHO global air quality guidelines: particulate matter (PM2.5
and PM10), ozone, nitrogen dioxide, sulfur dioxide and carbon monoxide: executive summary
WHO, 2021.
[9] Gerard Hoek, Ranjini M Krishnan, Rob Beelen, Annette Peters, Bart Ostro, Bert Brunekreef,
Joel D Kaufman. Long-term air pollution exposure and cardio-respiratory mortality:a review.
Environ Health, 2013;12(1):43
[10] Harvard T.C. Chan School of Public Health, SEDEMA-CDMX. Historical Analysis of Popu-
lation Health Benefits Associated With Air Quality in Mexico City During 1990 and 2014.
[11] Evans, J.S., Rojas-Bracho, L., Hammitt, J.K. and Dockery, D.W. (2021). Mortality Benefits
and Control Costs of Improving Air Quality in Mexico City: The Case of Heavy Duty Diesel
Vehicles. Risk Analysis, 41: 661-677. https://doi.org/10.1111/risa.13655.
[12] doi:10.1126/sciadv.abf4491, Christopher W. Tessum and David A. Paolella and Sarah E.
Chambliss and Joshua S. Apte and Jason D. Hill and Julian D. Marshall PM_2.5
polluters disproportionately and systemically affect people of color in the United States, Sci-
ence Advances,18, Vol. 7, 2021. Tessum20211,
[13] Krewski D, Burnett R, Goldberg MS, Hoover K, Siematicky J, Jerrett M, Abrahamowicz M,
White WH. Reanalysis of the Harvard Six Cities study and the American Cancer Society study
of particulate air pollution and mortality. Health Eff Inst Special Rep 2000;July. 97pp
[14] C A Pope, D V Bates, M E Raizenne Health effects of particulate air pollution: time for
reassessment? Environ. Health Perspect. 103(5);472-480, 1995.
[15] Abbey DE, Nishino N, McDonnell WF, Burchette RJ, Knutsen SF, Lawrence Beeson W, Yang
JX. Long-term inhalable particles and other air pollutants related to mortality in nonsmokers.
Am J Respir Crit Care Med. 1999 Feb;159(2):373-82. doi: 10.1164/ajrccm.159.2.9806020. PMID:
9927346.
[16] Osornio Vargas, Alvaro , Bonner, James , Alfaro-Moreno, Ernesto , Martínez, Leticia , García-
Cuellar, Claudia , Rosales, Sergio, Miranda, Javier, Rosas, Irma. Proinflammatory and Cyto-
toxic Effects of Mexico City Air Pollution Particulate Matter in Vitro Are Dependent on Particle
Size and Composition. Environ. Health Perspect., 2003. 111. 1289-93. 10.1289/ehp.5913.
 Chapter Bibliography 29

[17] Maher BA, Ahmed IA, Karloukovski V, MacLaren DA, Foulds PG, Allsop D, Mann DM,
Torres-Jardón R, Calderon-Garciduenas L. Magnetite pollution nanoparticles in the hu-
man brain. In Proc Natl Acad Sci U S A. 113,39,pp.10797-801, Sep 27, 2016. doi:
10.1073/pnas.1605941113.
[18] World Health Organization, International Agency for Research on Cancer. Press Release 213.
June 12, 2012.
[19] Gillon, J. The water cooler. Nature 404, 555, (2000). https://doi.org/10.1038/35007172
[20] Jazcilevich, A., Fuentes, V., Jauregui, E., Luna, E. Simulated Urban Climate Response to
Historical Land Use Modification in the Basin of Mexico. Climatic Change 44, 515–536 (2000).
https://doi.org/10.1023/A:1005588919627.
[21] Nandita D. Gnaguly, Chris G. Tzanis, Kostas Philippopoulos,Despina Deligiorgi. Analysis
of a severe air pollution episode in India during Diwali festival – a nationwide approach
Atmósfera , vol. 32, no. 3, pp. 225-236, 2019.
[22] Aron D. Jazcilevich, Agustín R. García, Ernesto Caetano, Locally induced surface air
confluence by complex terrain and its effects on air pollution in the valley of Mex-
ico. Atmospheric Environment , Volume 39, Issue 30,2005, Pages 5481-5489,ISSN 1352-2310,
https://doi.org/10.1016/j.atmosenv.2005.05.046.
[23] Benjamin de Foy„ A. Clappier, L. T. Molina, and M. J. Molina Distinct wind convergence
patterns in the Mexico City basin due to the interaction of the gap winds with the synoptic
flow. Atmos. Chem. Phys., 6, 1249–1265, 2006. www.atmos-chem-phys.net/6/1249/2006/.
[24] Aron D. Jazcilevich, Agustín R. García, Ernesto Caetano, A modeling study of air pollu-
tion modulation through land-use change in the Valley of Mexico. Atmospheric Environment,
Volume 36, Issue 14, Pages 2297 - 2307, 2002.
Chapter 2

The Atmosphere

2.1 Thermal Vertical structure of the atmosphere 32

2.2 Concentration units of chemical species 34

2.3 Constant chemical constituents of the atmosphere 37

2.4 Varying chemical constituents of the atmosphere 41

2.5 The two facets of ozone 46

2.6 The nonlinearity of photochemistry 49

Exercises 52

Answers to Exercises 55

Chapter Bibliography 56

The height of the upper most layer of the atmosphere represents only about 2.3 percent
of the Earth’s average diameter of 12,742 km. If the Earth would be the size of an apple,
then the atmosphere would be as thick as the apple’s peel. Nevertheless, this thin layer of
gases maintains the Earth’s temperature within ranges necessary for life. It also protects
us from harmful ultraviolet radiation and is where wind, rain and storms take place.
Perhaps, an appreciation of the atmosphere is better conveyed by the photograph taken
from the International Space Station, ISS, shown in Fig. 2.1.
As shown in the ISS’s photograph, the atmosphere is a strongly stratified fluid due
to the action of gravity. As we ascend severe changes are noticed in pressure, tempera-
ture, density, and composition. This stratification sets apart atmospheric flows from other

– 31 –
32 Chapter 2. THE ATMOSPHERE

Figure 2.1. View of the atmosphere from the International Space Station. In the
forefront, active cloud structures. In the background, shades of blue appear, point-
ing out the atmosphere’s layered character. Credit: Image Science and Analysis Lab-
oratory, NASA Johnson Space Center.

fluid studies in engineering and physics. Some of its most important implications will be
introduced here.
Besides specific citations, the material consulted on atmospheric chemistry and physics
is contained in the well known texts [11],[11],[1], [12] and[13].

2.1 Thermal Vertical structure of the atmosphere

At left of Fig. 2.2 is shown the vertical profile of temperatures based on the International
Standard Atmosphere, ISA, [2]. This is a static average of the atmosphere and is used as
reference. Based on the thermal variation of the atmosphere with height, the following
layers are formed:
The troposphere is the lowest thermal layer. It is in this layer that the important
meteorological phenomena collectively called weather takes place. In a standard at-
mosphere the temperature decreases linearly with height at a rate of 6.5 ◦ C per km.
This rate is called the environmental lapse rate. In shallow layers within the tropo-
sphere, the lapse rate is quite variable and is common that under certain conditions the
temperature may rise with height forming a temperature inversion. Also, its depth
 2.1 THERMAL VERTICAL STRUCTURE OF THE ATMOSPHERE 33

Figure 2.2. At left vertical temperature structure of the atmosphere and at right
its vertical structure by chemical composition. At center is shown the change in
molecular weight of air (Md ) due to composition. Adapted from [1].

varies with geographical location: in the tropics may reach 15 km, while in the polar
regions it reaches 9 km.
In the stratosphere the temperature remains constant until a height 20 km, where
temperature rises up to the stratopause encountered at a height of 50 km. The reason
for the rise of temperature is the presence of stratospheric Ozone and its corresponding
absorption of ultraviolet radiation from the Sun. This phenomena will be discussed
in Chapter 2, Section 2.5.
In the mesosphere localized at about a height of 80 km, the temperature decreases
with height to about −90◦ C. The air pressure at the bottom of this layer is below 1
percent of the pressure at sea level.
The highest thermal layer called the thermosphere reaches a height of about 150 km.
The relative high temperature of its extremely rarefied air found here is the result of
absorption of very short wavelength solar energy by atoms of oxygen and nitrogen.
The temperature rises to 1000◦ C. Since the probability of a fast moving particle col-
liding with an object is very low, the energy transferred would be insignificant. The
aurora, the Northern Lights and Southern Lights, occur here.
Some scientists consider that another layer above the thermosphere exists. It is called
the exosphere. Although considered by many as part of space, its action became
discernible by its drag effect on satellites. For example, even though the altitude of
the International Space Station is about 330 km, it loses about 2 km in altitude each
month due to drag, and must periodically be given a boost by rocket engines to keep
it in orbit.
34 Chapter 2. THE ATMOSPHERE

The exosphere gradually fades into outer space, so there is no clear upper boundary
of this layer. Some consider that it reaches up to 190,000 km, about halfway to the
Moon.

2.2 Concentration units of chemical species

To describe the chemical composition of the atmosphere shown in the right hand side of
Fig. 2.2, it is necessary to define the units used to express the amounts of chemical species.
Useful for this task is the ideal gas law,

P [milibars] V [m3 ] = n[mol] R[m3 milibars mol−1 K−1 ] T [K], (2.1)

where P is pressure, n number in moles of molecules, V volume and T is temperature.

As of 1986 the recommended value for the gas constant is R= 0.0831451 [m3 milibars
mol −1 K−1 ],[4], thus defining the corresponding SI units for n V, T and P . A mol is
the numerical value 6.02214076 × 1023 . By Avogadro’s Law, there is a mol of molecules
contained in a sample of a substance with same numerical value in grams of its corre-
sponding molecular mass. For atmospheric temperatures and pressures, the ideal gas law
gives an error of less than 0.2 percent for dry air and water vapor in comparison with an
expanded equation of state, [3].

Example 2.1
Example: How many molecules in grams of a substance

a The molecular mass of O2 is 32. If we take 32g of O2 , by Avogadro’s Law there

are 6.02214076 × 1023 , or one mol, of diatomic molecules of O2 .
b The molecular mas of nitrogen, N, is 14. How many molecules of N are contained
in 7, 14 and 28g of N?
7 14
Respectively, there are n = 14 × 6.02214076×1023 molecules , n = 14 ×6.02214076×
14
1023 molecules, and n = 28 × 6.02214076 × 1023 molecules of N.
c How many molecules are contained in 0.5, 0.1, 0.120g of nitrogen dioxide, NO2 ?

2.2.1 Mixing ratios

The most common unit to express the abundance of gases in the atmosphere is mixing
ratio by volume. It is defined as the fraction of volume of air occupied by a gas in an air
 2.2 CONCENTRATION UNITS OF CHEMICAL SPECIES 35

mixture. Mixing ratio by volume can be expressed as percent by volume, parts per million
by volume (ppmv), parts per billion by volume (ppbv) or even parts per trillion by
volume (pptv). For simplicity ppm, ppb and ppt will be used. It should be remembered
that a billion means 109 and that almost always mixing ratios are referenced to dry air.
Therefore, the mixing ratio ξi of a constituent gas ci in a given volume of air ctotal is given
by
ci
ξi =
, (2.2)
ctotal
where units in the numerator and denominator are of volume. When multiplied by a
hundred, a percentage is obtained. Occasionally, the mass ratio of ci and ctotal is used in
which case we denote this mixing ratio as parts per million by mass (ppmm), parts per
billion by mass (ppbm) or parts per trillion by mass (pptm). See [1]
From the ideal gas law given in Eq. 2.1, the moles per volume of an air sample con-
tained in volume V at pressure P is given by
n P
= . (2.3)
V RT
According to Dalton Law the partial pressure of a constituent gas i contained in the
sample is pi . Then,
ni pi
= . (2.4)
V RT
These equations into Eq. 2.2 give
ni /V pi /RT
ξi = = , (2.5)
n/V P/RT
or,
ni pi
ξi == . (2.6)
n P
A mixing ratio equivalently expressed as a molar ratio, mol mol−1 , and partial pressure
ratios pi /P has been obtained.
Therefore, the partial pressure of a specific gas in a mixture can be considered syn-
onymous with concentration in parts per million or billion. In this way a partial pressure
of 0.001 atm of a gas in air is equivalent to 1000 ppm, since 1000 × 10−6 = 0.001. By the
same token, a ppm is synonymous with a micro mol per mol, or µmol mol−1 , and a ppb
is synonymous with a nanomol per mol, or nmol mol−1 .
Mixing ratios are used to express the concentrations at various altitudes in the atmo-
sphere. Remembering again the ideal gas law, as total air pressure and hence total air
concentrations decreases with height, a constant mixing ratio does not mean a constant
concentration in terms of moles (or molecules) per volume or mole ratio. Also, molar
ratios are considered SI units, whereas ppm, ppb or ppt are not.
36 Chapter 2. THE ATMOSPHERE

Example 2.2
Example: How many molecules in a sample of air

a Obtain the concentration of air molecules in molecules per cm−3 in a sample at

T = 298 K and P = 1 atm.
Using Eq. 2.4,

1.01325 × 105 [N/m2 ]

c= = 40.897[mol/m3 ] , (2.7)
8.314[Nm/mol K] × 298[K]
or,
" #
mol m3 molecules
  
c = 40.897 × 6.02214076 × 1023
m3 106 cm3 mol
molecules
 
= 2.463 × 1019 . (2.8)
cm3
b Mexico City’s height is 2200 m ASL. Therefore, its average atmospheric pressure
is about 80% of the sea level pressure. If T=298 K, obtain the number of air
molecules per cm−3 .

0.8 × 1.01325 × 105 [N/m2 ]

c= = 32.72[mol/m3 ] , (2.9)
8.314[Nm/mol K] × 298[K]
or,
" #
mol m3 molecules
  
c = 32.72 × 6.02214076 × 1023
m3 6
10 cm 3 mol
molecules
 
= 1.97 × 1019 . (2.10)
cm3

Therefore, there are 20% less air molecules per cm3 in Mexico City
c Obtain number of molecules per cm3 in La Paz , Bolivia, when the temperature
is 10 ◦ C and 24.85 ◦ C.

2.2.2 Mass per volume

A common way to describe abundance of condensed materials and particles, or aerosols,
in the atmosphere is in terms of mass per unit volume of air, i.e., kgm−3 , gm−3 or µgm−3 .
These units are equivalent with mixing ratio by volume and can also be used used for
gases.
 2.3 CONSTANT CHEMICAL CONSTITUENTS OF THE ATMOSPHERE 37

Equation 2.7 provided the number of mols per cubic meter when a sample at T = 298
K and P = 1 atm. These are considered standard conditions. If this equation is multiplied
by the molecular weight, M, of the air sample in g mol−1 , the product will yield the desired
units in g m−3 . Therefore, for standard conditions, X ppm or ppb, yield Y µgm−3 via,
Y [µgm−3 ] = X [ppm] × 40.9 M[g mol−1 ] (2.11)
= X [ppb] × 0.0409 M[g mol−1 ] (2.12)

where M is the molecular weight of the compound or gas.

Example 2.3
Example: From mixing ratios to mass per volume and vice versa

a Obtain in µgm−3 a concentration of 10 ppb of ozone, O3 , in standard conditions.

There are 10 × 0.0409 × 3 × 16[µgm−3 ] = 19.6µg m−3 in 10 ppb of O3 .
b Obtain in µgm−3 a concentration of 75 ppb of Sulfur Dioxide, SO2 , in standard
conditions.
There are 75 × 0.0409 × (2 × 16 + 32)[µgm−3 ] = 0.0409 × 64µgm−3 = 26.18 µgm−3
in 75 ppb of Sulfur Dioxide.
c Obtain the concentration in µgm−3 of 75 ppb of Sulfur Dioxide in Denver, Co,
when T = 24.85◦ C and T = −10◦ C.

You can now transform to ppm or ppb the recommended levels of gases at standard
and other environmental conditions. See Exercise 2.2 in this chapter.
2.3 Constant chemical constituents of the atmosphere
We are now ready to describe the atmosphere in terms of its chemical constituents. The at-
mosphere contains gases whose proportional volume has remained constant for hundreds
of centuries and others whose proportional volume constantly vary in time, sometimes
in a matter of minutes, hours, days or years. The constant gases and their percent by
volume are shown in Table 2.1.

Table 2.1. Main constant gases in the atmosphere. Adapted from [5], Chapter 1.

Constituent Percent by Volume

Nitrogen (N2 ) 78.084
Oxygen (02 ) 20.946
Argon (Ar) 0.934
Neon (Ne) 0.00182
Helium (He) 0.000524
Krypton (Kr) 0.000114
Hydrogen (H2 ) 0.0005
38 Chapter 2. THE ATMOSPHERE

We can see that Nitrogen and Oxygen conform about 99 percent of the atmosphere
composition. This is true up to a height of 100 kilometers. The constant gases play a
minor role in meteorology. Due to evolutionary processes, as long as they remain in
these proportions they will provide for human, vegetation and animal life as we know it.
Although we have used the word “constant” for the gases in Table 2.1, their propor-
tional value was different millions of year ago. For example, 02 concentration variations
are shown in Fig. 2.3. About 360 millions years ago, during the Carboniferous Era, 02
reached almost 30%. Some scientist sustain that this provided the environment for giant
insects to exist, such as the Arthropleura shown in Fig. 2.4.

Figure 2.3. A schematic view of the changing concentration of atmospheric Oxygen

as a function of geological time in billions of years (Ga). Based on Tim Bertelink -
Own work. https://commons.wikimedia.org/w/index.php?curid=48915156

Figure 2.4. The Arthropleura centipede that existed during the Carboniferous Era
about 360 millions year ago, whose size is compared to a human. Insects breed
through their skin, so 30% of oxygen provided conditions to sustain this form of life.
 2.3 CONSTANT CHEMICAL CONSTITUENTS OF THE ATMOSPHERE 39

2.3.1 Vertical chemical structure of the atmosphere

In the absence of sources, sinks or turbulent macroscale mixing, the distribution of con-
stant gases, such as the chemically stable 02 , N2 and inert gases, is dominated by molecu-
lar diffusion due to random molecular motions. This mechanism is dominant starting at
a height of 100 km and going up, that is, above the turbopause or homopause, as shown
in Fig. 2.2. Since atmosphere above the turbopause is not well mixed, i.e., an heteroge-
neous mix, it is called heterosphere. Molecular diffusion tends to produce an atmosphere
in which the mean molecular mass of the mixture of gases decreases with height. This
condition is described by the Equation 2.13,
−z
c(z) = c(0)e H , (2.13)

where c(0) is the concentration at height z = 0, and H is the scale height for a given gas
defined as,
RT Nm/mol × K
 
H= , (2.14)
Mg mol kg−1 × ms−2
where R is the universal constant of gases, T is temperature, M is the molecular weight
and g is the gravity acceleration. Since H is inversely proportional to the molecular mass
of gas M, the concentration of gases with lower molecular mass (lighter) decreases slower
with height. It can be shown that that the concentration of a gas decreases by a factor e
for each increase in H in the height z. A depiction of changing molecular weight of dry
air Md with height is shown at center of Fig. 2.2.
40 Chapter 2. THE ATMOSPHERE

Example 2.4
Example: Vertical placement of gases in the atmosphere

a Obtain the scale height H of Hydrogen, H, and atomic Oxygen, 0.

Using Eq. 2.14 for H,

8.314[Nm/mol × 1000[K]
H= = 848000 m. (2.15)
1 × 10−3 [mol kg−1 ] × 9.8[ms−2 ]

For O,
8.314[Nm/mol × 1000[K]
H= = 530000 m. (2.16)
16 × 10−3 [mol kg−1 ] × 9.8[ms−2 ]
Note that the molecular weights are expressed in the MKS units system since R is
in those units. Temperatures were selected based on vertical thermal description
of atmosphere: T = 300K for O, and T = 1000K for H. These findings explain
why H, Helium,He, and monoatomic O dominate the higher altitudes in Fig. 2.2.
b Estimate the concentration of diatomic Oxygen, O2 , at 2,200 m. Use T = 300 K.
Assume c(0) = 21% at sea level.
Using Eq. 2.13,

8.314[Nm/mol × 300[K]
H= = 7900 m.
32 × 10−3 [mol kg−1 ] × 9.8[ms−2 ]

Therefore,
−2200
c(2200) = 21 × e 7900 = 21 × e−0.27 = 16 (2.17)

At 2,200 m the concentration of O2 is 16%, when at sea level is 21%. This is a

good estimation, but local variability in pressure and temperature also come into
play.
c Obtain the scale height H for Earth’s atmosphere in the lower 100 km, i.e., a gas
mixture of 21% O2 and 78% of NO2 . Relate these findings to Fig. 2.2.
 2.4 VARYING CHEMICAL CONSTITUENTS OF THE ATMOSPHERE 41

2.4 Varying chemical constituents of the atmosphere

Whereas in the turbopause the atmospheric vertical transport mechanism is random

molecular diffusion, below the turbopause the dominant mechanism that sets the vertical
motion of air parcels is the macroscale turbulent diffusion. This phenomenon, generated
by wind blowing over Earth’s surface and convection produced by heated land coverings,
overrides the molecular diffusion mechanism and produces a well mixed atmosphere, i.e.,
an homogeneous mix. That is why below the turbopause, the atmosphere is known as
the homosphere. The topic of atmospheric turbulence will be the subject of Chapter 7.
The reason why turbulent diffusion mechanisms dominate in the homosphere is be-
cause due to gravity, the atmospheric molecular arrangement is more compact here, re-
stricting the molecular mean free path. The time that it would take a molecule to travel
from high temperature to regions of lower temperatures would be many orders of mag-
nitude longer than the time required for macroscopic turbulent fluid motions.
Gases important for atmospheric chemistry and health effects are present in the homo-
sphere in small and varying concentrations. That is why they are called trace constituents.
Most of them have average concentrations below 200 ppb: So sensitive is the chemistry of
the atmosphere and its interaction with life and health. A partial list is given in Table 2.2.

Table 2.2. Volume fraction of some atmospheric trace gases at 1 atm. Adapted from [5].

Gas Chemical Formula Fraction of Volume of Air by the Species

Carbon Dioxide CO2 409.95 ppmv (Aug, 2019)
Methane CH4 1.8 ppmv
Hydrogen H2 0.56 ppmv
Nitrous Oxide N2 O 0.33 ppmv
Carbon Monoxide CO 40 – 200 ppbv
Ozone O3 10 – 200 ppbv (troposphere)
Formaldehyde HCHO 0.1 – 10 ppbv
Nitrogen Species NOx 10 pptv – 1 ppmv
Ammonia NH3 10 pptv - 1 ppbv
Sulfur Dioxide SO2 10 pptv – 1 ppbv
Dimethyl sulfide (CH3 )2 S several pptv – several ppbv
42 Chapter 2. THE ATMOSPHERE

Important reactive species are the regulated gases such as Ozone, O3 , Sulfur Dioxide,
SO2 , carbon monoxide, CO, Nitrogen Species, NOx , which are the sum of Nitrogen dioxide
and Nitrogen monoxide. Other compounds are toxics such as formaldehyde, HCHO, and
ammonia, NH3 .
Gases whose proportion is growing with time, mainly because of human activity, are
Carbon Dioxide, (CO2 ), and Methane (CH4 ) with 409.95 ppm and 1.8 ppm respectively.
Methane, is also a potent Greenhouse Gas. These gases, together with Nitrous Oxide,
N2 O, or laughing gas, are examples of important Greenhouse Gases whose concentrations
are responsible for the accelerated planet climate change, [9]. More on this subject will
be presented in Chapter 4, Section 4.5. As shown in Fig. 2.5, the concentrations of CO2
are growing at about 1.5 ppmv each year.
A gas not shown in Table 2.2 because its proportion strongly varies with time and
space is water vapor. Its mixing ratio may reach up to 4 percent by volume. It is the
source for clouds and precipitation. Its presence modifies the behavior of meteorological
phenomena.

2.4.1 Residence time, sinks and major sources of trace gases

If we were to follow on average the life history of the molecules of the trace gases shown
in Table 2.3 from their inception into the atmosphere until their removal, we would obtain
their residence time.

Table 2.3. Trace gases residence time and sources. Adapted from [5].

Gas Residence Time or Lifetime Major Sources

CO2 3 – 4 years Biological, oceanic, combustion, anthropogenic
CH4 10 years Biological,anthropogenic
H2 2 years Biological, HCHO photolysis
N2 O 150 years Biological, anthropogenic
CO 60 days Photochemical, combustion, anthropogenic
Ozone Days - Months Photochemical
HCHO 1.5 hours Photochemical
NOx variable Soils, anthropogenic, lightning
NH3 2 – 10 days Biological
SO2 Days Photochemical, volcanic, anthropogenic
(CH3 )2 S Days Biological, ocean
 2.4 VARYING CHEMICAL CONSTITUENTS OF THE ATMOSPHERE 43

Figure 2.5. Monthly mean atmospheric carbon dioxide at Mauna Loa Observatory,
Hawaii. The carbon dioxide data (red curve), measured as the mole fraction in dry
air, on Mauna Loa, constitute the longest record of direct measurements of CO2 in
the atmosphere. The black curve represents the seasonally corrected data in ppm.
Credit to C. David Keeling of the Scripps Institution of Oceanography.

A mathematical expression for the residence time is obtained by performing mass

balance analysis on a control volume placed in the atmosphere. If we let Q denote the
total mass of the substance in a given volume in the atmosphere, then the mass flux
through the volume boundary is given
dQ
= (Fin − Fout ) + (S − R) , (2.18)
dt
where Fin and Fout are the mass flux in and out of the volume, S is the source rate of
the substance inside the control volume, and R is the corresponding removal rate inside
the volume. If we consider an isolated volume, no mass escapes or is admitted, and then
Fin = Fout . Therefore,
dQ(t)
= S(t) − R(t) = P (t) . (2.19)
dt
If P (t) remains constant we have steady state case, so
dQ(t)
=P , (2.20)
dt
44 Chapter 2. THE ATMOSPHERE

which implies that Q increases or decreases at a constant rate (steadily), depending on

the sign of the constant P . Integrating this expression by separation of variables,
Z Q Z τ
dQ0 = P dt , (2.21)
0 0

so the residence time τ is then,

Q
τ= . (2.22)
P
If a compound concentration is constant in the atmosphere, it means that the emission
by sources S, is equal to removal processes R. Therefore P = R = −S, i.e., the compound
must be removed at the same rate that it is produced in the atmosphere.

Example 2.5
Example: Obtaining residence time of some gas compounds.

a The mass of the troposphere is about 4 × 1021 g. The constant mixing ratio in
mass of sulfur containing compounds is about 1 ppbm. Then the mass of sulfur
compounds in the troposphere is Q = 4 × 1012 g. Since sulfur compounds source
rate is S ≈ 200 × 1012 g yr−1 , the residence time is about,

4 × 1012 [g]
τ= = 1 week. (2.23)
200 × 1012 [g yr−1 ]

b Suppose that the removal rate of a compound is proportional to its concentration,

i.e., the more is present, the faster its removal. This is called first-order loss.
Obtain its residence time.
Assuming first-order loss R = kQ where k is a proportionality constant, then

Q 1
τ= = . (2.24)
kQ k

This may be the case for dry or cloud scavenging.

c Obtain τ when mass Q of a compound is removed by dry deposition, (k1 ), and
cloud scavenging,(k2 ).
 2.4 VARYING CHEMICAL CONSTITUENTS OF THE ATMOSPHERE 45

Table 2.4 shows the main sinks of the trace gases being considered. Although the
hydroxyl radical OH does not react with constant gases N2 , O2 , nor with CO2 , or H2 O,
it appears in 5 entries reacting with trace gases forming peroxy radical compounds. This
shows why OH is considered the most reactive species in the troposphere. That is why
OH plays an important role in eliminating some greenhouse gases such as methane and
other trace gases, thus deserving our attention.

Table 2.4. Major sinks of trace gases. Adapted from [5].

Gas Major Sinks

CO2 Photosynthesis
CH4 OH, microbial oxidation in soils
H2 soil uptake
N2 O O(1 D) in stratosphere
CO OH
O3 Photolysis
HCHO OH, photolysis
NOx OH
NH3 gas-to-particle conversion
SO2 OH, water-based oxidation
(CH3 )2 S OH

The hydroxyl radical OH is present in the troposphere in relative high concentrations,

although its lifetime is of about 1 s. At 1 atm and 298 K its concentrations range from
1 − 10 × 106 molecules per cm3 (or 0.04 to 0.4 × 103 ppb) during daytime and less than
2 × 105 molecules per cm3 (or 0.01 × 103 ppb) during nighttime. The most important
production routes of OH are via the following processes:
1. At wavelengths of < 319 nm, O3 photodissociates to form OH radicals,

O3 + hν → O2 + O (2.25)
O + H2 O → 2OH . (2.26)

2. Nitrous acid, HONO, is photolyzed at wavelengths of abouts 400 nm during daytime

to form OH radicals,

HONO + hν → OH + NO . (2.27)
46 Chapter 2. THE ATMOSPHERE

3. Nitrous acid is formed during nighttime thus maintaining a cycle via,

OH + NO → HONO + M . (2.28)

Thanks to OH, concentrations of CO are kept in check, and a removal mechanism of

CH4 , a greenhouse gas and other pollutants such as SO2 is in place.

2.5 The two facets of ozone

2.5.1 Ozone in the stratosphere: a shield against harmful solar

radiation
As shown in Fig. 2.2, in the troposphere of a standard atmosphere, temperature de-
creases with height 6.5 ◦ C per km, the environmental lapse rate. Nevertheless, after
the tropopause, temperature rises in the stratosphere. This is because of the presence of
Ozone in this layer.
In the higher part of the stratosphere the air is very thin. Most oxygen exists here
in its atomic form, having been dissociated from O2 molecules by solar UV radiation of
wavelength λ ∼ 200 nm (UV-C). Collisions among the atomic oxygen molecules leads
to re-formation of O2 which dissociate again photochemically as more UV-C light is ab-
sorbed.
Inside the stratosphere the intensity of UV-C light is lower since it was filtered above as
just described. Also, the air density is higher leading to higher concentrations of molecular
oxygen. Since the concentration of molecular oxygen is much higher than the atomic form,
chances are that the stratospheric oxygen atoms will collide with O2 molecules resulting
in the production of Ozone, as shown by the following exotermic reaction,

O + O2 −→ O3 + heat . (2.29)

To be more accurate, a third molecule is needed to carry away the heat energy generated
in the collision between atomic oxygen and its molecular form. This third molecule could
be N2 or H2 O or even another O2 molecule. This molecule will be designated generically
as M , so the above reaction becomes

O + O2 + M −→ O3 + M + heat . (2.30)

The release of heat by this reaction results in the increase of temperature in this layer,
thus forming a temperature inversion. Since vertical mixing is slow in the stratosphere,
there is a presence of a very stratified layer, ergo the name stratosphere.
 2.5 THE TWO FACETS OF OZONE 47

An ozone destruction process must be present in order to maintain its concentration

in check. The absorption of UV light in the range of 320 nm (UV-C and UV-B) results in
the decomposition of the O3 molecule in the middle and lower stratosphere:

O3 + UV(λ ≤ 320) −→ O∗2 + O∗ . (2.31)

The excited forms of oxygen produced by this reaction will react with O2 to reform ozone
but some excited forms will also react with ozone molecules, destroying them:

O3 + O∗ −→ 2O2 (2.32)

Since the activation energy for this reaction is relatively high for the atmosphere (17 kJ
mol−1 ) few of O3 and O collisions result in this reaction. If not disturbed, this cycle
explains the existence of the Ozone layer as depicted in Fig. 2.2.
The above chain of reactions comprise the Chapman mechanism.
There are other processes of ozone destruction not contemplated by the Chapman
mechanism. For example, the presence of Chlorine compounds act as catalysts in the
ozone destruction, leading to the ozone “holes” discovered in Antarctica. The chemical
mechanisms that lead to this man-made phenomena are described in Molina and Row-
land,[7].
To protect the ozone layer, the Montreal Protocol on Substances that Deplete the Ozone
Layer was signed in 1987. This is the first time that nations united to solve a global
environmental threat, [8].
The average lifetime of an ozone molecule in an altitude of 30 km is about 30 minutes,
but in the lower stratosphere it can last for months. Ozone concentrations in the strato-
sphere never exceeds 10 ppm. Nevertheless, these concentrations of ozone are sufficient
to filter UV-C and UV-B from sunlight before it reaches the surface layer, preventing dam-
age to life forms. Its formation took place about 400 million years ago, allowing living
organisms to migrate from the oceans and settle on land.

2.5.2 Ozone in the troposphere: harmful photochemical urban

pollution
Ozone in the troposphere is not formed as described by Chapman, because of the lack of
UV-C and UV-B already absorbed in the stratosphere. Urban ozone is the main pollutant
created by the Los Angeles type of smog. It is produced by the presence of anthropogenic
and biogenic emissions of Volatile Organic Compunds (VOC), NOx (= NO + NO2 ) and
sunlight. This phenomenon will be explained in a simplified form here.
48 Chapter 2. THE ATMOSPHERE

The NOx and VOC are the result of unburned and partially oxidized hydrocarbons.
VOCs are hydrocarbons and their derivatives whose boiling points are between 50◦ C and
260◦ C. These are emitted by unburned or partially oxidized fuel of internal combustion
engines or, for example, during vaporization while filling gas tanks. Also, as shown in
[6], highly reactive types of VOCs such as isoprene, pinene, and limonene emitted by
forest and shrubs surrounding a city, can enter the urban atmosphere enhancing ozone
production.
The emission of NOx is produced via two reactions. Some is produced by fuel NO.
About 30-60 % of fuel’s nitrogen is converted to NO during combustion, but since most
fuels do not contain much nitrogen, this process accounts for a small fraction of NO
emissions. The other form of NO emission is through thermal NO. At an activation
temperature above 2300 K, the following reaction takes place ,

N2 + O2 ⇐⇒ 2 NO (2.33)

The higher the temperature, the more NO is produced because of the high activation
energy needed for this reaction. These temperatures are encountered in combustion en-
gines of vehicular and thermoelelectric power plants. The inverse activation energy that
would restore N2 and O2 does not take place at temperatures present in the atmosphere.
The nitric oxide released into the atmosphere is oxidized to nitrogen dioxide, NO2 . This
reaction can take minutes to hours, depending on the concentration of other pollutants.
The yellow-brown color over the atmosphere of smog laden cities is due in part to the
presence of NO2 . This is because NO2 absorbs visible light in the range of 400 nm wave
length, filtering the purple component while allowing the transmission of yellow light.
As in the stratosphere, ground level ozone is produced by the reaction of oxygen atoms
with diatomic oxygen. However, the main source of oxygen atoms in the troposphere
come from the photochemical dissociation by UV-A sunlight of oxygen molecules:

O + O2 → O3 . (2.34)

Since most of UV-B and UV-B has been absorbed in the stratosphere, it is in the tropo-
sphere that UV-A light (315-400 nm) dissociates nitrogen oxide molecules providing the
supply of O needed for the formation of ozone,

NO2 + UV(λ = 315, 400nm) → NO + O . (2.35)

As metioned,Nitric oxide, NO, is predominantly emitted from combustion engine vehicles

and thermal electric power plants. During episodes of photochemical air pollution NO
is oxidized to NO2 through a complex reaction processes that involve the presence of
 2.6 THE NONLINEARITY OF PHOTOCHEMISTRY 49

free radicals as catalysts. Examples of a peroxy radical generically expressed as RO2 , are
CH3 O2 , HO2 . For sake of simplicity consider ethene, C2 H4 , the simplest alkene, reacting
with an hidroxil, OH, as shown in Fig. 2.6, converting NO to NO2 ,

RO2 + NO → NO2 + H2 O (2.36)

where the peroxy radical RO2 is produced by

RH + OH → R + H2 O (2.37)
R + O2 + M → RO2 + M. (2.38)

Figure 2.6. Formation of NO2 and peroxy radicals from hydrocarbons

This process shows that ozone is formed in considerable quantities when urban air is
rich in NOx and alkene hydrocarbons whose double C=C bonds allow fast reaction with
OH radicals.
Summarizing, ozone is formed when atomic oxygen dissociated from nitrogen dioxide
reacts with O2 . This atomic oxygen is provided by the presence of alkene hydrocarbons
that are first converted into RO2 according to equation 2.38 and then to NO2 . As seen
again, OH plays an important role in atmospheric photo-chemistry.

2.6 The nonlinearity of photochemistry

The above discussion on photochemical processes is a simplified explanation of complex

and nonlinear photochemical mechanisms, an active area of research. Its complexity in-
cludes a distinctive phenomenon called nonlinearity. In the present context it turns out
that we cannot consider the atmosphere, UV radiation and VOCs, reacting with NOx as
50 Chapter 2. THE ATMOSPHERE

a simple reactor, where we observe that an increase in O3 is commensurable with the

addition of its precursors. As a matter of fact, it is possible that adding NOx could cause,
at least locally, a decrease in O3 , or that a decrease in emitted VOCs may trigger higher
O3 production.
This situation is explained using Fig. 2.7 of next example, showing a graph relating
measured concentrations of NOx , VOCs and corresponding ozone concentration contours.
This is a typical EKMA diagram.

Example 2.6
Example: Environmental policies to reduce O3 .

Figure 2.7. Typical EKMA diagram

showing ozone isopleths.

a Refer to Fig. 2.7. Suppose a city whose atmosphere is situated in point A, wants to
reduce its O3 concentrations. If only VOCs are reduced leaving same concentra-
tions of NOx , we get near to point C. Virtually no reductions of O3 concentrations
have been achieved. If instead NOx is reduced, a sizable reduction of O3 is ac-
complished. The right policy is NOx reductions.
b Refer to Fig. 2.7. Suppose a city whose atmosphere is situated in point B, wants
to reduce its O3 concentrations. If only NOx concentrations are reduced, leaving
the same concentrations of VOCs, we get near to point C. As we travel along
this trajectory, we arrive at higher O3 isopoleths, increasing their concentrations
instead of lowering them! The correct policy is to decrease VOCs.
c Refer to Fig. 2.7. Suppose a city whose atmosphere is situated in point C, wants
to reduce its O3 concentrations. What would be the right environmental policy?
 2.6 THE NONLINEARITY OF PHOTOCHEMISTRY 51

The above exmaple, show that Point A is located in the NOx-limited region, where
concentrations of O3 are sensitive to NOX . In this region there are high concentrations of
VOCs providing an abundance of peroxy radicals quickly oxidizing NO to NO2 . Point B
is in the VOC-limited region, where O3 concentrations are sensitive to VOCs.
In this region if NOx concentrations are lowered, more OH radicals will be available
to react with VOCs, enhancing production of NO2 . In this case, ozone concentrations are
expected to increase even though a precursor emissions were lowered. It should also be
noted that it is possible that inside a city, the sensitivity to VOCs or NOx may change
from one area to another, depending on local emissions.
The knowledge of the nonlinearity in the O3 production by sunlight, NOX and VOCs,
was critical to correct policies in the 1990s. Three-way catalytic converters in vehicles in-
stead of two-way catalytic converters were introduced, among other environmental poli-
cies. Since then, the atmospheric environment in many cities around the world has been
markedly improved, as presented in Chapter 1, Section 1.2.
52 Chapter 2. THE ATMOSPHERE

Exercises

2.1 Mixing ratios

10 ppm of a pollutant gas in air means:
a) X molecules of pollutant in x molecules of air
b) X moles of pollutant in x moles of air
c) X atm partial pressure of pollutants per x atm of air
(Answer on Page 55)

2.2 Units of pollutants in the World Health Organization recommendations.

Consult in https://www.who.int/news-room/feature-stories/detail/what-are-the-who-air-
quality-guidelines the WHO pollution recommendations for criteria pollutants.
a) Transform to ppm or ppb the recommended levels of O3 , at standard conditions.
b) Transform to ppm or ppb the recommended levels of O3 , at 20 ◦C at a height of
2200 m ASL.
c) Which units are preferable to use for health purposes?
(Answer on Page 55)

2.3 Partial pressures

Obtain the partial pressure of N2 O at sea level, knowing that its mixing ratio is 330 ppbv.
(Answer on Page 55)

2.4 Residence time

a) Ammonia, NH3 , has a residence time of about 5 days. It comprises about 1× 10−8
by mass of Earth’s troposphere. Assuming steady state conditions, estimate its
removal rate.
b) Nitrous oxide, N2 O, comprises about 3× 10−5 by mass of Earth’s troposphere.
Its removal rate is about 1× 1010 kg yr−1 . Assuming steady state conditions,
estimate its time of residence.
(Answer on Page 55)
 EXERCISES 53

2.5 Convert concentration for 10 µgm−3 to ppb and atm

Convert a concentration of 100 µgm−3 to ppb at 27 C and 1 atm.
(Answer on Page 55)

2.6 Obtain scale heights

Obtain scale height H for N2 , 02 , Ar. Based on this accommodate this compounds in their
vertical place in the atmosphere.
(Answer on Page 55)
54 Chapter 2. THE ATMOSPHERE

2.7 Loschmidt Number

Using the Universal Law of gases, obtain the number of molecules of air in 1 m3 at 1 atm
and 0◦ C. This is the Loschmidt Number.
(Answer on Page 55)

2.8 Number of CO2 molecules in air

Using the Loschmidt Number, obtain the number of molecules of air in 1 m3 at 1 atm and
0◦ C. Consider that mixing ratio of CO2 is 360 ppm.
(Answer on Page 55)
 EXERCISES 55

Answers to Exercises

Answer to Exercise 2.1 (page 52)

Answer to Exercise 2.1

Answer to Exercise 2.1 (page 52)

Answer to Exercise 2.1

Answer to Exercise 2.3 (page 52)

Answer to Exercise partial pressures

Answer to Exercise 2.4 (page 52)

Answer to Exercise 2.3

Answer to Exercise 2.5 (page 53)

Answer to Exercise 2.4

Answer to Exercise 2.6 (page 53)

Answer to Exercise 2.5.

Answer to Exercise 2.7 (page 54)

Answer to Exercise 2.7

Answer to Exercise 2.8 (page 54)

Answer to Exercise 2.8
56 Chapter Bibliography

Chapter Bibliography

[1] Peter V. Hobbs. Introduction to Atmospheric Chemistry. Cambridge University Press, The Pitt
Building, Trumptington Street, Cambridge, United Kingdom. First Edition, 2000.
[2] International Organization for Standardization (ISO) International Standard Atmosphere. Inter-
national standard, ISO 2533:1975.
[3] Pruppacher, H.R. and Klett, J.D. (1997) Microphysics of Clouds and Precipitation. Kluwer
Academic, Dordrecht, 954 p. Second Edition 1997.
[4] Cohen, E. Richard, Barry N. Taylor. The 1986 CODATA Recommended Values of the Fun-
damental Physical Constants Journal of Research of the National Bureau of Standards, 92(2),
March-April 1987
[5] John M. Wallace, Peter V. Hobbs Atmospheric Science an Introductory Survey. Academic Press,
Second Edition. 2006. ISBN 13:978-0-12-732951-2
[6] Aron D. Jazcilevich, Agustín R. García, Ernesto Caetano. Locally induced surface air
confluence by complex terrain and its effects on air pollution in the valley of Mex-
ico. Atmospheric Environment,Volume 39, Issue 30, 2005,Pages 5481-5489, ISSN 1352-2310.
https://doi.org/10.1016/j.atmosenv.2005.05
[7] Molina, M., Rowland, F. Stratospheric sink for chlorofluoromethanes: chlorine atom-catalysed
destruction of ozone. Nature 249, 810–812 (1974). https://doi.org/10.1038/249810a0
[8] United Nations Environment Programme Ozone Secretariat. The Montreal Protocol on Sub-
stances that Deplete the Ozone Layer. https://ozone.unep.org/treaties/montreal-protocol. Re-
trived, June 2012.
[9] World Meteorological Organization (WMO) State of the Global Climate 2021 (WMO-No. 1290).
ISBN 978-92-63-11290-3.
[10] John H. Seinfeld and Spyros N. Pandis. Atmospheric Chemistry and Physics. John Wiley and
Sons, Inc.,Wiley Interscience, 1998. ISBN 0-471-17815-2
[11] Barbara J. Finlayson-Pitts and James N. Pitts, Jr. Chemistry of the Upper and Lower Atmopshere.
Academic Press, 1999. ISBN 0-12-257060-x
[12] C.A. Riegel Fundamentals of Atmospheric Dynamics and Thermodynamics. World Scientific
Co.Pte. Ltd., 1992. ISBN 9971-978-86-5
[13] Daniel Valero Fundamentals of Air Pollution. Academic Press, Fourth Edition, 2008. ISBN
978-0-12-373615-4
Chapter 3

Interactions between solar

radiation and the
atmosphere

3.1 Electromagnetic radiation, quanta and photons 58

3.2 Radiometry concepts 63

3.3 Radiation balance and fundamental laws 71

3.4 The Sun’s electromagnetic field spectrum and its physical effects 76

3.5 Radiation Phenomena in the Atmosphere 79

3.6 Attenuation (extinction) of solar radiation by atmospheric gases 96

3.7 Attenuation (extinction) of solar radiation by atmospheric 99

particles
3.8 Epilogue 103

Exercises 103

Answers to Exercises 105

Chapter Bibliography 106

In Fig. 3.1 the Sun appears low in the horizon, as seen from the International Space
Station. The Sun is made of hydrogen in a plasma state whose nuclear fusion forms
helium, generating Sun’s energy. This chapter explains the physical principles describing
the interactions between solar radiation and the atmosphere including the case of polluted
air. The involved processes determine the Earth’s temperature, the thermal and chemical
– 57
structure of the atmosphere, and its optical –
appearance.
58 Chapter 3. INTERACTIONS BETWEEN SOLAR RADIATION AND THE ATMOSPHERE

Figure 3.1. The Sun and Earth seen from the International Space Station on 24 May 2020.

We begin with an explanation of Sun’s radiation as a wave and quantum phenom-

ena. This duality provides the scientific background explaining the radiation phenomena
taking place in the atmosphere, necessary to understand solar radiation’s relation with
atmospheric physics and chemistry. Then, basic terminology and most important radi-
ation concepts and laws of atmospheric radiation are introduced. Once this material is
presented, the Sun’s radiation effects on the atmosphere are explained. Radiative en-
ergy transfer is covered by stating the Beer-Lambert-Bouguer law. This law can predict
the energy arriving to Earth surface just by plugging in attenuation coefficients. Further
knowledge is gained though, when the responsible scattering and absorption phenomena
are described with some detail. Not only this approach prepares the terrain to integrate,
at a latter stage, material on atmospheric thermodynamics and physical chemistry, but
also conveys how humans optically perceive the atmosphere.

3.1 Electromagnetic radiation, quanta and photons

As Maxwell proposed in 1864, a ray of electromagnetic radiation consists of an oscillating

electric and magnetic field, [1]. As shown in Fig. 3.2, these fields are placed at 90 degrees,
and travel through empty space at 299,792,458 m s−1 (approximately 3.0 × 10 8 m s−1 ).
This is the speed of light denoted by c. This form of radiation transfers energy from
one point in space to another regardless whether the points are separated by physical
media or vacuum. For example, the warmth we feel or the light we perceive from the Sun
 3.1 ELECTROMAGNETIC RADIATION, QUANTA AND PHOTONS 59

are forms of electromagnetic radiation that traveled through empty space and a gaseous
atmosphere.
The height of the oscillations is named amplitude. Its units are volts, V, or Teslas,
T, depending on weather we refer to the electric or magnetic field, respectively. Since
the amplitudes of the electric and magnetic fields correspond in time, they are said to
be in phase. As shown in Fig. 3.2, the wavelength denoted by λ is the distance between
crests of the wave. Its units are usually expressed in micrometers, µm, nanometers, nm,
or meters, m. The number of cycles per second is called frequency, denoted by ν. Its
units are Hertz, Hz, and is defined as one cycle per second: 1 Hz = 1 s−1 .

Figure 3.2. An electromagnetic plane wave. The actual magnitude of the electric
field is much larger than the magnetic field, although they contain same amount of
energy.

Frequency and wavelength are reciprocal quantities, since the longer the wavelength
the more it takes for a complete wave to pass by, i.e., the longer the wavelength the lower
the frequency, and vice versa. The exact relationship is:

λν = c , (3.1)

where c is the speed of light in a vacuum. We will assume that this is also the speed of
light in the atmosphere.
60 Chapter 3. INTERACTIONS BETWEEN SOLAR RADIATION AND THE ATMOSPHERE

Example 3.1
Example: Wavelengths and frequencies

a Heat from the Sun is transmitted as an electromagnetic wave with a frequency

of about 3 × 1014 Hz. What is its wavelength?
From Eq. 3.1,

c 2.998 × 108 m s−1 2.998 × 108

λ= = = m = 0.998 × 10−6 m ,
ν 3 × 1014 s−1 3 × 1014
or about 1000 nm. As will be defined later, this range of frequencies are in the
infrared spectrum.
b Ultraviolet radiation is transmitted by the Sun. Its frequency is about 8.6 × 1014
Hz. What is its wavelength?
Answer: 350 nm.
c Visible light is transmitted by the Sun to the Earth. Corresponding range of
frequencies are between 4.3 × 1014 Hz and 7.1 × 1014 Hz. Therefore visible light
occupies a narrow range of frequencies. Obtain the corresponding wavelength
range.

The electric field E of the electromagnetic wave represented in Fig. 3.2, is described
mathematically by

x x
   
E = E0 cos ω t − = E0 cos 2πν t − , (3.2)
c c
or in exponential (complex) notation

x x
E = E0 eiω(t− c ) = E0 ei2πν(t− c ) , (3.3)
where E0 is the amplitude of the electric field in volts and ω = 2πν is angular frequency
in radians s−1 . Both equations state that the wave of amplitude E0 travels a distance
∆t
∆x in time c . The exponential notation becomes handy when the electromagnetic wave
interacts with a medium as will be shown in Section 3.7, where scattering and transmission
of solar radiation is discussed.
Although many experiments involving electromagnetic radiation can be explained
using Maxwell wave-like approach, there are other cases in which this explanation is
not possible. One of them is when a metal is illuminated by ultraviolet light. In this case
electrons are rejected only if the light frequency is above a threshold that is characteristic of
the metal. This is the photoelectric effect. For its explanation Albert Einstein was awarded
 3.1 ELECTROMAGNETIC RADIATION, QUANTA AND PHOTONS 61

the physics Nobel prize in 1921. The fact that a threshold dependent on a material must
be acquired to observe an effect, contradicts the wave-like behavior. Einstein proposed
that radiation is composed of particles or packets of energy that he called photons. The
energy or quanta of a single photon is related to the frequency ν of the radiation by

Ep = hp ν, (3.4)

where hp is Planck’s constant with value 6.626 × 10−34 [J s]. Equation 3.4 states that
frequency and energy are directly proportional with hp as the constant of proportionality.
Therefore, the larger the frequency, the stronger the energy. The subindex p is used to
stress photonic energy in which case a a commonly used unit of energy is the electron-
volt (eV) rather than the joule (J). An electron volt is the energy required to raise an electron
through 1 volt, thus a photon with an energy of 1 [eV] = 1.602 × 10−19 [J].

Example 3.2
Example: From wavelength to electron volt

a Consider a photon with wavelength of 470 nm, i.e., visible blue light. What is its
energy Ep ?. In Eq. 3.4, the energy carried by each photon is given by,
c
Ep = hp ν = hp ,
λ
Then,
Ep = (6.626 × 10−34 [Js]) × (6.4 × 1014 [Hz]) = 4.2 × 10−19 [J] .

In eV’s,
4.2 × 10−19 [J]
Ep = = 2.62[[eV]].
1.602 × 10−19 [J]
b Consider visible red light. Its wavelength is 700 nm. Obtain corresponding pho-
ton Ep .
Answer: 2.0 × 10−19 [J]= 1.25 [eV].
Note that the red photon has less energy and slower frequency than the blue
photon.
c Compare Ep photon energies, of infrared (λ= 1000 nm) and ultraviolet radiations
(λ= 350 nm).
62 Chapter 3. INTERACTIONS BETWEEN SOLAR RADIATION AND THE ATMOSPHERE

We can conclude that according to Einstein we can visualize a beam of red light as a
stream of photons where each photon is carrying an energy of Ep = hp × 4.310−14 [Hz] =
2.8 × 10−19 [J], whereas in a beam of green light each photon is carrying an energy of 3.8
× 10−19 [J].
Regarding the photoelectric effect, to observe the ejection of electrons, photons must
acquire the correct energy given by a work function given by,

Ek = hp ν − Φ , (3.5)
|{z} |{z}
energy supplied by photon energy required to eject electron

where Ek = 12 me v2 is the kinetic energy of an ejected electron. Here, me is the mass of an

electron and v its velocity. Basically, Eq. 3.5 states that if the energy of the photon, hp ν, is
greater than Φ, the electron is ejected with kinetic energy Ek .
The expression in Eq. 3.5 describes a plot where the kinetic energy Ek against the
frequency of radiation should be straight line of slope hp , the same for all metals, and
an extrapolated intercept with the vertical axis at −Φ, which is different foe each metal.
The intercept with horizontal axis (corresponding to zero kinetic energy of the ejected
Φ
electron) is at hp in each case. This is shown in Fig. 3.3. Once the proper energy for a
specific material is achieved, the energy of ejected electrons varies linearly with incident
photon frequency

Figure 3.3. Required work function frequency and corresponding kinetic energy of
ejected photons for Rubidium, Potassium and Sodium.
 3.2 RADIOMETRY CONCEPTS 63

The electromagnetic radiation wave-particle duality is necessary to explain the inter-

action of solar radiation with atmospheric constituents. In some cases, solar radiation
will behave like a wave when producing heat on a body surface by moving back and
forth its molecules. In other cases, it will behave like a stream of photons colliding with
gases atoms in the atmosphere producing photochemical reactions, and energy release or
absorption by dislocating electron orbitals.

3.2 Radiometry concepts

The following are standard concepts to describe radiative energy transfer. We take into
account definitions provided by the AMS Glossary of Meteorology.
Radiant energy, qe , is the energy transported towards or away from a surface by
millions of photons. Its units are Joules, [J], or Watts per second, [W s−1 ].
Radiant flux, Φe , is the amount of emitted or received radiant energy per unit time.
Its units are Joules per second, [J s−1 ], or Watts.
Actinic flux, F . In photochemistry, where the energy interchanges between photons
and molecules is necessary to describe, we need to know the photon flux incident
on a molecule form all directions. Actinic flux is defined as the radiative flux from all
directions on a volume. The word actinic, means ”capable of causing photochemical
reactions”, [11]. When we need to specify the wavelength of actinic flux, it will be
denoted by F (λ). This will be then a spectral actinic flux. The units used for F are
number of photons per [cm−2 s−1 ], and for F (λ),[cm−2 s−1 nm−1 ] .
Irradiance or radiant flux density, Ee , is the radiant flux passing normal to a unit area
per time,
Φe
Ee = , (3.6)
A
where A is the normal or effective area encountered by Φe . The resulting units are
[W m−2 ] or equivalently, [J m−2 s−1 ]. Since the definition of irradiance Ee includes a
vector normal to a surface, it should be treated as a vector.
Solid angle and steradians.The concept of solid angle appears prominently in radiom-
etry and optics. Consider the following analogy: In two dimensions, the magnitude
in radians, rad, of an angle θ is the length S subtended by θ on circle divided by the
radius, r, of the circle. This is shown in Fig. 3.4(a). In three dimensions, a solid angle
Ω is the surface area subtended by Ω over the square of the radius, r2 of a sphere:

A
Ω= . (3.7)
r2
64 Chapter 3. INTERACTIONS BETWEEN SOLAR RADIATION AND THE ATMOSPHERE

This is shown in Fig. 3.4(b). Angles are measured in radians, [rad], while solid angles
in steradians, [sr]. In a full circle of radius r the angle has 2 Sr =2 πr
r = 2π radians,
2
while in a full sphere of radius r, the solid angle has 4 rA2 =4 πr
r2
= 4π steradians.

Figure 3.4. In (a) angle θ subtending S in a circle of radius r. The magnitude in

radians of the angle θ is the corresponding arc length over radius r . In (b) solid
angle subtending an area on surface of sphere with radius r2 . The magnitude of Ω
in steradians is the projected area A over the square of radius, r2 .

Example 3.3
Solid angles and steradians

a Obtain the case when a solid angle subtends one steradian.

A A
Since Ω = r2
, then 1sr = r2
and A = r2 .
b Obtain the area subtended by one steradian of the surface of a sphere.
1
From its center a sphere subtends 4π steradians. Then, one steradian is 4π = 0.08,
or 8% of the sphere area.
c The Moon has a radius of 1737 km, and the distance between the Earth and the
Moon is 384,400 km. Obtain the solid angle subtended by the Moon in the night
sky.
Answer: Solid angle subtended by Moon is 0.000064 sr.

Radiance or radiant intensity, Ie is the rate at which radiant energy in a set of direc-
tions confined to a unit solid angle around a particular direction is transferred across
 3.2 RADIOMETRY CONCEPTS 65

unit area of a surface (real or imaginary) projected onto this direction. Its units are
[J m−2 sr−1 ]. If the radiance is expressed as a function of wavelength λ, i.e., Ie (λ, the
units are [J m−2 nm−1 sr−1 ]. Then,

dΦe
Ie = . (3.8)
dA dΩ cos θ dt
The chromatic or spectral version is,

dΦe
Ie,λ = .
dA d λ dΩ cos θ dt
Its units are [J m−2 nm−1 sr−1 ].
The radiance or intensity Ie and irradiance Ee are related by

dEe
Ie = , (3.9)
cos θ dΩ
where dΩ represents the differential of solid angle and θ the angle between the beam
of radiation and the direction normal to the surface (usually horizontal) on which
the radiance is measured, as shown in Fig. 3.5. In other words, the angle θ is used to
obtain the normal or effective radiation area.
Relationships between irradiance and radiance. Consider a receptor or emitter lying
in space according to the geometry of Fig. 3.6. Relative to the receptor, θ is the zenith
and φ the azimuth angle. Therefore, by integrating on all the zenith range,
Z
Ee = Ie cos θ dΩ. (3.10)
2π

For isotropic or Lambertian radiance, i.e., when radiation is independent of direction,

it can be shown that dΩ = 2π sin θdθ . Then,
π
Z Z 2π Z
2
Ee = Ie cos θ dΩ = Ie cos θ sin θdθdφ
2π φ=0 θ=0

π
1 h 2 i π2
Z
2
= 2πIe cos θ sin θ dθdφ = 2πIe cos θ .
θ=0 2 0

By evaluating the limits in last expression, Eq. 3.9 becomes

Ee = πIe , (3.11)

or in spectral form
Ee,λ = πIe,λ . (3.12)
66 Chapter 3. INTERACTIONS BETWEEN SOLAR RADIATION AND THE ATMOSPHERE

Example 3.4
Relationship between irradiance (E) and radiance (I)
A surface A1 = 3 cm2 emits 750 Wm−2 in the infrared, at λ = 400 nm. This radi-
ation strikes surface A2 = 5 cm2 , at a distance r=75 cm, as shown in Fig. 3.5(a).
Assume Lambertian case.
a Using geometry shown in Fig. 3.5 (a)determine solid angle subtended by A2
when viewed form A1 .
The corresponding solid angle Ω is,

A2 5
Ω= = 2 = 8.88 × 10−4 sr .
r2 75
b Obtain radiant flux (or radiation rate) Φe,λ at which radiation emitted by A1
strikes A2 .
First, we obtain the radiance or radiance intensity Ie,λ emitted by A1 ,

750
Ie,λ = = 238.7 [Wm−2 m−2 sr−1 ] .
π
Since both surfaces are parallel, θ = 0 in Eq. 3.9 ,

Φe,λ = 238.7 × (3 × 10−4 × 8.888 × 10−4 ) = 0.6365 × 10−4 W.

c Obtain radiant flux at which radiation emitted by A1 strikes tilted surface A2 , as

shown in Fig. 3.5(b), where θ2 =40 ◦ . Interpret results.

Figure 3.5. In (a), surface A1 =3 cm2 emits 750 Wm−2 at λ = 400 nm.This emission
is intercepted by surface A2 =5 cm2 . In (b) surface A2 is tilted by an angle of θ2 = 40◦ .
 3.2 RADIOMETRY CONCEPTS 67

Figure 3.6. Consider a receptor or emitter lying in the center of an hemisphere.

Relative to the receptor, θ is the zenith and φ the azimuth angle. Radiation is passing
through a surface element δφ by δθ of the hemisphere.

The relations in Eqs. 3.11 and 3.12 should be understood in the following way: Con-
sider an area element dA emitting or receiving an irradiance Ee . This is an emission in all
directions above the area element. If the emissions is same in all directions (isotropic),
the source radiance is obtained by dividing the irradiance by the solid angle subtended
by the hemisphere above the source, i.e., dΩ = 2π sr, and by projecting the outgoing
beams flux to the normal vector of dA via the cos θ term, thus obtaining Eq. 3.11 via
Eq. 3.9. Therefore, Eq. 3.11 refers only to an isotropic or Lambertian receptor or emitter
area element.
An important feature of irradiance Ee , as opposed to radiance Ie , is that Ee obeys the
inverse square law, i.e, as the relative distance r between source and receptor increases,
the value of Ee decreases proportionally to r2 , and vice versa. Consider an isotropic radiant
flux source with radiant energy Φe at the center of a sphere of radius r. The irradiance
Φe 2
at the sphere surface will be Ee = 4πr 2 , since 4πr is the sphere surface area. If the sphere

radius is increased, the corresponding surface area is increased proportionally to r2 , and

therefore Ee decreases in this same proportion. Therefore, as a receptor moves away from
the source, Ee decreases as r2 , and vice-versa: if a receptor gets closer to the source, Ee
increases as r2 . This is because the total power is constant and it is spread over an area
that increases with the distance squared from the radiation source.
68 Chapter 3. INTERACTIONS BETWEEN SOLAR RADIATION AND THE ATMOSPHERE

On the other hand, consider the radiance Ie with a radiant surface element dA. In
1
Eq. 3.8, the quadratic terms in the numerator and denominator cancel, since Φ ∝ r2
and
dA
dΩ = r2
. Therefore, irradiance Ee incorporates a quadratic dimming effect with distance,
while radiance Ie is constant. Under this conditions,
Ie
Ee ∝ . (3.13)
r2
In other words, irradiance decreases proportional to r−2 as the area element moves farther
away from the source, while the amount of radiance is constant. If a point radiation source
emits radiation uniformly in all directions and there is no absorption, then the irradiance
drops off in proportion to the distance squared from the source, since the total power is
constant and it is spread over an area that increases with the distance squared from the
radiation source.
Important considerations in the choice of independent variable for irradiance. The
irradiance energy at a specific frequency ν (or wavelength λ = νc ) is denoted by Ee,ν
(or Ee,λ ), and is called monochromatic or spectral irradiance. By integrating Ee,ν
over all frequencies or wavelengths E is obtained,
Z ∞
Ee = Ee,ν dν, (3.14)
0
or Z ∞
Ee = Ee,λ dλ. (3.15)
0
By using Eqs. 3.1 and 3.14 it follows that,

c
 
Ee,λ = Ee,ν . (3.16)
λ2
The fact that a square value of λ appears en Eq. 3.16 means that the transforma-
tion of Ee from wavelength to frequency units is not linear, and “distortion” of abcisa
axis (independent variable) takes place. Thanks to the form of Eq. 3.16, Ee is obtained
indistinctly using Eqs. 3.14 or 3.15 even though the function shapes in the correspond-
ing integrands are different. Care most be exercised when interpreting energy density
graphs, as explained by Soffer and Lynch in [5]. See next example.
 3.2 RADIOMETRY CONCEPTS 69

Example 3.5
Example: Value of Ee in terms of λ and ν and resulting paradoxes.

1. As will be shown in Section 3.4 and 3.6, solar irradiation at sea level peaks to
approximately 1.8 W m−2 nm−1 at λ = 500 nm. What is the corresponding
irradiation in W m−2 nm−1 Hz−1 ?
From Eq. 3.16

λ2 (0.50 × 10−6 )2 m2
Ee,ν = Ee,λ = (1.8 × 109 Wm−2 m−1 s).
c 300 × 106 m s−1
Then,

Ee,ν = (0.833 × 10−12 )(1.8 × 109 Wm−2 m−2 Hz−1 ) = 1.49 × 10−12 Wm−2 Hz−1 .

The corresponding frequency for this value is,

c
ν= = (300 × 106 ms−1 ) (0.5 × 10−6 m) = 6.0 × 1014 Hz.
λ
Therefore,
Ee,ν=6.0×1014 Hz = 1.49 × 10−12 Wm−2 nm−1 Hz−1 .

2. Human eye is most sensitive to light around λ= 560 nm. From above, peak
solar radiation is at λ = 500 nm. Apparently, peak solar radiation and peak
eye sensitivity are close when compared using λ. Compare now when using
corresponding frequencies.
Answer: When expressed in Hz, form above example solar radiation peaks around
6.0 ×1014 Hz. Peak of eyes sensitivity is around 3.41 ×1014 Hz. When using fre-
quencies, eye sensitivity and solar peak radiation do not coincide.
3. Explain the paradox found above.
Hint: Compare Eqs. 3.16 and 3.1.
Comment: From above example, it should be clear that the peak position in the
ordinates for Ee , depends on choice of independent variable. Also, note that the
relative magnitude of the abscissas, when the ordinate is expressed in Hz or nm,
varies by many orders of magnitude.
70 Chapter 3. INTERACTIONS BETWEEN SOLAR RADIATION AND THE ATMOSPHERE

Example 3.6
Example: Devices that use radiant flux, radiance and irradiance

To further understand radiometry concepts, we examine how some electronic detec-

tors are used to measure, or extract energy from solar radiation.
1. A photodidode is an electronic device that converts radiant flux into electric
current. The current is generated when photons, coming form all directions and
surrounding sources, are absorbed in the photodiode. A crucial parameters of
a photodiode is its spectral responsivity, r, expressed as the ratio of generated
electric current, Ic in Amperes (A), to incoming radiant flux, Φe , in Watts, (W):

Ic
r= .
Φe

Consider a silicon photodiode, for which r = 0.3 W

A at λ =500 nm. Obtain the
generated current if corresponding Φe = 9 W.

Ic = 0.3 × 9 = 2.7A.

2. Consider an array of thousands of similar devices called photovoltaic cells. They

are connected in such a way, that their current is summed. The power conversion
efficiency of a solar cell is a parameter which is defined by the fraction of incident
power converted into electricity. Consider a solar panel of 1 m2 with 20% of power
efficiency, operating at sea level. Obtain generated current at sea level, if the
operating wavelength of the solar panel is in the nearinfrared. The surface of the
panel, makes a 70 degree angle with incoming radiation.
At sea level, the irradiance in the near infrared is about 1000 W m−2 . Since the
angle θ is with respect the normal to the surface, the effective area is 1m2 ×
cos(90 − 70) =0.94 m2 . The effective available incoming power is then 0.94 m2 ×
1000 W m−2 =940 W. The solar panel generates 940 W × 0.2= 188 W.
Notice that the irradiance incorporates parameters such as area, or energy density,
and orientation.
3. Suppose a surface A1 = 1[m2 ] emits a radiance Ie [J m−2 sr−1 ] towards another
surface A2 = 10[m2 ], 10 [m] away. Obtain best orientation of A2 to maximize
received power.
 3.3 RADIATION BALANCE AND FUNDAMENTAL LAWS 71

Figure 3.7. Incident radiation flux decomposed while passing through a semitrans-
parent material into absorbed, reflected and transmitted portions. Notice that ab-
sorbed radiation stays within the control volume, while transmitted radiation goes
through the medium. Reflected radiation doesn’t enter the control volume.

3.3 Radiation balance and fundamental laws

When radiation flux Φe strikes a body surface, part is reflected, absorbed or transmitted
as shown in Fig. 3.7. The ratio (ρ) of reflected radiation Eref
e to total incident radiation Ee
on a surface is called reflectance. Likewise, transmittance (τ ) is the ratio of transmitted
radiation Etr abs to E . As
e to Ee , and absorptance (α) is the ratio of absorbed radiation Ee e
done with Ee , the corresponding monochromatic or spectral expression for these ratios is
obtained by specifying specific wavelength or frequency. Therefore,

Eabs
e,λ
Spectral absorptance: αλ = , 0 ≤ αλ ≤ 1 (3.17)
Ee,λ

Etr
e,λ
Spectral transmittance: τλ = , 0 ≤ τλ ≤ 1 (3.18)
Ee,λ

Eref
e,λ
Spectral reflectance: ρλ = , 0 ≤ ρλ ≤ 1 (3.19)
Ee,λ
72 Chapter 3. INTERACTIONS BETWEEN SOLAR RADIATION AND THE ATMOSPHERE

Stemming from the energy conservation principle, these three ratios are related by

ρλ + τλ + αλ = 1 (3.20)

For isotropic conditions Eqs. 3.11 and 3.12 show that Ee and Ie are related by a constant
π. Therefore, absorptance, reflectance and transmittance can be also be defined using Ie .
A black-body is such that ρλ = τλ = 0 and αλ = 1 for all wavelengths. Body is a gas or
solid with uniform temperature and well defined borders. Since a black-body has perfect
absorptance (αλ = 1) and no reflectance nor transmittance (τλ = ρλ = 0), all incident
electromagnetic energy is absorbed. This idealization can be physically approximated
by a cavity, e.g., a box of insulating material with a small hole, (Fig. 3.8). Radiation,
with wavelengths sufficiently smaller than the size of the hole, will enter the box. If its
interior is sufficiently complex and in thermal equilibrium with the enclosure, most of the
radiation will scatter and will not be able escape back.

Figure 3.8. Conceptual box scheme of a black-body. Radiation enters cavity through
opening but no energy leaves. The enclosure is a perfect insulator.

A fundamental property of matter is that a molecule which absorbs radiation of a

particular wavelength also is able to emit radiation of the same wavelength. This leads to
the statement of Kirchoff Law:

α λ = λ or α = , (3.21)

where  is the emissivity. It is defined as as the ratio of the emitted radiation to the maximum
possible emitted radiation at that temperature. Kirchoff’s law applies to all bodies including
a black-body. In this case α =  = 1. Therefore a black-body is also a perfect emitter
or radiator, since it emits the maximum possible radiation at a given temperature. If a
black-body is a perfect emitter, then we can also define emissivity  of an object as the ratio
of the emitted radiation to the radiation emitted by a black-body at that temperature.
 3.3 RADIATION BALANCE AND FUNDAMENTAL LAWS 73

Two facts arise from the black-body definition:

1. The value of the emissivity, , tells us how far the radiation properties of an object are
from a black-body. For example, at 290 K  for grass is 0.90 and dull (not polished)
stainless steel is  =0.21. Therefore, grass behaves ”more like” a black-body than dull
stainless steel.
2. A black-body radiation behaves perfectly isotropically. Therefore, the more the ra-
diation of an object behaves like a black-body, the more isotropic is its radiation, as
shown in Fig. 3.9.

Figure 3.9. Radiation from black and grey body. Radiation of black body is isotrop-
ical while for gray body is not isotropical.

The term surface albedo, (Latin for white), represented by a, is the emissivity of a
surface in the shortwave radiation (λ = 0.15- 4 µm), whereas emissivity λ is reserved for
the longwave radiation (λ = 3-100 µm ). Albedo includes part of the visible spectrum
where surfaces with low albedo appear dark and with large albedo ∼ 1 appear white,
justifying its name.
Table 3.1 shows some radiative properties of natural surfaces. Note that albedo and
emissivity may be dependent on zenith angle due to optical properties and phenomena
that will be explained in Section 3.5.2 on radiation scattering and transmission.
In our everyday experience we have observed objects glowing due to its temperature
with different colors (frequency or wavelength): from a mellow yellow (λ ≈ 0.54 to 0.60
µ m) in a campfire at lower temperatures, to an intense white-blue (λ ≈ 0.44 to 0.49 µ
m) in a steel factory at higher temperatures. Thus, at different temperatures objects emit
electromagnetic energy at different frequencies or wavelengths. What are the radiation
frequencies emitted at different temperatures by the surface of a black-body? The theory
of the energy distribution of black-body radiation was developed by Planck and was first
74 Chapter 3. INTERACTIONS BETWEEN SOLAR RADIATION AND THE ATMOSPHERE

Table 3.1. Radiative properties of natural surfaces.

Surface type Other specifications Albedo (a) Emissivity ()

Water Small zenith angle 0.03 - 0-10 0.92 - 0.97

Large zenith angle 0.10 - 1.00 0.92 - 0.97
Snow Old 0.40 - 0.70 0.82 - 0.89
Fresh 0.45 - 0.95 0.90 - 0.99
Ice Sea 0.30 - 0.45 0.92 - 0.97
Glacier 0.20 - 0.40
Bare sand Dry 0.35 - 0.45 0.84 - 0.90
Wet 0.20 - 0.30 0.91 - 0.95
Bare soil Dry clay 0.20 - 0.40 0.95
Moist clay 0.10 - 0.20 0.97
Wet fallow field 0.05 - 0.07
Paved Concrete 0.17 - 0.27 0.71 - 0.88
Black gravel road 0.05 - 0.10 0.88 - 0.95
Grass Long (1 m) 0.16 0.90
Short (0.02 m) 0.26 0.95
Agricultural Wheat, rice, etc. 0.18 - 0.25 0.90 - 0.99
Orchards 0.15 - 0.20 0.90 - 0.95
Forests Deciduous 0.10 - 0.20 0.97 - 0.98
Coniferous 0.05 - 0.15 0.97 - 0.99

published in 1901, [6]. If radiation energy is isotropic, the irradiance energy per unit
wavelength emitted by a black-body E∗e,λ as function of its surface temperature is given
by Planck’s law:

2
(2πhp λc 5 )
E∗e,λ = c (3.22)
(ehp bλT ) − 1)
where hp is Planck’s constant 6.626 × 10−34 [J s], b is the Boltzmann constant = 1.381 ×
10−23 [J K−1 ], λ is wavelength in meters and c is the speed of light in vacuum.
 3.3 RADIATION BALANCE AND FUNDAMENTAL LAWS 75

The total irradiance E∗e of a black-body is found by integrating Eq. 3.22 over all wave-
hp c
lengths. Upon substituting x = λbT , the radiance is given by
∞ 2πb4 T4 ∞ x3
Z Z
E∗e = E∗e,λ dλ = dx. (3.23)
0 c2 h3p 0 ex − 1
R∞ x3 π4
It can be shown that 0 ex −1 dx = 15 , then

2π 5 b4 4
E∗e = T ≡ σT4 . (3.24)
15c2 h3p

This is the Stefan-Boltzmann law and σ is the Stefan-Boltzmann constant whose

value is 5.67 × 10−8 [W m−2 K−4 ]. Therefore the irradiance of a black-body varies to the
fourth degree with its temperature. To obtain the irradiance Egray
e of a body that is not
black, i.e., gray,  is introduced into the Stephan-Boltzmann law:

Egray
e = σT4 . (3.25)

Therefore, the irradiance of grass, whose emissivity is  = 0.9, is 10 percent less than
corresponding black-body irradiance E∗ for same temperature T.
Solar radiation energy is mainly centered in the visible range (390 to 700 nm). For
this wavelength we can use absorptance, transmittance and reflectance values shown in
Table 3.1.
To obtain the wavelength λmax at which maximum black-body irradiance is obtained,
hp c
Planck’s law is differentiated with respect to λ and equated to zero. Again let x = λbT .
dE∗e,λ
When setting dλ = 0 gives
x
−5 + ex = 0,
(ex − 1)
whose solution is x = 5(1 − e−x ) ≈ 5. Therefore,

hp c 2897
λmax ≈ = µm (3.26)
T 5b T
This result is called Wien’s displacement law. This law indicates that λmax is inversely
proportional to T: the shorter the emitted wave length (higher frequency), the higher the
temperature of the black-body surface. Also, using Wiens’s law if λmax is observed, the
corresponding temperature of the black body surface can be inferred.
The above basic radiation laws have been presented in reverse historic order for clarity.
Actually, Stefan-Bolztmann and Wien laws were first obtained experimentally. Years later
Planck postulated Eq. 3.22, from where Stefan-Boltzmann and Wien laws can be obtained,
as was shown.
76 Chapter 3. INTERACTIONS BETWEEN SOLAR RADIATION AND THE ATMOSPHERE

3.4 The Sun’s electromagnetic field spectrum and its physical effects

The Sun is the main provider of energy to Earth: For every 100 units of the Sun’s energy,
stars provide 10 to 15 units, the full moon 0.01, the Earth’s interior 0.005, and human
actions about 0.01. The Sun’s energy is formed at its nucleus by the fusion of hydrogen
atoms into helium, thus providing a photon. It takes a photon about 140,000 years to
make its journey form the nucleus to the surface of the Sun. Once there, photons travel
towards Earth at speed of light c, carrying their dual identity as a particle and wave, with
preferred wavelengths (or frequencies).
The preferred wavelengths of the expelled photons conform the Sun’s electromagnetic
radiation spectrum. To a good approximation, the distribution of electromagnetic radia-
tion emitted by the sun is of a black-body at a temperature of about 5 778 K, as shown
in Fig. 3.10. Due to its black-body temperature, Wien displacement law indicates that
most of the Sun’s emitted electromagnetic power is centered between 380 nm and 780
nm, within the visible human range. It is considered that the Sun emits in shortwave
radiation range, loosely defined to lie between 100 nm to 500 nm. This range includes
from the near-infrared to the near-ultraviolet spectrum as shown in Fig. 3.11.

Figure 3.10. Blackbody spectrum curves for the Sun and Earth.
 3.4 THE SUN’S ELECTROMAGNETIC FIELD SPECTRUM AND ITS PHYSICAL EFFECTS 77

Figure 3.11. Schematic of the electromagnetic spectrum.

Example 3.7
Power emitted by Sun and Earth.

1. Obtain the power or energy rate of Sun, considering Fig. 3.10.

Using Eq. 3.24,
Ee = σT4 = 63 × 106 [Wm]−2 ,

Considering the Sun as a sphere with radius rs =696 340 [Km], the surface of Sun
is Ss = 6,098 × 1015 [m]. Then the power emitted by the Sun is,

P hie = 3.68 × 1026 [W] .

2. Obtain power emitted by Earth considering data in Fig. 3.10.

3. Obtain [W m−2 ] reaching the top of Earth’s atmosphere from the Sun.

The Sun’s total irradiation for all wavelengths reaching the top of Earth atmosphere is
called the Solar constant. It is denoted by E∞ and its value is 1 362 [kW m−2 ]. In reality,
this is an average that considers eleven year periods of solar maximums and minimums.
The maximum or minimums do not vary more than 0.1 percent from E∞ .
As Fig. 3.10 shows Sun’s incoming radiation overwhelms Earth’s outgoing electro-
magnetic in all wavelengths. Due to Earth’s black-body temperature of 275 [K] as shown
in Fig. 3.10, Wien displacement law places main Earth radiation between 0.04 nm and 1
nm. Therefore Earth’s radiation, mainly due to Sun’s reflection and scattering, is centered
around longwave radiation. This range is mostly in the infrared, as shown in Fig. 3.11.
We defer its study until next chapter, concentrating here on Sun’s shortwave incoming
radiation.
78 Chapter 3. INTERACTIONS BETWEEN SOLAR RADIATION AND THE ATMOSPHERE

A convenient way to learn about the nature and effects of Sun’s electromagnetic radia-
tion spectrum, is by considering how they were discovered. Circa 1800, Frederick William
Herschel, a British-German scientist, performed an experiment that consisted on decom-
posing Sun’s light using a prism, [2]. He placed thermometers to measure heat produced
on each of the the obtained colors, and other ones outside the visible ”rainbow” (or color
bands), for experimental control on prevailing environment temperature . A schematic of
Herschel’s experiment is shown in Fig. 3.12. First he noticed that temperature increased
from the violet to the red part of the spectrum. To his surprise, the control thermome-
ter placed outside the visible rainbow near to the red color showed higher temperature
than the ones in the visible spectrum. He concluded that heat is mainly transmitted in
an invisible part of the solar spectrum that we now call the infrared. This portion of the
spectrum is between wavelengths of 700 nm to 1000 nm, or respective photon energies of
1.24 [meV] to 1.7 [eV].

Figure 3.12. Herschel experiment where thermometers were placed at each of the
visible color bands of decomposed solar light by a prism. The purpose was to mea-
sure corresponding heat at each band. In a thermometer placed outside the visible
bands but near to the red color, substantial heat was measured. Herschel concluded
that heat is mainly transmitted in the invisible infrared.

A year later, a similar experiment was performed by the German physicist Johann
Wilhelm Ritter. In this case he placed one silver chloride-soaked paper on the violet
color and another outside the visible rainbow, near the violet side. He observed that the
paper darkened more quickly there than in the violet light. He called this part of the
invisible spectrum ”oxidizing rays” to emphasize chemical reactivity, to distinguish them
from Herschel’s ”heat rays”. We call this part of the spectrum ultraviolet, and is placed
between 10 nm to 400 nm, or respective photon energies of 3.3 [eV] to 124 [eV].
Photons in their particle-like behaviour provide the energy for photochemical mech-
anisms by dissociating atmospheric gases molecules into highly reactive fragments. This
 3.5 RADIATION PHENOMENA IN THE ATMOSPHERE 79

in turn depends on the probability of a photon, at an appropriate energy (or frequency),

encountering a molecule atom. The radiative quantity pertinent for photochemical reac-
tions is the photon-flux incident on the molecule from all directions and sources in an
air volume. This is called actinic flux, denoted by Ia , in number of photons cm−2 s−1 .
When specifying actinic-flux wavelength dependence, the spectral actinic flux is denoted
by Ia (λ), in cm−2 s−1 nm−1 .
Herschel and Ritter’s experiments provided observable proof of the existence of en-
ergy outside the visible spectrum and their effects: heat on a body due to infrared, and
chemical reactivity due to ultraviolet radiation. Moreover, in the 1960s, the effect of ul-
traviolet radiation on DNA was established, [3]. Section 3.1 on electromagnetic radiation
explained that these effects are manifestations of radiation’s wave-particle duality behav-
ior.
Nowadays we know that the whole electromagnetic spectrum contains energy beyond
the ultraviolet and infrared: from micro-pulsations (106 nm), to Gamma-rays (10−20 nm).
As shown in Fig. 3.11, the visible spectrum occupies a small portion of the total spec-
trum, 390 nm to 700 nm, or respectively, photon energies between 1.7 [eV] and 3.3 [eV].
Some texts comment that ”the human eye has evolved to take advantage of the Sun’s
maximum output at these visible wavelengths,” or ”most plants are green because the
Sun’s maximum output is near to green wavelengths”. As shown by Eq. 3.15 (also check
the corresponding example and exercise), this is not the case since the maximum output
depends on whether Ee is expressed using λ, ν, or another independent variable. Rather,
it appears that evolution of the human eye is related to wavelengths where water is most
transparent. Refer to the paper by Soffer and Lynch, [5].

3.5 Radiation Phenomena in the Atmosphere

Consider a solar beam propagating downward through the atmosphere. The presence
of atmospheric gases and aerosols affect the path and energy of this beam. It will be
necessary to invoke the particle-wave duality to explain the involved physical phenomena.
We start by stating the Beer-Lambert-Bouguer Law, which determines the amount of
radiation reaching the Earth’s surface. This law hinges on determining the correspond-
ing radiation extinction or attenuation, which is the sum of scattering and absorption
phenomena, to be explained in the next subsections. The following synoptic diagram will
help the reader navigating through the rest of the material:
80 Chapter 3. INTERACTIONS BETWEEN SOLAR RADIATION AND THE ATMOSPHERE

  








 Rayleigh

 
 


 
 


 
 







 Scatter- Mie
  
ing

 
 


 
 

  
Opti-

  

Extinc-



cal

 
tion

 


 


 





 


 


 


 

 



 Absorp-

tion

Radiation 














 Trans-

mission


These phenomena are function of the electromagnetic frequency travelling through the
atmosphere. This means that the path of an electromagnetic wave is affected differently by
the presence of a particle or gas molecule depending on the beam’s frequency: scattering
or absorption only takes place at certain selected frequencies. Scattering will mainly imply
deviations of beam path as in a pinball game. Absorption and translation will take place
when a photon interacts with a gas molecule disappearing and its energy transformed
into internal or other form of energy.

3.5.1 Radiative Transfer: Beer-Lambert-Bouguer Law.

Consider a solar beam propagating vertically downward through a slab of effective atmo-
sphere as shown in Fig. 3.13. Let the irradiances at z and z + dz above Earth’s surface be
Ee,λ (z) and Ee,λ (z) + dEe,λ (z) respectively. Theoretically and experimentally it has been
shown that in the section dz of air, the scattered and absorbed electromagnetic radiation
is proportional to dz and to Ee,λ (z). Therefore,
beλ
z }| {
dEe,λ (z) = (bsλ + baλ ) Ee,λ (z)dz + (baλ E∗e + Pscattering )dz , (3.27)

where bsλ is the spectral directional scattering coefficient and baλ is the spectral direc-
tional absorption coefficient. Overall, beλ = bsλ + baλ is the spectral directional extinc-
tion coefficient of air. All these coefficients have m−1 dimensions. For short, the spectral
 3.5 RADIATION PHENOMENA IN THE ATMOSPHERE 81

Figure 3.13. Scattering of a solar beam through a section of atmosphere dz. Note
that the positive direction of z axis is upwards.

and directional terminology will be dropped. The other term in parenthesis in Eq. 3.27
includes the black body emission of the medium itself (atmosphere) and scattering com-
ing form all directions. For the case of a solar beam operating in the visible spectrum, the
effects contained in last parenthesis will be neglected.
Therefore, Eq. 3.27 simplifies to

dEe,λ (z) = beλ Ee,λ (z)dz. (3.28)

Gases and aerosols have their own scattering, absorption and extinction coefficients,
as will be shown in Sections 3.6 and 3.5, where attenuation of atmospheric gases and
particles is discussed.
If scattering, absorption and extinction are divided by mass concentration of a gas or
aerosol, we can define their mass scattering, absorption and extinction efficiencies,

bxλ
αxλ = , x = e, s, a , (3.29)
ρ
where bx=(e,s,a) λ , is the extinction, scattering or absorption coefficient for air, and ρ is air
density. The units of αxλ are [m2 kg−1 ]. The most important absorbing aerosol for solar
radiation is black carbon, as will be discussed in Section 3.6 where solar attenuation by
particles is discussed.
Integrating Eq. 3.28 form a height z to the top of the atmosphere, where the radiative
flux is that of the Sun, Ee,λ,∞ ,
82 Chapter 3. INTERACTIONS BETWEEN SOLAR RADIATION AND THE ATMOSPHERE

Z Ee,λ,∞ dEe,λ(z)
Z ∞
= beλ dz (3.30)
Ee,λ Ee,λ(z) z

The optical depth at wavelength λ between a height z and the top of the atmosphere
is defined as
Z ∞
τλ ≡ beλ dz . (3.31)
z

Despite its name τλ is a dimensionless quantity since beλ has units of [m−1 ]. For urban
atmospheres the mid-visible wavelengths typical values for τλ are between 0.3 and 0.5.
For clean air τλ ' 0.2.
Using Eqs. 3.30 and 3.31, the transmissivity or transmittance, (tλ ), of the atmosphere
above a height z can now be related to optical depth through,

Ee,λ (z)
tλ ≡ = e−τλ , (3.32)
Ee,λ,∞
which is the the Beer-Lambert-Bouguer Law. The steps followed to obtain this law can
be repeated using radiation flux Φe , or radiance Ie , instead of Ee .
 3.5 RADIATION PHENOMENA IN THE ATMOSPHERE 83

Example 3.8
Transmittance and irradiance in the atmosphere.

1. Obtain an expression for the change in irradiance when the solar radiation vertical
beam traverses an atmospheric slab of width ∆z.
Using Eq. 3.32,
Eλ (z) = Eλ∞ e−τλ . (3.33)

If the distance covered by the radiation beam is ∆z, an equivalent expression of

Beer-Lambert-Bouguer Law is,

Eλ (z) = Eλ∞ e−beλ ∆z . (3.34)

2. Obtain change in irradiance at height z when τλ =1 and 0.2.

For τλ =1 and using Beer-Lambert-Bouguer Law,

Eλ∞
Eλ (z) = . (3.35)
e
Therefore at height z the irradiance will be almost a third of the irradiance at the
top of the atmosphere.
3. If the total optical depth of the atmosphere, i.e., the optical depth from the Earth’s
surface to the top of the atmosphere at mid-visible wavelengths is 0.4, what is
the percentage reduction in solar irradiance between the top of the atmosphere
and sea level for a vertical Sun?

The electromagnetic energy that is scattered and absorbed between the top of the
atmosphere and a height z is 1 − tλ . This is the case only for a clear atmosphere where
ρλ = 0. If so the absorptivity, αλ , also can be related to optical depth if it is defined as the
fraction of Ee,λ∞ that is absorbed between the top of the atmosphere and a height z. Therefore,

αλ = 1 − tλ = 1 − e−τλ . (3.36)

Absorptivity values go from the maximum value of 1 asymptotically with increasing

values of τλ , to a minimum of 0 for τλ = 0.
In Fig. 3.14 is shown a typical relationship between αλ and τλ with λ. For certain λ
values absorptivity (transmissivity) is maximum (minimum). In this example, for a 50 m
path length, absorptivity saturation may be reached. Regions in the absorption spectrum
where values of αλ are low are called windows and where αλ are high are absorption
bands.
84 Chapter 3. INTERACTIONS BETWEEN SOLAR RADIATION AND THE ATMOSPHERE

Figure 3.14. Idealized absorption and transmittance spectrum for three path lengths

The optical depth of an atmospheric layer between heights z1 and z2 (z2 > z1 ) is
defined by Z z2
τλ = beλ dz (3.37)
z1

If beλ is independent of z between z1 and z2

τλ ' beλ ∆z (3.38)

where ∆z = (z2 − z1 ) becomes the path length of the layer. Very small path lengths can
produce large values for τλ if for a certain wavelength beλ is large and vice-versa. The
value for τλ depends on the material composition of the atmospheric layer. Also, the value
of the extinction coefficient beλ can be due to scattering, absorption or a combination of
both. This will be discussed in sections 3.5 and 3.6.
In the next paragraphs we follow Horvath, [7]. An extension of atmospheric trans-
mittance defined in Eq. 3.32 for varying substance concentrations inside the path length
has been proposed by Beer. If c0 is a reference concentration and b0eλ is the extinction
coefficient at this concentration, then transmittance at concentration c is obtained by
c
tλ = e−beλ c0 ∆z . (3.39)

For gases such as water vapor this law holds outside the absorption bands. In the
absorption bands, transmittance is approximated using

c

γ
tλ = e−beλ c0
∆z
, (3.40)

where γ is between 0.5 and 1.

In Fig. 3.15 is shown the corresponding absorption coefficients for O3 and NO2 . Note
that the absorption of O3 in the band 200-300 nm is responsible for very low transmittance
 3.5 RADIATION PHENOMENA IN THE ATMOSPHERE 85

(high absorption) in that wavelength range. For the case of NO2 , transmittance is low
(absorption is high) in the blue part of the spectrum.

Figure 3.15. Absorption coefficients for pure ozone and nitrogen dioxide at stan-
dard conditions. Based on Horvath, [7].
86 Chapter 3. INTERACTIONS BETWEEN SOLAR RADIATION AND THE ATMOSPHERE

Example 3.9
Transmittance of some atmospheric gases.

The transmittance, tλ , of some gases at specific concentration and wavelength is ob-

tained.
1. Using Fig. 3.15 the absorption coefficient, baλ of pure O3 at λ = 600 [nm] is 10
[m−1 ]. Obtain corresponding tλ for an O3 concentrations of 100 ppb for a layer
10 km deep.
Considering that baλ = 10 [m−1 ] and that c
c0 = 100 × 10−9 into Eq. 3.39,
−9 ×103
tλ = e−10×100×10 = 0.99 . (3.41)

Therefore, the layer is almost transparent to radiation at λ = 600 [nm].

2. Using Fig. 3.15 obtain tλ for a concentration of 100 µgm−3 of NO2 at λ =450 nm,
for a layer 10 km deep.
Hint:100 µgm−3 of NO2 is 52 ppb.
3. Obtain tλ of a mixture of O3 and NO2 at λ =600 nm with same above concentra-
tions. The layer depth is 10 km.

For the idealized case of spherical particles as constituents of an atmospheric parcel,

the attenuation parameter be can be estimated as follows:
For a gas containing N idealized spherical similar particles of greater size than the
incoming radiation wavelenght, the attenuation be is proportional to addition of the
cross section of particles of radius r,

be = πr2 N . (3.42)

To estimate the extinction coefficient of a gas containing N alike spherical particles

smaller than the incoming radiation wavelength the cross section of the particle πr2
is added, and multiplied by an efficiency factor Qe ,

be = Qe πr2 N . (3.43)

The efficiency factor Qe can be understood as the amount of light attenuated by the
particle relative to the amount of light incident on the cross-section. It can be due to
scattering plus attenuation, i.e., Qe =Qs + Qa . The efficiency factor Qe for spheres are
calculated using Mie theory, as presented in Subsection 3.5.2. The efficiency factors are
 3.5 RADIATION PHENOMENA IN THE ATMOSPHERE 87

dependent on relative size of particle to the wavelength λ, given by the size parameter
2πr
χ= λ .
The efficiency factors Qe , Qs or Qa , compactly expressed as Qx , x=e,s,a, can be di-
vided by mass of suspended particles per volume, mv to obtain an effectiveness per
mass of certain materials to interact with incoming radiation, i.e., the mass extinction
coefficients as discussed in subsection 3.7.1. For spherical particles of radius r,

bx,λ 3Qx
αx,λ = = . (3.44)
mv 4rρ
The Beer-Lambert-Bouguer Law can be applied once the extinction coefficient be,λ
is known. Nevertheless, leaving the explanation for atmospheric radiation transfer at
this point will be incomplete. Deeper knowledge on scattering, transmission, absorption
and translation processes is necessary to understand connections between atmospheric
physics and chemistry. These processes also explain the optical human perception of the
atmosphere.

3.5.2 Scattering and Transmission.

The Rayleigh and Mie theory results presented herein are obtained by solving a wave
equation derived from Maxwell Equations under specific boundary conditions and sim-
plifications, such as isotropy and assuming spherical scattering particles. Therefore, these
phenomena together with optical theory appeal to the wave-like behavior of radiation,
whereas transmission depends on quanta for its definition.

Scattering.
Consider an idealized spherical particle or gas molecule of diameter d. Depending on the
relative size of the impinging wave length λ and d, the following scattering regimes are
defined:
88 Chapter 3. INTERACTIONS BETWEEN SOLAR RADIATION AND THE ATMOSPHERE

Figure 3.16. Rayleigh (left) and Mie (right) scattering patterns as seen by an ob-
server. Respective phase angles θ are also represented.

1
Rayleigh or molecular Scattering. This regime is valid when d ≤ 10 λ, i.e., when
visible electromagnetic waves interacts with atmospheric gas molecules. No energy
exchange of the electromagnetic ray and the medium takes place in this regime. Con-
sider an incident electromagnetic ray with radiance Ie (0) impinging on a gas molecule,
(Fig. 3.16). The perceived radiance Ie of the scattered ray in a direction with an angle
θ to the incident direction by an observer is given by,

π 2 (n2 − 1) σ (h) 1
Ie (λ, θ, h) = Ie (0) (1 + cos2 θ) , (3.45)
2 N λ4
where λ is the wavelength of the incoming light, h the altitude of the point, m=1.00029
the refractive index of air (to be defined when optical scattering is presented below),
Nm =2.504 ×1025 m−3 the molecular number density (number of molecules per cubic
meter) of the standard atmosphere, σ (h) the density ratio defined by

− Hh
σ=e R .

Therefore σ will be equal to one at the surface and zero at the top of the atmosphere.
HR ≈ 8000 m is considered the Earth’s scale height, as presented in Chapter 1.
The version of Rayleigh scattering formula in Eq. 3.45 is used in the area of rendering
and computer graphics. It applies specifically to Earth’s atmosphere. As is the case of
the general version of Rayleigh theory, it expresses the fact that scattering intensity is
inversely proportional to λ4 . This strong wavelength dependence enhances scattering
 3.5 RADIATION PHENOMENA IN THE ATMOSPHERE 89

of short wavelengths. The dependence with θ provides the typical Rayleigh scattering
pattern shown in Fig. 3.16. Notice the typical shape of Rayleigh scattering in which a
receptor will perceive more light in certain directions than others. Note that since the
perceived radiation is a function of position angle θ, the involved radiation parameter
is radiance Ie and not the irradiance Ee .

Why is the sky blue?

Ie (0)
Using Eq. 3.45 the attenuation Ie due to Rayleigh is
I 1
= 4.
Ie (0) λ
If we consider same intensities Ie (0) for all light colors and that the blue wavelength
is circa 470 nm and in the other part of the visible spectrum orange wavelength is
circa 600 nm:
Example obtain scattering pattern. Back front side scatter blue light at the surface?

Mie Scattering. This regime is valid when d ' λ. This is the case when visible
electromagnetic waves interact with atmospheric particles (aerosols) such as pollen,
dust, smoke, water droplets in a cloud or fog. These particles mainly exist near the
Earth’s surface. Mie scattering is responsible for the white appearance of the clouds.
Its typical shape is shown in Fig. 3.16. Among the characteristics of Mie scattering
are: forward scattering is favored while backward scattering tends to be reduced; the
amount of scattering is dependent on λ; energy absorption takes place.
The radiance of the observed scattered radiation at an angle θ to the incident direction
at a distance r is given by,
i1 + i2
Ie = Ie (0) , (3.46)
2k2 r2
where k= 2π
λ , i1 and i2 are the intensity Mie parameters. They are functions of a,
λ, refractive index m (to be defined below when introducing optical scattering) and
the phase angle θ shown in Fig. 3.16 . The Mie parameters refer to the intensity of
light scattering perpendicularly and parallel to the plane through the directions of
propagation of the incident and outgoing light waves. The mathematical procedure
to obtain i1 and i2 and corresponding values obtained by Mie in 1902, can be read in
published tables.
For Mie-scattering due to a a conglomerate of N individual particles, intensity super-
position is invoked,
N
X 1
Ie = Ie (0) (i1 (aj ) + i2 (aj ))n(aj ). (3.47)
j=1
2 k2 r2j
90 Chapter 3. INTERACTIONS BETWEEN SOLAR RADIATION AND THE ATMOSPHERE

Numerical methods are employed to obtain scattering for different aerosol conglomer-
ates. These methods provide visual rendering of atmospheric conditions for different
cases of haze, fog, smog and rain.
Example using values to obtain scattering attenuation See note on HG note Obtain
percentage of back Mie scattering ( θ = 180◦ ) for λ = 555 nm, a= 200 µ m. This
particle size corresponds to a water droplet.
Ie 1
= (i1 (a) + i2 (a))n(a).
Ie (0) 2 k2 r2
where n(a)= 1, and so on.
Optical Scattering. If d  λ such as in the interaction of visible electromagnetic
waves (light) and water drops, the laws of optical geometry are followed. This regime
introduces the refraction effect with the consequence that the amplitude and speed
of the wave entering the medium are altered. Remember the complex form of the
electric field E in Eq. 3.3:
x x
E = E0 eiω(t− c ) = E0 ei2πν(t− c ) .

If the medium dampens the amplitude of the incoming wave, the complex from of
E must be multiplied by an exponential with imaginary exponent. Physically this
will constitute absorption. If the medium also affects the velocity of the incoming
wave, the complex form of E must be multiplied by an exponential with a real expo-
nent. Therefore, a complex number with an exponent m=Re(m) − iIm(m) is used to
accomplish these effects. This exponent is called the refraction index of an optical
medium.
The real part of m, which affects velocity and thus direction of the wave entering the
medium, is the ratio of the speed of light c and the wave velocity v of light in that
medium,
c
Re(m) =
.
v
The wave velocity or phase velocity is the speed at which the crests of the wave move.
Thus, a refractive index with real part m=Re(1.33) (water) means that the wave of
light travels 30 percent faster in vacuum than in the medium. The change in wave
velocity v1 to v2 is dependent on the material and varies with the wavelength (fre-
quency) of light. This change in velocity introduces a refraction angle, deviating the
light wave. The relative angles, shown in Fig. 3.17, are given by Snell law:

m1 sin θ1 = m2 sin θ2 .

This is why prisms divide light into its wavelength constituents (or spectral colors)
as used by Herschel and Ritter in their experiments of Section 3.4. Each color travels
 3.5 RADIATION PHENOMENA IN THE ATMOSPHERE 91

Figure 3.17. Wave light traveling through medium with refractive index m1 entering
medium m2 . Corresponding refraction angles θ1 and θ2 are shown.

at different velocity while traveling the prism, and by Snell Law each color acquires
a different traveling angle thus spreading the light components. Once the separated
waves leave the prism, they resume their speed c, experiencing a delay depending on
each wave frequency.
Based on above discussion, the dampening (absorption) and delay effect phenomena
are expressed by the following mathematical expression,

∆x −iω(Re(m)−1) ∆x −iω(t− xc )
Eafter medium = e−ωIm(m) e c E e
0 B . (3.48)
| {z c}| {z }
B
A

The factor marked B is Eq. 3.3 delayed by factor ω(Re(m) − 1) ∆x

c in traversing the
medium. Factor marked A is an exponential with a negative real exponential, de-
scribing a decrease in magnitude E0 as a function of ω and ∆z. Thus, the wave is
attenuated and leaves the medium with less energy. The term Im(m) is called ab-
sorption index.

Transmission.
Occurs if the photons of solar beam pass through matter unchanged. If photons find
their way to a measuring point via transmission, it is said that the medium (atmosphere)
is transparent. This transparency is dependent on photon energy (or frequency), i.e.,
atmosphere can be transparent for certain photon frequencies and opaque for others, as
will be shown in Section 3.5.1.
92 Chapter 3. INTERACTIONS BETWEEN SOLAR RADIATION AND THE ATMOSPHERE

3.5.3 Absorption and Translation

The electromagnetic radiation energy can be stored in an air molecule by absorption or
translation,

Estored = Eabsorption + Etranslation .

Each of these situations are now explained.

Translation or Kinetic Energy (Temperature).

This energy corresponds to the relative velocity of molecules or atoms in the air and is
not quantized. The average translational kinetic energy of a freely moving particle in a
system with temperature T will be
3bT 1
= mv2 ,
2 2
where v is the root-mean-square speed of molecules, T is temperature, b is the Boltz-
mann constant constant, defined in Planck’s Law by Eq. 3.22 . The root mean square
is obtained from the speed distribution of molecules. For an ideal gas, molecules are
subject to a Maxwell Speed Distribution, so v occupies the root mean of this distribu-
tion. Figure 3.18 shows the corresponding velocity probability density function of differ-
ent temperatures for diatomic nitrogen, N2 . Clearly, kinetic energy and temperature are
equivalent variables: for a gas, temperature is a measure of mean molecular speed.

Figure 3.18. Probability density functions of the speeds of a few noble gases at a
given temperature of 298.15 K (25 C)

For temperatures found in the atmosphere, the kinetic energy of a molecule is gen-
erally small compared to required energy for vibrational transitions. Therefore, energy
 3.5 RADIATION PHENOMENA IN THE ATMOSPHERE 93

is available instead to provoke molecular collisions, carrying away or adding energy in

photon-molecule interactions. This will be reflected in the broadening effect described in
Section 3.5.2.

Absorption.
This phenomenon is a process in which energy of an electromagnetic wave is lost when a photon
impinges on a molecule of air in the atmosphere and causes a transition. By transition is meant
that the arriving photon may cause molecular vibration, rotation, translation, or elec-
tronic transition by modifying the energy level of an atom by changing the orbit, or more
correctly, the orbital of an electron. For all these processes, the energy lost by absorbing
the disappearing photon is gained by the air molecule rising its temperature or internal
energy. Since absorption processes are photonic, they only take allowed quantized dis-
crete values, as was discussed when presenting the concepts of quanta and work function
in Section 3.1. The hydrogen atom and H2 molecule will be used as a simple model to
exemplify the concepts and energetic implications of vibration, rotation, translation and
electronic transition.
Electronic Transition.∗ Consider the case of electronic transition in an hydrogen atom.
The hydrogen allowed energy levels En , n=1,2,3,..., are given by Balmer and Lyman
series obtained experimentally. In 1927 Schrodinger obtained a general formula based
only on quantum physics principles:

hp Rr m e e4
En = − , Rr = n = 1, 2, ... , (3.49)
n2 8h3p 20

where Rr is the Rydberg constant, e is the electron potential charge, 0 is , and me is

the electron mass. At the end, Rr = 3.29 × 1015 Hz, as was first obtained experimentally
by Rydberg. The values n=1,2,3,.., are the principal quantum numbers. When n=1,
it is said that hydrogen is at ground state. A photon is absorbed (or emitted) only at
electronic transitions with wavelengths (or energy) provided by Eq. 3.49.

Example
Calculate the wavelength of the photon absorbed by hydrogen atom causing an elec-
tron transition from level n=2 to n=3.
Solution:
Using and 3.49,

1 1 5
ν = Rr ( 2
− 2 ) = Rr .
3 2 36
From λν = c,
94 Chapter 3. INTERACTIONS BETWEEN SOLAR RADIATION AND THE ATMOSPHERE

c c 36 c
λ= = = .
ν 5/36 Rr 5 Rr
Substituting values for Rr and c:

36 × (2.998 × 108 ms−1 )

λ= = 6.57 × 10−7 m.
5 × (3.29.29 × 1015 s−1 )
This wavelength, 657 nm, corresponds to red color. Photons of this ”color” are ab-
sorbed by hydrogen atom promoting transition from level n=2 to level n=3, and are
given off when the electron falls back to the n=2 level.
Rotational Transition.∗ A molecule, composed of atoms, rotates about an axis through
it center of gravity. Fig. 3.19 shows the axes of rotation for diatomic molecules (H2 ,N2 ,
O2 ), triatomic molecules ( CO2 ) and an asymmetric top molecules such as H2 O and
O3 . If a photon provokes the rotation of a molecule it will lose its energy and disap-
pear. In exchange, rotational energy Enj is gained by the molecule.

Figure 3.19. Molecular masses m1 and m2 with center of mass cm

The H2 molecule is used to exemplify rotational transition for a rigid rotating dipole.
Under this condition, the allowed energies for rotational transition are,

Enj = Bhp c nj (nj + 1) (3.50)

h
where B= 8π2pIr c is the rotational constant,nj , nj =0,1,2,3.., is the rotational quantum
number, hp is Planck constant, Ir is the moment of inertia of the molecule and c speed
of light in vacuum. Ir is found using,

Ir = µr2

m1 m2
where µ = m1 +m2 is the reduced mass, and m1 and m2 are the molecular masses
placed at each end of the molecule axis at a distance r.
The change in energy ∆Enj when rotational transition goes from one state nj to another
nj±1 is given by
 3.5 RADIATION PHENOMENA IN THE ATMOSPHERE 95

h̄2 nj
∆Enj = − (3.51)
Ir
hp
where h̄ = 2 π.

Example:
Vibrational Transition.∗ For vibrational states, the allowed (quantized) energy levels
Eνk are,

1
Eνk = hω0 (νk + ) (3.52)
2
where hp is the Planck constant, νk = 0, 1, 2, ... are the vibrational quantum numbers.
The sub-indexes k =1,2 and 3, denote the vibration modes, i.e., the three different
types of vibrations (Fig. 3.20), which are: symmetric, bending, and antisymmetric,
respectively.

Figure 3.20. Vibration modes: symmetric, bending and antisymmetric.

The change in energy ∆Eνk when vibrational transition goes from one state νk to
another νk±1 is given by

∆Eνk = hω0 (3.53)

Since the quantum rule contains 12 , this formula cannot be applied to νk = 0.

Example
Rotational and Vibrational Transition.∗ Rotational and vibrational transitions occur
simultaneously since molecular vibrations produce oscillating electric dipole sufficient
to cause rotation. The resulting energy level is the sum of both transitions.

Enj ,νk = Enj + Eνk (3.54)

96 Chapter 3. INTERACTIONS BETWEEN SOLAR RADIATION AND THE ATMOSPHERE

Vibrational energy levels are dominant, so rotational energy levels are located in the
vicinity of vibrational levels. The change in energy ∆Enj ,νk for transition from one
state to another is given by

h̄2 nj
∆Enj ,νk = hω0 − . (3.55)
Ir

Example
Ionization potential. It is the characteristic amount of energy relative to the ground
state capable of removing an electron from the atom. For atoms with more than one
electron, the ionization potential usually refers to the most loosely bound electron.
Dissociation potential. It is the characteristic amount of energy relative to the ground
state capable of separating atoms forming a molecule. Its amount is related to the
strength of the bonds forming the molecule.
The concepts introduced here using mainly the H2 molecule, will become handy when
studying radiation absorption effects in the atmosphere. As shown, the quantum aspect
of radiation is prominent to explain radiation absorption.

3.6 Attenuation (extinction) of solar radiation by atmospheric gases

In preceding Section 3.5 was shown that molecular vibration, rotation, electronic transi-
tion, ionization, and dissociation occur at a quantum molecular level. These phenomena
takes place in the atmosphere when gases molecules react with solar radiation. Similar
rules as Balmer and Lyman’s have been found theoretically and experimentally for atmo-
spheric gas molecules such as O2 , O3 , and even more complex molecules such as H2 O,
CH4 , and N2 O. Therefore, specific photon energy levels are necessary for certain atmo-
spheric gases to interact with solar radiation to elicit attenuation. The quantic attenuation
effects on gases are used to explain atmospheric radiation effects.

3.6.1 Main absorptivity effects by atmospheric gases

Figure 3.21 shows the spectrum of solar irradiation at the top of the atmosphere (blue
upper curve) and at sea level (blue lower curve) for an aerosol-free atmosphere. The
mechanism present here diminishing solar irradiance Ee , is due to the absorption by
gaseous molecules. The result of this attenuation effect is the area between the blue
curves showing the irradiance deficit between the top of the atmosphere and sea level
radiation. The main gaseous absorbers are indicated.
3.6 ATTENUATION (EXTINCTION) OF SOLAR RADIATION BY ATMOSPHERIC GASES 97

Figure 3.21. The black curve shows the electromagnetic radiation of a black-body
temperature at 5900 K representing the Sun. If only scattering by gaseous molecules
is present, the upper blue curve shows the irradiation Eλ at the top of the atmo-
sphere and the lower blue curve at sea level . The main absorbers are indicated.
98 Chapter 3. INTERACTIONS BETWEEN SOLAR RADIATION AND THE ATMOSPHERE

In Fig. 3.22 the overall absorptivity at sea level due to CH4 , N2 O, O2 , O3 , and H2 O
is shown. The role played by a selection of most important gaseous absorbers and corre-
sponding quantum physics explanations are described next.

Figure 3.22. Absortivity at sea level due to different gases of the atmosphere.

1. Water vapor, H2 O. This compound is responsible for about 70 percent of of all at-
mospheric radiation, mainly in the infrared region. Since its sinks and and sources
(condensation and evaporation), are determined by climate, its presence is highly
variable.
The vibration modes and quantum numbers for water molecules and respective en-
ergy levels for first level orbitals are shown in Fig. 3.20. Symmetric and asymmetric
vibrations occur at selected wavelengths of 2.73 µm and 2.66 µm, whereas bending
takes place at 6.27 µm. Also water vapor has vibration-rotation band near 6.4 µm and
rotation at wavelengths longer than 12 µm.
Figure 3.22 shows that water vapor absorption occurs in three regions of the spectrum:
Rotational transitions are responsible for absorption in the microwave and far-infrared,
and vibrational transitions in the mid-infrared and near-infrared.
2. Carbon dioxide, CO2 . Strong interaction is found in the infrared range at 4.25 nm
due to symmetrical stretching and at 15 nm due to bending. Therefore, a significant
amount of solar radiation is absorbed by CO2 .
 3.7 ATTENUATION (EXTINCTION) OF SOLAR RADIATION BY ATMOSPHERIC PARTICLES 99

Its mixing ratio rises about 0.4 percent per year, increasing atmospheric optical depth
in the infrared. This modification, mainly due to anthropogenic activities, has crucial
consequences to climate change.
3. Ozone, O3 . This compound has a stretching band near 9.6 µm that is important in
the infrared. Most importantly, has dissociation between 200 and 300 nm. Therefore,
a photon near these wavelengths is absorbed and ozone heats the middle atmosphere
conforming the ozonosphere, attenuating Sun’s ultraviolet bands B and C. This fea-
ture, protects life on Earth as will be explained in Chapter 4.

3.6.2 Line broadening

It is clear form Figs. 3.21 and 3.22 that atmospheric attenuation manifest as bands rather
than precise spectrum transition lines at specific wavelengths (or frequencies). This
phenomenon is called broadening. After broadening, substantial absorption is present
around transition line centers. This is due to the following processes:
Natural broadening. It is related to the intrinsic quantum process of the uncertainty
principle: If we now the energy exactly, we can only know the frequency up to a
finite precision. Due to its magnitude, this process is the least important broadening
process.
Doppler broadening. It results from the relative velocity of a molecule and a photon,
which causes a Doppler frequency shifting. Kinetic energy provided by translational
energy (temperature) plays an important role in increasing molecular speed. This
allows a larger frequency range to produce a particular transition. This process is
dominant at high altitudes.
Pressure or collision broadening. Collisions between molecules or atoms, can supply
or remove small amounts of energy while emitting or absorbing radiation due to
a radiative transition. This allows photons with a broader range of frequencies to
produce a transition. The width of the spectral bands is enhanced by larger density
and pressure, since the probability of collisions increases. Therefore, this broadening
process is dominant in the troposphere.

3.7 Attenuation (extinction) of solar radiation by atmospheric particles

The interaction between solar radiation and polluted environments due to the presence
of anthropogenic or natural particles is now discussed. Although idealized spherical
particles are used to illustrate the involved phenomena, they are good physical models to
100 Chapter 3. INTERACTIONS BETWEEN SOLAR RADIATION AND THE ATMOSPHERE

describe real life pollution situations. The reasons of the importance of black carbon in
attenuation and climate change is accentuated.

3.7.1 Extinction coefficients of gases and particles

In Fig. 3.23 is shown the attenuation and extinction coefficients of particles and gases in
a representative urban atmospheric mixture. Particles in this test mixture have a mean
diameter of 0.5 µm, standard deviation of 1.7 and mass concentration of 100 µg m−3 .
Gases concentrations are: 50 ppb of NO2 ,100 ppb of O3 , 20 percent of O2 , and water
vapor corresponding to 57 percent relative humidity at 20 ◦ C. At most wavelengths in the
visible range, extinction is dominated by particles. Note that the only gas with comparable
extinction is NO2 at 400 nm.

Figure 3.23. Attenuation and extinction coefficients of particles and gases in a rep-
resentative urban atmospheric mixture. Particles containing black carbon exhibit
largest attenuation values spanning the wavelength domain.
 3.7 ATTENUATION (EXTINCTION) OF SOLAR RADIATION BY ATMOSPHERIC PARTICLES 101

Extinction is mainly due to particles containing elemental carbon also called black
carbon. Since elemental carbon is a conductor, its electrons can absorb energy at broad
wavelength bands. This is why extinction is linear within visible wavelengths, and the
decrease with wavelength is mainly due to particle size. In the case of elemental carbon,
absorption takes place when the electromagnetic wave is dampened within the material,
and energy conversion to heat takes place. This is why black carbon is considered an
important contributor to global climate change. The fact that it exists in the atmosphere
for a relative short time (in the order of weeks), provides an opportunity: by cutting
its emissions, mainly coming form incomplete combustion of fossil fuels, biofuels, and
biomass, relatively fast reduction in the rate of climate change can be accomplished.

Example.
At a distance of 10 km how much is due to particles etc. use fig. 5 Horvath
As seen in Subsection 3.5.2, the Mie parameters are dependent on the refractive index
m. The magnitude of m for elemental carbon is large: Depending on the data source, the
real part varies form 1.2 to 2, and from -0.1 to -1.0 for the complex part, thus explaining
its high attenuation in the visible range. For example: Considering elemental carbon
with Im(m)=-1 and thickness 10 µm, corresponding transmission is 10−99 at 550 nm. For
comparison, the Im(m) for crustal material (mainly silica, aluminium and iron) is ≈ -0.02
at 300 nm and ≈ -0.004 at 700 nm.

3.7.2 Attenuation in typical particle polluted atmospheres

As the particle material containing carbon ages, it may be incorporated into water drops.
Also, through coagulation, other materials may adhere forming an internal mixture, i.e.,
particles containing carbon and other substances. Absorption and scattering calculation
for these complex mixtures is difficult and only spherical particles are considered to obtain
the corresponding Mie parameters.
As stated in Subsection 3.5.1, the efficiency factors Qx=a are related with the particles
mass and size as specified in Eq. 3.44 to define mass attenuation. Some values for mass
attenuation contained in the refractive index m for different type of idealized spherical
particles are as follows:
Non absorbing particles. Refractive index: m=1.5 - 0 i and ρ=1500 kg m−3 . This is usu-
ally the case of sulphates or organic materials.
Slightly absorbing aerosol. Refractive index: m=1.5 - 0.01 i and ρ=1500 kg m−3 .
Graphite particles. Refractive index: m=2.0 -1.0 i and ρ=2 250 kg m−3 .
102 Chapter 3. INTERACTIONS BETWEEN SOLAR RADIATION AND THE ATMOSPHERE

Homogeneous mixture of graphite and air. Refractive index: m=2.0 -0.5 i and ρ=1 250
kg m−3 . It assumed a structure containing 50 percent of empty space.
Homogeneous mixture of graphite and air. Refractive index: m=1.25 -0.25 i and ρ=625
kg m−3 . It assumed a structure containing 75 percent of empty space.
The corresponding mass extinction coefficients at a wavelength of 550 nm for these
refractive indexes are shown in Fig. 3.24. A ”resonance” or oscillation for non or slightly
absorbing particles(m= 1.5-0.01 i, and m= 1.5 -0.0 i) is present, but in general mass
extinction decreases inversely proportional to particle diameter above 1 µm. For particles
smaller than 0.1 µm in diameter, two types of behavior are expected: On one hand, the
extinction coefficient for the non-absorbing particles rapidly decreases, since Rayleigh
scattering is proportional to the square of the volume. On the other hand, the extinction
coefficient for small absorbing particles is independent of size and much higher than for
non-absorbing particles. Therefore, small absorbing particles are efficient visible radiation
attenuators.

Figure 3.24. Mass extinction coefficients as a function of particle diameter for a

given wavelength of 550 nm.

As mentioned, black carbon emissions are mainly due to incomplete combustion and
diesel powered transport plays an important role. This is shown in Fig. Patrick where
black carbon scattering in cities are defined by traffic patterns. Morning rush hour (0900 to
1000 LST) concentrations are reflected in these time series. Evening rush hour concentra-
tions are not as large because of higher mixing height and therefore moreconcentrations
dilution. These concepts are to be studied in next chapters.
 3.8 EPILOGUE 103

3.8 Epilogue

An underlying concept in this chapter has been the wave-particle duality of radiation and
how it applies to atmospheric radiation phenomena.

Exercises

3.1 Exercise electron volt to wavelength

0.5 um 2.4796 [eV] 3.97 -19 Joules.
0.7 um 1.7711 [eV] 2.84 -19
(Answer on Page 105)

3.2 to be added
(to be added)
(Answer on Page 105)

3.3 Effective area using cosine theta

to be added
(Answer on Page 105)

3.4 Solar peak

The solar peak, when irradiance is expressed as a function of frequency, is about 2.6 ×
10−12 W m−2 Hz−1 . Obtain the corresponding irradiance when expressed as a function of
wavelength.
(Answer on Page 105)

3.5 Solid angle

What is the solid angle Ω of a sphere?
(Answer on Page 105)

3.6 Electromagnetic energy

(on the electromagnetic energy relationships between an emitter and a receptor)
(Answer on Page 105)
104 Chapter 3. INTERACTIONS BETWEEN SOLAR RADIATION AND THE ATMOSPHERE

3.7 Black-body Venus

Obtain black-body temperature of planet Venus.
(Answer on Page 105)

3.8 Black-body Mars

Obtain black-body temperature of planet Mars.
(Answer on Page 105)
 EXERCISES 105

Answers to Exercises

Answer to Exercise 3.1 (page 103)

Here is the answer for exercise 3.1.

Answer to Exercise 3.2 (page 103)

Answer for exercise 3.2

Answer to Exercise 3.3 (page 103)

Answer for exercise 3.3

Answer to Exercise 3.4 (page 103)

The answer to the solar peak exercises is: Ee,λ=880nm ≈ 1.3 W m−2 nm−1 .

Answer to Exercise 3.5 (page 103)

Answer to: What is the solid angle Ω of a sphere? 4π sr.

Answer to Exercise 3.6 (page 103)

(electromagnetic energy relationships)

Answer to Exercise 3.7 (page 104)

(black-body temperature of planet Venus.)

Answer to Exercise 3.8 (page 104)

(black-body temperature of planet Mars.)
106 Chapter Bibliography

Chapter Bibliography

[1] Maxwell, James Clerk. A dynamical theory of the electromagnetic field. Philosophical Transac-
tions of the Royal Society of London. 155: 459-512, (1865). doi:10.1098/rstl.1865.0008. OL
25533062M. S2CID 186207827. Paper read at a meeting of the Royal Society on 8 December
1864.
[2] Herschel, W. Experiments on the refrangibility of the invisible rays of the Sun. Phil. Trans. Roy.
Soc. London 90, pp284-292, (1800).
[3] Smith, Kendric C. Physical and Chemical Changes Induced in Nucleic Acids by Ul-
traviolet Light. Radiation Research Supplement, vol. 6, pp. 54-79, (1966). JSTOR,
https://doi.org/10.2307/3583551.
[4] John H. Seinfeld and Spyros N. Pandis. Atmospheric Chemistry and Physics. John Wiley and
Sons, Inc.,Wiley Interscience, 1998. ISBN 0-471-17815-2
[5] Bernard H. Soffer,David K. Lynch. Some paradoxes, errors, and resolutions concerning
the spectral optimization of human vision. American Journal of Physics 67, 946, (1999).
https://doi.org/10.1119/1.19170.
[6] Planck, M. Uber das Gesetz der Energieverteilung im Normalspektrum. Annalen der Physik. 4
(3): 553-563: (1901). Bibcode:1901AnP...309..553P. doi:10.1002/andp.19013090310.
[7] H. Horvath Atmospheric light asborption: A review. Atmospheric Environment 27A, 3, pp.
293-317, (1993).
Chapter 4

Interactions between
terrestrial radiation and the
atmosphere

4.1 Earth’s energy budget 108

4.2 Energy Fluxes Near the Surface 111

4.3 Cooling and Warming in the Surface Layer 120

4.4 Empirical schemes to establish the surface radiation budget 123

4.5 A simple Model of the Greenhouse Effect 125

4.6 The first description of Greenhouse Effect 127

Exercises 129

Answers to Exercises 131

Chapter Bibliography 132

Modern satellite technology provides an invaluable tool to observe and measure from
space the Earth’s energy budget, as shown in Fig. 4.1. This will be the starting point to
understand how Earth’s radiation influences the atmosphere. The main focus of this chap-
ter will be determining the energy interactions between energy reflection and absorption
near the Earth’s surface. This allows the description of phenomena such as Urban Heat
Island and the Greenhouse Effect on a cause-effect basis.

– 107 –
108 Chapter 4. INTERACTIONS BETWEEN TERRESTRIAL RADIATION AND THE ATMOSPHERE

Figure 4.1. False color photographs of short and longwave radiation emitted by
Earth. The Clouds and the Earth’s Radiant Energy System (CERES) instrument aboard
NASA’s Aqua and Terra satellites measures the Earth’s reflected shortwave radiation
and longwave radiation emitted into space. Thanks to satellites, it is possible to ac-
curately determine the Earth’s total radiation budget. Credit: NASA/Goddard Space
Flight Center Scientific Visualization Studio.
Parts of the material presented in this chapter benefited greatly from the excellent text
by S. Pal Arya, [1].

4.1 Earth’s energy budget

The Earth’s atmosphere maintains the balance between the energy that reaches Earth,
mainly from the Sun, and the energy that flows from Earth back out to space. As shown
in Fig.3.10 in Chapter 3, the Earth radiates energy similar to that of a black body at approx-
imately 275-300 K. By Wien Law as expressed in Eq. 3.26, the corresponding wavelength
is centered at 11 µm. This lies in the infrared portion of the Electromagnetic Spectrum.
At this wavelength, the electromagnetic wave transmits kinetic energy to the molecules of
a receiving body and heat is generated. Therefore, Earth’s surface radiates back the Sun
energy upwards which is absorbed as heat by the atmosphere.
As shown in Fig. 4.2, the atmosphere further emits infrared radiation upwards and
downwards. As discussed in Chapter 3, absorption and emission of radiation by various
atmospheric gases occur in a series of discrete wavelengths. Figure 4.2 shows that about
50% of solar radiation is able to get to Earth’s surface. The rest is absorbed or scattered by
gases, clouds, and aerosols. About 30% percent of solar radiation is reflected because of
albedo provided by the atmosphere itself, clouds and surface (especially polar regions).
 4.1 EARTH’S ENERGY BUDGET 109

As pointed out above, the Earth’s surface (land and oceans) emit back radiation in the

Figure 4.2. Incoming, outgoing and absorbed fluxes by the atmosphere are shown
in this NASA poster. It is based on the work of many scientists over more than 100
years, with the most recent measurements from the Clouds and the Earth’s Radi-
ant Energy System (CERES; http://ceres.larc.nasa.gov) satellite instrument provid-
ing high accuracy data of the radiation components (reflected solar and emitted
infrared radiation fluxes).

infrared, perceived as heat. About 20% of the incoming radiation is emitted back by
the surface. Only 6% make it back to outer space. The rest is absorbed by atmospheric
gases such as water and CO2 . Heat emission in the form of sensible, and latent heat
contribute with 6% and 24%, respectively. For their definitions see Appendix A. Also
infrared outgoing radiation is emitted by water vapor and CO2 with 38% of contribution.
Clouds emit 26% of outgoing infrared radiation.
All layers participate in absorption/emission, but the near surface region is the most
important because here is where exist the largest concentrations of highly absorbent chem-
ical compounds in the infrared: water vapor, CO2 , methane, and aerosols. In great mea-
sure, these are directly or indirectly produced by living organisms. When left alone, life
provides a thermally steady environment conductive for live itself.
In Fig. 4.3 is shown the infrared radiance spectrum emitted by Earth received by
a satellite. Also shown are the corresponding black body different temperatures. The
area beneath the spectrum represents the total power per unit surface emitted by Earth.
110 Chapter 4. INTERACTIONS BETWEEN TERRESTRIAL RADIATION AND THE ATMOSPHERE

Around wavelength between 13 to 18 µm there is a window, corresponding to a CO2 band,

and another around 8 to 10 µm corresponding to O3 . Close to 7 µm there is a relatively

Figure 4.3. Earth radiance measured from space (solid line) by satellite Nimbus 4
above the Sahara region. Dotted lines are black body temperatures. Gases respon-
sible for main absorption bands are indicated. Adapted from [3].

weak absorption band of methane, CH4 . These windows imply that the atmosphere is
absorbing infrared energy by these gases. If the concentrations of CO2 increases and CH4 ,
the depth of the window will increase, decreasing even more the IR outward emission
because the area under it is reduced. Therefore, in order to maintain equilibrium between
power absorbed from the Sun and terrestrial emission, the surface temperature must in-
crease. This is the Greenhouse effect that provides the thermal conditions for life on
Earth. If altered by humans, a rapidly changing set of climate conditions will develop,
challenging adaptive responses.
In the wavelength range below 7.5µm, water vapor is shown to act as a powerful
Greenhouse effect gas. It effectively absorbs IR energy. Since its scale length H̄ is between
2-3 km, its role in the radiation balance in the lower troposphere is very important.
The above description of Earth’s energy budget describes how the same amount of
energy it receives it must lose, and explains how the thermal equilibrium principle has
been sustained for many centuries. The involved processes control the temperature struc-
ture of the atmosphere, and the temperature at the surface. These will be explained in
next section.
 4.2 ENERGY FLUXES NEAR THE SURFACE 111

4.2 Energy Fluxes Near the Surface

On a flat and homogeneous surface the balance of energy fluxes is,

RN = H + HL + HG + ∆HS (4.1)

where RN is the net radiation flux, H is the sensible heat flux, HL is the latent heat flux
and HG is ground heat flux and ∆HS is the change in energy storage per unit time, per
unit horizontal area, over the whole depth of a layer. This layer may be in the ocean,
atmosphere, or is a soil slab. The arithmetic symbol convention to be used in Eq. 4.1 is:
1. Radiative fluxes directed towards a layer or slab, are positive.
2. Non-radiative energy fluxes directed away from a layer or slab, are positive.
Remember from Chapter 3, that radiative fluxes do not need a physical medium to prop-
agate, while non-radiative fluxes need a physical medium to propagate.
The net radiation RN in Eq. 4.1 is the external force whose response are the sensible,
latent, storage and ground heat fluxes. The energy storage term ∆HS is expressed as,

∂
Z
∆HS = (ρcT )dz , (4.2)
∂t
where ρ is the mass density, c is the specific heat, T is the absolute temperature of the
layer material at some level z, and the integral is over the whole depth of the layer. The
term ∆HS in Eq. 4.1 can be interpreted as the difference between incoming (Hin ) and
outgoing (Hout ) energy from the layer. The terms Hin and Hout are formed by appropriate
combinations of RN , H, HL and HG . A schematic representation of the energy budget in
a layer is presented in Fig. 4.4(a). Note that since HG is a non-radiative flux, its symbol
is positive in the example of Fig. 4.4(a).
When sensible heat energy input to the layer exceeds the output, as shown in Fig. 4.4(b),
by conservation of energy there is an accumulation of energy or flux convergence (
∆HS > 0) which results in warming of the layer. When energy output exceeds input,
as shown in Fig. 4.4(c), energy is lost and the layer cools. In this case, we have flux di-
vergence (∆HS < 0). If the specific heat c of the medium is constant, Eq. 4.2 relates the
rate of energy storage and the rate of warming or cooling of the layer.
The net radiation RN in Eq. 4.1 is a result of the action of shortwave (RS ) and longwave
radiation (RL ) at he surface,
RN = RS + RL . (4.3)

In turn,
RS = RS ↓ +RS ↑ , (4.4)
112 Chapter 4. INTERACTIONS BETWEEN TERRESTRIAL RADIATION AND THE ATMOSPHERE

Figure 4.4. Representation of (a) energy budget of a layer, (b) flux convergence, and
(c) flux divergence

and
RL = RL ↓ +RL ↑ . (4.5)

Therefore the overall net radiation is composed of

RN = RS ↓ +RS ↑ +RL ↓ +RL ↑ , (4.6)

where the downward and upward arrows denote incoming and outgoing radiations, re-
spectively.
The incoming radiation RS ↓ consists of direct-beam solar radiation and that scattered
from the direct solar beam by molecules or particulates in the atmosphere, i.e., diffuse at-
mospheric radiation. This cumulative effect is called insolation. The outgoing shortwave
radiation RS ↑ is actually,
RS ↑= −aRS ↓ , (4.7)

where a is the surface albedo as defined in Chapter 3, Section 3.3. Then RS = (1 − a)RS ↓
and is determined by the insolation at the ground.
A typical diurnal cycle for a grass surface is shown in Fig. 4.5 and for a water surface
in Fig. 4.6. The incoming longwave radiation RL ↓ from the atmosphere in the absence
 4.2 ENERGY FLUXES NEAR THE SURFACE 113

Figure 4.5. Observed radiation budget over a 0.2 m stand of native grass at Mata-
dor, Saskatchewan, July 1971.

Figure 4.6. Observed radiation budget over Lake Ontario under clear skies on Au-
gust 28th., 1969
114 Chapter 4. INTERACTIONS BETWEEN TERRESTRIAL RADIATION AND THE ATMOSPHERE

of clouds, depends on the distribution of temperature, water vapor and CO2 . It does
not show significant diurnal variation. By the Stefan-Boltzmann law, Eq.3.24 in Chapter
3, the outgoing longwave radiation RL ↑ is proportional to T 4 where T is the surface
temperature. Therefore it has a stronger diurnal variation, with a maximum near early
afternoon and minimum value at down. RL ↑ and RL ↓ have comparable magnitudes so
RL is generally small.

4.2.1 Energy Balance During a Diurnal Cycle

Equation 4.1 is followed through a day cycle. During the day, the Earth’s surface receives
radiant energy (RN > 0) partitioned ino sensible and latent heat fluxes to the atmosphere
and heat flux to the submedium. Typically, H, HL and HG are positive during the day.
Schematically this situation is shown in Fig. 4.7(a).

Figure 4.7. Representation of typical surface energy budgets during (a) daytime and
(b) nightime

As shown in Fig. 4.7(b), during the night the surface loses energy by outgoing radi-
ation. This loss is compensated by gains of heat from air and soil media and from latent
heat of condensation released during dew formation. During partially or clear conditions,
 4.2 ENERGY FLUXES NEAR THE SURFACE 115

surfaces loses energy by outgoing radiation. This loss is compensated by gains of sub-
medium heat, and latent heat of condensation released during dew formation. At night
all terms of energy balance are negative, with much smaller magnitudes as those of the
daytime fluxes, except HG . Usually it is considered that HG is proportional to RN or that
there exists an empirical regression relationship between these quantities.

Example 4.1
Example: Estimation of surface energy budget and energy fluxes.

The following measurements were obtained over a dry rural location, on a clear and
calm night:
Outgoing longwave radiation from surface= 400 Wm2 .
Incoming longwave radiation from the atmosphere= 350 Wm2 .
a Calculate the equivalent blackbody surface temperature, and the actual surface
temperature if surface emissivity is 0.95.
Using Eq.3.24 in Chapter 3 and since the outgoing longwave radiation determines
the surface equivalent temperature,Te , RS ↑= σTe = −400 Wm2 . Then, Te =
 1
400 4
σ =289.8 K.
The actual surface temperature Ts is obtained by using the emissivity value =0.9.
 1
400 4
Then, RS ↑= −σTs = −400Wm2 . Then, Ts = σ =293.6 K.
b Estimate the ground heat flux, making appropriate assumptions about other
fluxes.
Using the balance of energy equation Eq. 4.7 with ∆HS ≈ 0, HG = RN −(H +HL ).
Since clear night
h
implies
i
(H + HL ) ≈ 0, then HG = RL ↑ +RL ↓. Then, HG =
(−400 + 350) Wm2 =-50 Wm2 is the ground heat flux.
c What flux would you add if the night is not clear, and what effect does this have
on the results. Draw a sketch of the fluxes directions. Hint: Consider the clouds
reflection.

An important quantity that allows for the comparison between sensible and latent
H
heat is the Bowen ratio, B = HL , that can be estimated independently, [2]. If this ratio is
known H and HL can be obtained by

RN − HG
H= (4.8)
1 + B −1
116 Chapter 4. INTERACTIONS BETWEEN TERRESTRIAL RADIATION AND THE ATMOSPHERE

RN − HG
HL = (4.9)
1+B
The latent heat flux can also be estimated as HL = Le E where Le ∼ 2.45 × 106 J Kg−1
is the latent heat of evaporation/condensation and E is the rate of evaporation/conden-
sation. In [4], a widely used and simple energy balance is presented to obtain H. First
it is assumed that HG = 0.1RN . Substituting this into the energy balance equation and
considering the case when ∆HS ≈ 0,

0.9 RN
H= . (4.10)
(1 + B −1 )

4.2.2 Energy Balance of Surfaces

Examples of observed energy balances of different surfaces are presented and discussed.

Energy Balance for Bare Surfaces

Since mainly a flat and slim surface is considered, ∆HS ' 0. In Fig. 4.8 is shown the energy
balance for a dry lakebed. Here HL = 0, i.e., no evaporation or condensation fluxes. This
an example of extreme thermally climatic environment: At 2m the temperatures diurnal
fluctuations had an amplitude of up to 28◦ C.

Figure 4.8. Diurnal energy budget over a dry lakebed at El Mirage, California, June
10-11, 1950.

For this case it can be observed that the net radiation RN is balanced by the ground
heat and sensible fluxes. The relative values change by large amounts during a diurnal
cycle.
 4.2 ENERGY FLUXES NEAR THE SURFACE 117

Energy Balance for Water Surfaces

Water bodies are fluids with dynamically active surface and boundary layer. Convective
and advective heat transfers within the surface determine HG . Measurement of this flux
is complicated by the fact that shortwave radiation penetrates to a considerable depth in
water. It is necessary to consider a thick enough layer such that convective and radiative
heat exchanges are negligible, i.e., ∼ 10m. Latent heat HL should dominate over sensible
heat H in most cases, and Bowen ratio is much less than unity. By day the water layer is
a heat sink ∆HS > 0, and by night it is a heat source ∆HS < 0.

4.2.3 Energy Balance of Canopies

The term canopy is applied to the atmospheric layer contained between the surface to the
uppermost spreading branchy layer of a forest, or the layer from the surface to the mean
height of roofs in a city. The canopy is where most of life occurs. The energy balance
when the presence of this layer is considered, is discussed next.

Energy Balance for Vegetation Canopies

In the presence of vegetation, surface temperature is no longer an appropriate datum to
carry energy balance calculations. Inside the vegetative canopy the radiative, sensible
and latent heat fluxes are all spatially variable. As a matter of fact, inside the canopy of
a rainforest microclimates niches are defined depending on height, and different animal
species occupy each niche.
To perform an energy balance, it is necessary to consider the whole canopy as a layer.
Measurements of RN , H and HL are needed well above plants or trees. For these to be
representative, horizontal variations of fluxes should be negligible.
The rate of energy storage ∆HS consists of the rate of physical heat storage and the
rate of biochemical heat storage as a result of photosynthesis and CO2 exchange. It is not
easy to measure ∆HS in this case. It can have small values for plants, but larger values
for forests.
The latent heat exchange occurs not only due to evaporation but to a large extent due
to transpiration form plant leaves. The combination of evaporation and transpiration is
called evapotranspiration. It produces fluxes of water vapor that can be nearly constant
or sudden, during the early morning, depending on the type of vegetation.
An example of heat flux measurements for this case is shown in Fig. 4.9. Latent heat
flux due to evapotranspiration is dominant and approximately balances the net radiation,
118 Chapter 4. INTERACTIONS BETWEEN TERRESTRIAL RADIATION AND THE ATMOSPHERE

while H and HG are an order of magnitude smaller. In late afternoon HL exceeds RN

and H becomes negative, i.e a downward flux. In this case, ∆HS ' 0.

Figure 4.9. Observed energy budget of a Douglas fir canopy in Haney, British
Columbia, July 23, 1970.

Energy Balance for Urban Canopies

An appropriate energy balance for an urban canopy can be expressed as

RN + HF = H + HL + HG + ∆HS (4.11)

where the new term HF takes into account the heat flux associated with anthropogenic
energy consumption. The magnitude of HF as a forcing term can be estimated by the per
HF
capita energy use and population density. The ratio RN has been estimated for a number
of cities. The largest values for this ratio is found in densely inhabited cities in middle
HF
and high latitudes. Annually averaged values of RN for Los Angeles is about 0.2 and for
Moscow 3 with typical values of 0.35, stressing the importance of HF . This ratio varies
with season, being larger in winter. The central business and industrial areas tend to be
the ”hot spots”.
Surface albedos in a city can vary between 0.1 to 0.2 due to materials and colors of
streets, roofs and highways , favoring absorption of solar radiation. Through enhanced
turbulence and increased temperatures in the urban canopy, a large portion of available
energy is transferred to the atmosphere via sensible heat. Bowen ratio should be large,
due to the lack of surface water available for evaporation. This situation changes locally
for the case of watered lawns and green parks.
 4.2 ENERGY FLUXES NEAR THE SURFACE 119

Although ∆HS is important, is very difficult to determine from direct measurements.

It is usually determined as the residual term from the energy balance Eq. 4.11. The
contribution to the balance equation by ∆HS is especially important during the night.
Differences between urban and suburban energy balances can be seen in Fig. 4.10.
Even though the suburban case has about 36 percent of constructed area and 64 percent
of lawns and gardens, ∆HS is relatively much higher in the suburban case and HL is
relatively higher in the rural case.

Figure 4.10. Monthly averaged energy budgets at suburban and rural sites in
Greater Vancouver, Canada, during summer.

The differences between urban and suburban cases observed in Fig. 4.10 reveal that
urbanization increases the energy storage capacity ∆HS , and that sensible heat flux H
surpasses the latent heat flux HL . These factors cause the Urban Heat Island, (UHI). It
is defined as the maximum difference in the urban peak temperature and the background rural
temperature, ∆Tu−r . A more complex lists of factors for this phenomenon follows:
120 Chapter 4. INTERACTIONS BETWEEN TERRESTRIAL RADIATION AND THE ATMOSPHERE

Increased incoming longwave radiation (RL ↓) due to absorption of outgoing long-

wave radiation and re-emisson by polluted urban atmosphere.
Decreased outgoing longwave radiation loss (RL ↑) from street canyons due o a re-
duction in their sky view factor by buildings.

Increased shortwave radiation(RS ) absorption by the urban canopy due to the effect
of street canyons albedo.

Greater daytime heat storage (∆HS ) due to the thermal properties of urban materials
and heat release at nighttime.

Addition of anthropogenic heat (HA ) in the urban area in process emission (heating
cooling), transportation, and industrial operations.

Decreased evaporation and, hence, the latent heat flux (HL ) due to removal of vege-
tation and surface waterproofing of the city.
In general all these factors modify the energy balance generating the UHI . Some of
these factors help storing heat during the day in the urban canopy that is released at
nighttime, keeping the urban air warmer.
In Figs. 4.11 and 4.12 are shown surface isotherms in Mexico City. It can be seen that
as urban sprawl expands and urbanization density increases, so is the intensity of the
UHI.

4.3 Cooling and Warming in the Surface Layer

Using the principle of conservation of energy the rate of cooling and warming of a layer
due to change of net radiation with height is calculated. Consider a thin layer between
the levels z and z + ∆z with the corresponding fluxes RN and RN (z + ∆z), as shown in
Fig. 4.13. Then,

∂T ∂RN
   
ρcP ∆z = RN (z + ∆z) − RN (z) ' ∆z , (4.12)
∂t R ∂z
or,
∂T 1 ∂RN
    
= , (4.13)
∂t R ρcP ∂z
   
∂T ∂RN
where ∂t is the rate of change of temperature due to radiation and ∂z > 0 is the
R   R
∂RN
convergence or divergence ∂z < 0 of net radiation. The former leads to warming
R
and the latter to cooling of air.
 4.3 COOLING AND WARMING IN THE SURFACE LAYER 121

Figure 4.11. Average annual temperature (C) in Mexico City for year 1986. Note
the size and strength of the UHI. Credit: Personal communication by Dr. Ernesto
Jauregui.

Figure 4.12. Average annual temperature (C) in Mexico City for year 2006. Note the
size and strength of the UHI compared with year 1986. Credit: Personal communi-
cation by Dr. Ernesto Jauregui.
122 Chapter 4. INTERACTIONS BETWEEN TERRESTRIAL RADIATION AND THE ATMOSPHERE

Example 4.2
Example: Cooling rate due to sensible heat and longwave net radiation.
The following measurements were obtained inside the surface atmospheric layer
at night:
Net radiation at 5 m level= -75 Wm2
Net radiation at 100 m level= -150 Wm2
Sensible heat at surface= -45 W m2
a Calculate the average rate of cooling in the PBL due to divergence of net radiation.
Since at night RN = RL , changes in temperature are due to longwave radiation
directed towards the atmosphere, as shown in Fig. 4.4(b). Using Eq. 4.13,

!
1∂T ∂RN
  
=
R ∂t
ρcp ∂z R
1 −150 + 75
 h i
= h i h i Wm2
1005 kgJK × 1.225 mkg
−3
100 − 5

= −6.41 × 10−4 Ks−1

= −2.31 Kh−1 .

b Calculate the average rate of cooling in the PBL due to divergence of sensible heat
flux. Using Eq. 4.13 but having in mind that that we are dealing with sensible
heat H,

∂T 1 ∂H
    
− =−
∂t H ρcP ∂z H
1 0 + 45 h
  i
=− h i h i W m2
1005 J
× 1.225 kg −3
100
kg K m
= −3.66 × 10−4 Ks−1
= −1.32 Kh−1 ,

c Obtain total rate of cooling

Other factors that influence warming or cooling are turbulent exchange of sensible
heat and advection of warm or cold air. Even though sensible heat is non-radiative,
similar formulas as in Eq. 4.13 can be used, as long as the layer of interest contains a
medium where heat is transmitted.
 4.4 EMPIRICAL SCHEMES TO ESTABLISH THE SURFACE RADIATION BUDGET 123

Figure 4.13. Radiative flux convergence in an atmospheric layer.

In the daytime with clear skies net radiation is dominated by RS that does not vary
much with height in the lower troposphere. Temperature changes with time due to radi-
ation are not significant. We can say that daytime warming in the lower troposphere is
largely due to convergence of sensible heat.
As seen in the Example 4.2, at night the net radiation is entirely RL . Since in the lower
troposphere the concentrations of CO2 , water vapor and other gases are larger, absorption
in the infrared is large and temperature varies strongly with height. In clear sky nights
there are significant amounts of radiation divergence implying heat loss with height. Also
the sensible heat divergence should not be ignored.

4.4 Empirical schemes to establish the surface radiation budget

Parts of the material contained in this section follows closely Section 3 in [5].
Consider the case of a bare thin surface, i.e., ∆Hs ≈ 0. Then,

RN = H + HL + HG . (4.14)

According to De Bruin and Holstslag, a good estimate for HG is

HG = cG RN , (4.15)

where cG = 0.1 was chosen in Eq.4.10. It remains to find H and HL .

124 Chapter 4. INTERACTIONS BETWEEN TERRESTRIAL RADIATION AND THE ATMOSPHERE

If the Bowen ratio B is known through field measurements, then

0.9RN
H= , (4.16)
(1 + B −1 )
H
and by definition, HL = B.
If measurements of RN are not available, it can be approximated using reference en-
vironmental air temperature measurements Tr and the ground balance method in [4]
via,

(1 − r(φ)) R + c1 Tr6 − σTr4 + c2 N

RN = , (4.17)
1 + c3
where c1 = 5.31 × 10−13 Wm−2 , c2 = 60 Wm−2 , c3 = 0.12, σ is the Stefan Boltzman
constant, N is cloud cover, r(φ) is current albedo, and φ is average of solar elevation at
current time t and previous time t − 1,

φt−1 − φt
φ= . (4.18)
2

To obtain R, the estimated solar radiation corrected by cloud cover n, in [7] is pro-
posed that,  
R = R0 1 − 0.75N 3.4 , (4.19)

where R0 is the clear sky insolation. It can be calculated by means of,

R0 = 990(sin(φ) − 30) , (4.20)

where φ is obtained in Eq.4.18. If cloud cover information is not available, a value of

N =0.5 is assumed.
Another method to estimate H and HG is given by Penman- Monteith, simplified by
De Bruin and Holstlag in [8]. The simplified parametrization is given by,

(1 − α) + ( γs )
H= (RN − HG ) − β , (4.21)
1 + γs
α
HG = γ (RN − HG ) + β . (4.22)
1+ s
∂qs Cp
Here, s = ∂T , where qs is the saturation specific humidity, γ = Le , where Cp is the specific
heat of air at constant pressure, Le is the latent heat of water evaporation/condensation.
γ
γ
The ratios s and s
1+ γs
are tabulated as a function of environmental temperature at standard
pressure in Table 4.1. Satisfactory results are obtained using α = 1 and β = 20 Wm−2 .
 4.5 A SIMPLE MODEL OF THE GREENHOUSE EFFECT 125

γ
γ
Table 4.1. Dependence on temperature of s and s
1+ γs
for standard pressure P= 1000 mb.

γ
γ
T(C) s
s
1+ γs

-5 2.01 0.67
0 1.44 0.59
5 1.06 0.51
10 0.79 0.44
15 0.60 0.38
20 0.45 0.31
25 0.35 0.26
30 0.27 0.21
35 0.21 0.17

4.5 A simple Model of the Greenhouse Effect

With the concepts acquired so far it is possible to build a simple Greenhouse Model as
shown in Chapter 2 in [9]. Consider an atmospheric slab shown in Fig. 4.14, that contains
a Greenhouse gas such as CO2 . Therefore, it is mostly transparent to short wavelength
solar radiation, but absorbs longwave terrestrial radiation with an emissvity/absorptivity
. The surface albedo is denoted by α. Both concepts were defined in Chapter 3, Section
3.3. The Earth’s surface temperature is Ts and that of the atmospheric slab is Ta . The
S0
Solar Constant, S0 =1366 W m2 and corresponding incoming solar energy flux is 4 , as
noted in Section 4.1. Having in mind the arithmetic sign convention for radiative and
non-radiative fluxes presented in Section 4.2, two balance equations are obtained:

S0 (1 − α)
= (1 − )σTs4 + σTa4 , (4.23)
4
and

σTs4 = 2σTa4 . (4.24)

Combining these equations,

S0 (1 − α) 
 
= (1 − )σTs4 + σTs4 = 1 − σTs4 , (4.25)
4 2
126 Chapter 4. INTERACTIONS BETWEEN TERRESTRIAL RADIATION AND THE ATMOSPHERE

Figure 4.14. Idealized greenhouse model. The blue arrows denote shortwave (so-
lar) radiative flux density and the red arrow denotes longwave (terrestrial) radiative
flux density. The atmospheric slab, which interacts only with the longwave radiation,
is indicated.

and solving for Ts ,

!1
S0 (1 − α) 4
Ts = . (4.26)
4σ(1 − 2 )

Equation 4.26 relates Earth’s surface temperature Ts with surface albedo α and an
atmospheric slab emissivity/absorptivity, : The radiation properties of the surface or
atmosphere can be changed by varying α or  and observe its effects on the surface tem-
perature Ts .
 4.6 THE FIRST DESCRIPTION OF GREENHOUSE EFFECT 127

Example 4.3
Example: Greenhouse cause and effect
By using the simplified Greenhouse model given in Eq. 4.26, the effect of changing
radiation parameters in the atmosphere will be seen.
a Consider that there is no atmosphere, i.e.,  = 0. Obtain the surface temperature,
Ts .
The corresponding surface temperature under this condition is called the effec-
tive emission temperature , Te . Then,
1
S0 (1 − α)

4
Te = . (4.27)
4σ
Using Solar constant S0 = 1360 W, and a adequate value α = 0.3 we obtain that
Te = 255 K or Te = −15C. This is much cooler than the Earth’s measured surface
temperature of 288 K or ≈ 15 C. This shows the warming effect of the atmosphere,
or the Greenhouse effect.
b Consider  = 0.78 and α = 0.3. These values reflect current atmospheric condi-
tions. Obtain Ts .
Substituting values in Eq. 4.26,

Ts = 1.13Te = 288.15K . (4.28)

Now,Ts = 16.15 C, a value nearer to reality.

c Consider  = 0.82 and α = 0.3. This correspond to an atmosphere doubling
carbon dioxide and the associated positive feedback of water vapor. Obtain Ts
and comment.

4.6 The first description of Greenhouse Effect

For a century and a half the world considered that John Tyndall, an Irish physicist, was
the person who discovered the warming potential of carbon dioxide and water vapor,
even though he published his findings three years after the experiments performed by
scientist Eunice Newton Foote.
The following paragraphs were taken form Physics Today, [10], which describes Foote’s
presentation to the American Association for the Advancement of Science (AAAS) meet-
ing in Albany, New York, on 23 August 1856. She did not present her research. It was
128 Chapter 4. INTERACTIONS BETWEEN TERRESTRIAL RADIATION AND THE ATMOSPHERE

read instead by Joseph Henry, secretary of the Smithsonian Institution. This situation
probably had to do with gender issues.
Foote’s experiment was simple: She placed two identical thermometers in identical
glass cylinders, 30 inches long and 4 inches in diameter. Using an air pump, she exhausted
air from one cylinder and added air into the other. After the temperatures equalized, she
placed the jars next to each other in the Sun and recorded the resulting temperature every
two to three minutes. She also conducted the experiment with both jars in the shade. In
comparing the temperature changes, she observed that “ the (thermal) action increases
with the density of the air, and is diminished as it becomes more rarified.” She repeated
the experiment using moist and dry air by adding water to one cylinder and dehydrating
the other using calcium chloride. She discovered that damp air became significantly hotter
than dry air.
Last, she measured the effect of different gases against “ common air ” (the ambi-
ent atmosphere) and found “ the highest effect of the Sun’s rays . . . to be in carbonic
acid gas.” She noted that after being removed from direct sunlight, carbon dioxide main-
tained its high temperature much longer than other gases did. She also tested hydrogen
and oxygen but listed only their final temperature values. “ What really struck me was the
elegance of her experimental design,” particularly her careful attempt to reduce experi-
mental errors through control groups, says Joseph Ortiz, a climate scientist who recently
analyzed Foote’s research.
The penultimate paragraph of her short paper summarized her groundbreaking con-
clusion: Additional carbon dioxide in the atmosphere would cause global warming, “ and
if as some suppose, at one period of (Earth’s) history the air had mixed with it a larger
proportion than at present, an increased temperature from its own action as well as from
increased weight must have necessarily resulted.”
 EXERCISES 129

Exercises

4.1 Relationship between path length and aλ

For τλ << 1 obtain a linear relationship between the path length and aλ . Discuss the
obtained relationship
(Answer on Page 131)

4.2 Bowen Ratio

Describe the typical conditions in which you would expect the Bowen ratio to be as follows:
a) much less than unity
b) much greater than unity
c) negative
(Answer on Page 131)

4.3 Sensible and Latent Heat Fluxes

Over an ocean surface, the Bowen ratio is estimated to be about 0.3. Estimate the sensible
and latent heat fluxes to the atmosphere if HG = 20 Wm2
(Answer on Page 131)

4.4 Energy Budgets

What are the major differences between energy budgets of a bare soil surface and vege-
tative surface
(Answer on Page 131)

4.5 Measurements at Night

The following measurements were made at night form a meteorological tower in the mid-
dle of a large farm:
a) Net radiation at 5 m level= -75 Wm2
b) Net radiation at 100m level= -150 Wm2
c) Sensible heat at surface= -45 W m2
d) Planetary boundary layer height at = 90 m2
130 Chapter 4. INTERACTIONS BETWEEN TERRESTRIAL RADIATION AND THE ATMOSPHERE

e) Calculate the average rate of cooling in the PBL due to divergence of (a) net
radiation and (b) sensible heat flux.
(Answer on Page 131)

4.6 Measurements on Winter Night

The following measurements were made over a short grass surface on a winter night when
no evaporation or condensation occurred:
a) Outgoing long-wave radiation form surface = 365 Wm2
b) Incoming long-wave radiation form atmosphere = 295 Wm2
c) Ground heat flux form the soil= 45 Wm2
d) Calculate the apparent (equivalent blackbody) temperature of the surface
e) Calculate the actual surface temperature if surface emissivity is 0.92
f) Estimate the sensible heat flux to or from air
(Answer on Page 131)

4.7 Measurements of Radiative Fluxes

The following measurements were made of the radiative fluxes over a short grass surface
during a clear sunny day:
a) Incoming shortwave radiation= 675 Wm2
b) Incoming longwave radiation= 390 Wm2
c) Ground surface temperature= 35 C
d) Albedo of surface= 0.20
e) Emissivity of the surface= 0.92
f) From the radiation balance equation, calculate the net radiation at the surface
g) What would be the net radiation after the surface is thoroughly watered so its
albedo drops to 0.10 and its effective surface temperature reduce to 25C?
h) Qualitatively discuss the effect of watering on the other energy fluxes to or from
the surface
(Answer on Page 131)
 EXERCISES 131

Answers to Exercises

Answer to Exercise 4.1 (page 129)

Answer for exercise 4.1.

Answer to Exercise 4.2 (page 129)

Answer for exercise 4.2.

Answer to Exercise 4.3 (page 129)

Answer for exercise 4.3.

Answer to Exercise 4.4 (page 129)

Answer for exercise 4.4.

Answer to Exercise 4.5 (page 129)

Answer for exercise 4.5.

Answer to Exercise 4.6 (page 130)

Answer for exercise 4.6.

Answer to Exercise 4.7 (page 130)

Answer for exercise 4.7.
132 Chapter Bibliography

Chapter Bibliography

[1] S. Pal Arya Introduction to Micrometeorology. Academic Press, Second Edition. 2001. ISBN
13:978-0-12-059354-5
[2] Bowen Ratio Instrumentation. Campbell Scientific, Inc. Instrument Manual,1987-2005.
[3] Fred Ortenberg Ozone: Space Vision (Space monitoring of Earth Atmospheric Ozone). Graphic
Touch, Ltd. First Edition, 2002. Editor: Prof. M. Guelman.
[4] Oke, T.R. Boundary Layer Climates. John Wiley and Sons, New york, New York, 372 pp, 1987.
[5] Cimorelli, Alan J., Steven G.Perry, Akula Venkatram, Jeffrey c. Weil, Robert B. Wilson, Russell
F. Lee, Warre D. Peters, Roger W. Brode, James O. Paumier. AERMOD: Desciprition of Model
Formulation. EPA-454/R-03-004, September 2004.
[6] Holstlag, A.A.M. and A.P. van Ulden. A simple scheme for daytime estimates for the surface
fluxes from routine weather data. J. Climate Appl. Meteor. 22, 517-529, (1983).
[7] Kasten, F., and G. Czeplak. Solar and terrestial radiation dependent on the amount and type
of cloud. Solar Energy 24, 177-189, (1980).
[8] De Bruin, H.A.R., and A.A.M. Holstlag. A simple parametrization of the surface fluxes of
sensible and latent heat during daytime compared with the Penman-Monteith concpet. J.
Appl. Meteor., 21, 1610-1621, (1982).
[9] Marshall J., Plumb R.A. Atmosphere, Ocean and Climate Dynamics: An Introductory Text. Aca-
demic Press, 2008.
[10] Maura Shapiro. Eunice Newton Foote’s nearly forgotten discovery. Physics Today: People and
History, 23 Aug 2021. DOI:10.1063/PT.6.4.20210823a
Chapter 5

The Atmospheric Boundary

Layer

5.1 A definition of the Atmospheric Boundary Layer 134

5.2 Convective and Stable Boundary Layer 136

5.3 Dry Atmospheric Boundary Layer 139

5.4 Moist Unsaturated Atmospheric Boundary Layer 146

5.5 Saturated Atmospheric Boundary Layer 151

5.6 Local and Nonlocal Static Atmospheric Stability 151

5.7 Pasquill Stability Classes 157

5.8 Pollution, Tropospheric Layers and Inversions 159

5.9 A Simplified Pollution Box Model 163

Exercises 169

Answers to Exercises 171

Chapter Bibliography 172

The atmosphere is arranged in stacks. Using the scale height H, it was shown in Chapter
2 that these stacks obey a minimal potential energy equilibrium state. Nevertheless this
state of equilibrium is continuously disturbed especially in the lowest layer because of
turbulence exerted by Earth’s surface. Figure 5.1 shows a photograph depicting these

– 133 –
134 Chapter 5. THE ATMOSPHERIC BOUNDARY LAYER

stacks. It is in the lower-most stack called the Atmospheric Boundary Layer (ABL), where
pollution accumulates, and where we will focus our attention.

Figure 5.1. View from an airplane, flying north (left), leaving the Valley of Mexico
on February, 2019. The ABL (mixing layer), is clearly marked by the morning haze.
Above the ABL, the exhalation of Popocatepetl Volcano at a height of 5,400 m de-
lineates other layers in the free atmosphere. Just below the Popocatepetl, it can be
seen a forest fire whose smoke emission is contained inside the ABL. Credit: Aron
Jazcilevich.

5.1 A definition of the Atmospheric Boundary Layer

In fluid dynamics a boundary layer is defined as the layer of a fluid in the immediate vicinity
of a material surface in which significant exchange of momentum, heat, or mass takes place between
the surface and the fluid. The atmospheric boundary layer (ABL) or planetary boundary
layer (PBL) is a consequence of the interaction between the atmosphere and the planetary
surface. The interaction is not only mechanical as in aerodynamics, but also through heat
exchanges. Therefore it can be defined as the part of the troposphere that is directly influenced
by the presence of the Earth’s surface and responds to surface forcing (mechanical and thermal)
with a timescale of less than an hour. This definition is based in Stull, on [1]. The surfaces can
be urban, crop fields, woods, desert, lakes or oceans. Therefore, forcing include frictional
drag, terrain induced flow modification, evaporation and condensation and heat transfer.
Pollutant emissions and sinks constitute additional forms of interaction.
 5.1 A DEFINITION OF THE ATMOSPHERIC BOUNDARY LAYER 135

The time scale of one hour allows the introduction in the ABL definition of thun-
derstorms, cumulus, tornadoes, plumes, wakes, dust devils and small scale turbulence.
These phenomena occur within spatial scales below 20 km. Exclusion is made of atmo-
spheric phenomena such as synoptic cyclones, hurricanes and global general circulation.
Many of these have their genesis on the planetary surface, but their effects take long time
to become apparent in another region.
The ABL structure is described differently by authors depending on their research
goals. We will combine Stull [1] and Kaimal [2] conceptions. A schematic of the ABL is
shown in Fig. 5.2. Three layers form the ABL:
1. A roughness layer. This bottom layer is conformed of the specific land use or orog-
raphy where the atmosphere is located. If the land use is urban, principles of urban
meteorology are applied to study flows due to buildings and houses, if crops, agricul-
tural meteorology is used, or if oceans, atmospheric-ocean interactions are studied and
applied. Examples of other roughness elements are forests, grasslands or mountains.
2. A surface layer. This layer, that is included in the roughness layer, plays an important
role in atmospheric pollution. Its depth varies from 50-100 m. Here, flow is insensitive
to Earth’s rotation and wind structure is determined by surface friction and vertical
gradient temperature. Since Earth’s rotation has no influence, shearing stress is con-
stant in the vertical. While the interior of the ABL is homogenized by turbulence, the
surface exhibits strong vertical gradients of temperature, humidity, pollutants concen-
trations and wind due to the strong coupling with the ground. This takes place in the
bottom 10% of the ABL.
3. An outer layer placed above the surface layer extending to a height 500-1,000 m or
above, where shearing stress is variable and vertical wind profile is influenced by
surface friction, temperature gradient and Earth’s rotation.
136 Chapter 5. THE ATMOSPHERIC BOUNDARY LAYER

Figure 5.2. Depiction of the planetary boundary layer.

Vertically, the ABL can reach only a few meters during the night, to about 2-3 km dur-
ing day time and may span the whole troposphere. Observations during the MILAGRO
campaign has shown that in Mexico City the ABL can reach about 3.5 km, [3]. As shown
in Fig. 5.3, a stark difference between the ABL and free atmosphere can be appreciated
by observing diurnal temperature variations: In the free atmosphere diurnal temperature
variations are barely registered while in the ABL the variation is clearly noticed. This
shows the direct influence of surface thermal behavior on the atmospheric layer above it
in hourly time scales.

5.2 Convective and Stable Boundary Layer

The daily surface thermal variability induces the formation of the dynamic structures in
the ABL summarized in Fig. 5.4. After sunrise, the atmosphere is warmed by turbulent
heat flux from the ground and a well mixed layer replaces the stable layer left during pre-
vious night. The mixed layer is bounded above by an entrainment zone, where exchanges
of energy and material take place between the ABL and the troposphere. The entrainment
layer has a depth of about 10 percent the depth of the ABL. These daytime layers com-
prise what is known the Convective Boundary Layer or CBL. Just before sunset a stable
layer starts forming to later become the night time layer nearest to the ground. Above it,
a residual layer is formed containing material left that reached sufficient height during
 5.2 CONVECTIVE AND STABLE BOUNDARY LAYER 137

Figure 5.3. Daily temperature variation at two heights: near the ground (97.5 kPa)
and roughly at 1,155 m (85 kPa).

sunlight . This layer is much less turbulent than the daytime mixed layer. It is capped by
an inversion that signals the reaches of the ABL. These nighttime layers comprise what
is known as the Stable Boundary Layer or SBL.

Figure 5.4. Daily cycle of the ABL. From [1].

After sunrise, a new mixed layer from the ground starts to be substituted for the pre-
vious, renewing the cycle. If meteorological conditions are favorable for a highly stratified
138 Chapter 5. THE ATMOSPHERIC BOUNDARY LAYER

atmosphere (cold days, low wind), the residual layer may partially survive. Otherwise
the mixed layer rises, substituting , the residual layer completely. The material in the
residual layer becomes part of the mixed layer, participating in all physical and chemical
processes taking place there. This phenomenon is called entrainment. When low pres-
sure synoptic conditions are dominant, the transition from an a Stable Boundary Layer
(SBL) to a Convective Boundary Layer (CBL), may take an hour or less. When a sta-
tionary high pressure system is dominant, this transition may take more than four hours,
providing conditions for adverse pollution episodes.
The transition periods from CBL to SBL and viceversa are of importance for air quality.
During these periods a change of sign in the heat flux H takes place, and therefore net
radiation is near zero. Atmospheric mechanisms changing mixing and pollution dilution
conditions are taking place. The transition period can be isolated in terms of the solar
angle φ, as described next, [4].
An empirical scheme to obtain RN was presented in Chapter 3,

(1 − r(φ)) R + c1 Tr6 − σTr4 + c2 N

RN = , (5.1)
1 + c3

where φ is solar elevation, N is the cloud cover, c1 = 5.31 × 10−13 W m−3 K−6 , c2 = 60
W m−2 , c3 = 12, and σ = 5.67 × 10−8 Wm−2 K−4 is the Stefan-Boltzmann constant. The
albedo is calculated using,
r(φ) = r0 + (1 − r0 )eaφ+b , (5.2)

where r0 = r(φ = 90),a = −0.1 and b = −0.5(1 − r0 ). By setting R0 in Eq. 5.1 to zero,
and solving for sin(φc ), the critical transition solar angle φc between CBL and SBL can
be determined by,
" #
1 −c1 Tr6 + σTr4 − c2 N
sin(φc ) = + 30 . (5.3)
990 (1 − r(φ))(1 − 0.75N 3.4 )

In broad terms, the transition from stable to convective conditions occurs when φc = 13,
and when overcast φc = 23. Some air quality models such as AERMOD, [5], determine
φc , using an estimate of cloud cover. In Eq. 5.3, an equivalent cloud cover Ne is calculated
using Eq. 5.2, i.e.,
1
1 − R/R0
 
3.4
Ne = . (5.4)
0.75
This cloud cover Ne , is substituted in Eq. 5.3, and a better estimate of φc is obtained.
The meaning of stability, mixing, inversions, entrainment and the physics behind these
surface-atmosphere mechanisms are explained in the remaining part of this chapter. Fur-
ther atmospheric processes such as wind, temperature and mass distribution profile with
 5.3 DRY ATMOSPHERIC BOUNDARY LAYER 139

altitude will be studied in Chapter 5. A more formal treatment of turbulence will be

presented in Chapter 7.
The case of a dry atmosphere is discussed next. It provides the first building block in
understanding the thermodynamics of the near surface atmosphere.

5.3 Dry Atmospheric Boundary Layer

Near the surface the temperature changes with height. The knowledge of the rate of
change of temperature T with height z provides critical information about atmospheric
structure and stability. Here, we consider the case of a dry atmosphere. With modi-
fications, we will then be able to handle the moist unsaturated case presented in next
section.
The main thermodynamics formulas needed are:

1. Hydrostatic Equation.
The pressure p in the lower atmosphere decreases with height according to the hy-
drostatic equation
∂p
= −ρg , (5.5)
∂z
where ρ [kg m−3 ] is the air density, z [m] is the altitude and g [m s−2 ] is the gravita-
tional acceleration. We will consider for convenience g = 9.8[ m s−2 ]. The term ρg in
Newtons,[N], accounts for the weight of the fluid: an increase of weight with depth
causes an increase of pressure and vice versa, therefore the minus sign in Eq. 5.5.
This equation in the ABL is only valid when the atmosphere is at rest, i.e., in hydro-
static equilibrium. Otherwise changes in pressure due to vertical wind components
must be taken into account.
2. Adiabatic Conservation Law.
An adiabatic process occurs when there is no transfer of heat or mass between the volume of
interest and its surroundings. This is the case in a closed system. So when we consider
an air parcel in the atmosphere, it will be isolated from radiation heating or cooling, or
any other energy transfer mechanism with the medium outside. Mass will not cross
the boundaries of the parcel, but the volume occupied by the parcel may expand or
contract thus providing a mechanism to contract or expand the gas inside the parcel.
Under these conditions, pressure and density are related by,
γ
p ρ

= (5.6)
p0 ρ0
140 Chapter 5. THE ATMOSPHERIC BOUNDARY LAYER

Cp
where γ = Cv , Cp =1.005 [J kg−1 K−1 ] is the heat capacity at constant pressure and
Cv = 718 [J kg−1 K−1 ] is the heat capacity at constant volume, when at 300 K. Finally,
p0 [kPa] and ρ0 [kg m−3 ] are reference pressure and densities. This is the Adiabatic
Conservation Law.
3. Law of Perfect Gases
We will consider the atmosphere as a perfect gas. The equation of state for these
gases is given by,
R
p= ρT = Ra ρT , (5.7)
Ma
where R= 0.8314 × 10−3 [m−3 kPa mole−1 K −1 ] is the Universal Gas Constant, T [K]
is absolute temperature, ρ is the air density, Ma is the molecular mass of air, Ra = [kJ
K−1 kg−1 ] is the specific gas constant for air. For pressures and temperatures present
in the atmosphere, the error of this equation is less than 0.2% when compared with
the more complex Van der Waals Equation, [3]. This error is well within the accuracy
provided by atmospheric instruments.
Expressing pressure and density in terms on temperature, the adiabatic conservation
law can be written in terms of the temperature,

γ
p T
 
γ−1
= , (5.8)
p0 T0
or,
1
ρ T
 
γ−1
= . (5.9)
ρ0 T0
Combining the preceding equations to eliminate p and ρ we obtain,

∂T γ−1 g g
=− =− ' −10 K km−1 . (5.10)
∂z γ R Cp

This result says that as we ascend in a dry atmosphere under adiabatic conditions, a
reduction in temperature of about 10 K (or ◦ C) takes place for every km. This is called the
dry adiabatic lapse rate and is denoted by Γd . Therefore, Γd = 10 [K km−1 ]. Remember
that for a Standard Atmosphere mentioned in Section 2.1 of Chapter 2, the Environmental
Lapse Rate is Γe = 6.5 [K km−1 ]. Its purpose is to provide a useful average reference
value valid in many locations, and is obtained using other suppositions to obtain Γd .
The adiabatic conservation law can be written

 γ−1
T p

γ
= , (5.11)
T0 p0
 5.3 DRY ATMOSPHERIC BOUNDARY LAYER 141

or,
 γ−1
p

γ
T = T0 , (5.12)
p0
where T0 is the temperature corresponding to the reference pressure p0 . This is the Pois-
son Equation.
The potential temperature, denoted by θ, is defined as the temperature that a parcel
would have if it were brought adiabatically to a given reference pressure p0 , as shown in Fig. 5.5.
So if − (γ−1)
γ = κ = 0.286, and considering as a reference sea pressure level, p0 ≈100 kPa,
then,
κ κ
p0 100
 
θ=T = T . (5.13)
p p
The advantage of using potential temperature is that the systematic subtraction from
the adiabatic gradient is avoided. In this way the adiabatic or neutral case is easily recog-
nizable as a vertical line in the θ − z plane as shown if Fig. 5.6. Also, since the potential
temperature of an adiabatic parcel is conserved during vertical movements, the parcel
may be tagged using its potential temperature value. These features will be appreciated
when atmospheric stability is discussed in Section 5.5.
142 Chapter 5. THE ATMOSPHERIC BOUNDARY LAYER

Figure 5.5. Potential temperature of a parcel at 70 kPa is T = -4 ◦ C. At 1000 kPa is 25 ◦ C.

Figure 5.6. In (a) environmental temperature with height. Adiabatic or neutral case
is shown with a 10 ◦ C slope . In (b) potential temperature with height. Neutral or
adiabatic case corresponds to a vertical line.
 5.3 DRY ATMOSPHERIC BOUNDARY LAYER 143

Example 5.1
Example: Pressure coordinates to height coordinates.

a Estimate the height in meters of the pressure coordinate of 85 kPa displayed in

Fig. 5.3.
As a rough estimate, a constant ρ = 1.225 kg m−3 will be used in Eq. 5.5. It cor-
responds to sea level conditions at 15◦ C. These temperature and pressure values
are frequently used. They are called Standard Temperature and Pressure, STP,
conditions. Integrating Eq. 5.5,

p − p0 = −ρg(z − z0 ) , (5.14)

or,

kgm−2
(101.32 − 87.5) [kPa] × 101.97 [ ] = 1.225 [kgm−3 ] (z) [m] . (5.15)
kPa
Then z=1,155 m.
b Same as above, but consider an average temperature of 25◦ C.
Using Eq. 5.7 the density of air at T = 298.15 K and pressure p = 101.32 kPa
is ρ =1.204 kg m−3 . These temperature and pressure values are frequently used
and are called normal temperature and pressure, or NTP conditions. With this
value of ρ,

kgm−2
(101.32 − 87.5) [kPa] × 101.97 [ ] = 1.204 [kgm−3 ] (z) [m] , (5.16)
kPa
so z=1,170 m.
c Consider STP surface conditions to estimate height in meters of the 65 kPa pres-
sure coordinate by first using an intermediate pressure coordinate at 85 kPa, and
from this information obtain final result. Compare results when obtaining height
without using intermediate point.

Another important advantage of using potential temperature, is that when comparing

temperature data between locations at different heights, the obvious orographic (height)
difference is eliminated by referring temperatures to a common reference point. This
allows observing other phenomena influencing temperature behavior, besides height.
144 Chapter 5. THE ATMOSPHERIC BOUNDARY LAYER

Example 5.2
Example: Detecting Urban Heat Island (IHU) using potential temperature

a The average night temperature reported by a meteorological station in the center

of a city at a height of 2,220 m ASL is ◦ C. At the same time, another station in
a rural area outside the city reports ◦ C at a height of 2400 m ASL. Obtain the
potential temperature at both stations.
b Compare and comment on differences of environmental and potential tempera-
tures at both stations.
c Refer potential temperatures to 2,220 m ASL. Comment.

Most of the time, the observed or environmental potential temperature does not re-
main constant as we ascend, i.e., the atmosphere is non-adiabatic or non-neutral. Some
handy formulas to obtain potential temperature from measured data are presented next,
based on Arya ref.
Equation 5.13 is used to obtain an approximation of the vertical variation of potential
temperature, ∂θ if the corresponding variation of environment temperature ∂T is known,

∂θ θ ∂T ∂T
= ( + Γ) ' +Γ. (5.17)
∂z T ∂z ∂z
This equation works well when potential and actual ambient temperatures do not
differ by more than 10 %. Using finite differences, see Appendix II, an approximation to
this equation is,

∆θ = ∆T + Γ∆z . (5.18)

This approximation is useful to obtain the difference in potential temperatures between

two height levels.
An integral version of Eq. 5.18 is,

θ − θ0 = T − T0 + Γ∆z . (5.19)

This expression is useful to convert a temperature sounding into a potential temperature

if no aloft pressure information is known. If surface pressure is known, θ0 is obtained
using Eq. 5.13 and T0 can be inferred.
 5.3 DRY ATMOSPHERIC BOUNDARY LAYER 145

Example 5.3
Example: Obtaining potential temperature profiles.

a Assume a dry atmosphere. Obtain potential temperature vertical profile from the
environmental temperature in Table 5.1.

Table 5.1. Environmental temperatures vs. height. Data for Example 5.3.

Height above surface(m) 2.0 3.5 6.5 11.5

Air temperature 27.81 27.72 27.62 27.61

Using Eq. 5.19 with θ0 = T0 , results are displayed in Fig. 5.7

b Interpret results.
At a first glance, the potential temperatures profile in Fig. 5.7 indicates a stable
or subadiabatic atmosphere at the top, something that might be missed while
glancing through the environmental temperatures profile.
c Can you describe the adiabatic motion of a parcel below the stable layer?

Figure 5.7. Environmental a temperature T and potential temperature θ for

data in Table 5.1
146 Chapter 5. THE ATMOSPHERIC BOUNDARY LAYER

As will be shown next, the material presented for the dry atmosphere can be modified
to tackle the moist unsaturated case, and obtain similar thermodynamic expressions.

5.4 Moist Unsaturated Atmospheric Boundary Layer

The presence of water vapor in the atmosphere is now considered. The material in this
section is valid as long there is no phase change, i.e, no water condensation or evapora-
tion. This is because either will respectively imply latent heat release or absorption, and
therefore adiabatic conditions are not maintained. Moreover, when condensation occurs,
droplets will fall crossing the boundary and the system cannot be considered closed.
The most often used variables to express the moisture content of air, or its degree of
saturation, are:

1. Vapor pressure. In Appendix III the concepts of saturation and vapor pressure are
illustrated. Formally, vapor pressure of a substance is the pressure exerted by its vapor
when the vapor is in dynamic equilibrium with the condensed phase.
The water vapor pressure, e, in air can be obtained from the equation of ideal gases
written in the form,
V
e = eα = Rv T , (5.20)
n
V
where V is volume, n number molecules, α = n becomes the specific volume of water
vapor and Rv =461 [J K−1 kg−1 ] is the gas constant of water vapor. The saturated
case is treated similarly, and
es α = Rv T , (5.21)

where es is the saturated vapor pressure.

For typical temperature and pressure atmospheric conditions, the August-Roche-
Magnus formula provides an approximation of vapor pressure:
17.625 T
es = 0.61094 e T +243.04 , (5.22)

where es is in kPa and T is in Celsius.

Note that in meteorology, the term vapor pressure is used to mean the partial pressure
of water vapor in the atmosphere, even if it is not in equilibrium. The equilibrium
vapor pressure, es , will be specified as such.
 5.4 MOIST UNSATURATED ATMOSPHERIC BOUNDARY LAYER 147

Example 5.4
Example: Water evaporation at different heights and temperatures.

a Obtain es for water vapor at a standard atmosphere.

Using Eq. 5.22 with T =15 ◦ C, es =0.170 kPa. Since es is larger than standard
barometric pressure at sea level of 101.325 kPa, evaporation is in process.
b Same T as above but at 2 200 m above sea level.
Using the hydrostatic equation 5.5, atmospheric pressure at 2 200 m is p =0.755
kPa. Since es is larger than estimated atmospheric pressure and the differ-
ence is larger than the case at sea level, a faster evaporation rate is expected.
c Obtain the height of a location given the corresponding boiling temperature
of water. Check your findings knowing that the height of Mexico City is
2,200 m and water boils there at about 92.4 ◦ C.

The most often used variables to express the moisture content of air are,

2. Mass mixing ratio.

This is defined as the ratio of the mass of water vapor, mv , to the mass of dry air, md ,
contained in an air sample,
mv [g]
r= . (5.23)
md [kg]
Although r is a dimensionless quantity, since the mass contents of water in the atmo-
sphere is very low, the numerator is usually expressed in g and the denominator in
kg.
According to Dalton’s Law, p = pd + e, vapor and dry air occupy same volume, as we
see by multiplying and dividing r by V,

mv /V ρv
r= = , (5.24)
md /V ρd
where ρv and ρd are the densities of vapor and dry air respectively.

3. Specific humidity.
We will be using specific humidity, Q, as our choice of moisture variable. This is
defined as the ratio of mass of water vapor to the total mass of moist air,

mv [g]
Q= . (5.25)
(mv + md ) [kg]
Although Q is a dimensionless quantity, since the mass contents of water in the atmo-
sphere is very low, numerator is usually expressed in g and denominator in kg. By
148 Chapter 5. THE ATMOSPHERIC BOUNDARY LAYER

Dalton’s law and carrying same procedure to find Eq. 5.24, we can get
ρv
Q= . (5.26)
ρv + ρd
Since mv  md , in practice the values for r and Q do not differ significantly and their
magnitude is much less than one, so r ' Q  1. By writing the Universal Law of
Gases in the form,
V e V pd e
mv = and md = , (5.27)
Rd T Rd T
where  = Mv /Md =0.622 is the ratio of vapor and dry air molecular masses, and
remember that pd is pressure of dry component, e is vapor pressure and Rd = 287
[J K−1 kg−1 ] is the gas constant for dry air. It is used for the dry and vapor state
equations, since contents of water in atmospheric applications is less than 4 %. Then,
mv V e Rd T e
Q= = = (5.28)
md Rd T V p d p−e
or,
e e
Q = 0.622 ≈ 0.622 . (5.29)
p−e p

4. Relative humidity.
This is defined as the percentage ratio of the actual mixing ratio, r, to saturation mixing
ratio, rs , for a given temperature and pressure, or equivalently, the ratio of mass of
water vapor, mv , and mass of saturated vapor, ms ,
r mv
RH = 100% × = 100% × (5.30)
rs ms
In practice, it is convenient to use
r
RH = 100% × , (5.31)
rs
where rs is saturation mixing ratio. This is because mixing ratios are usually available
from soundings plotted in meteorological diagrams. Relative humidity is the most
commonly measured and reported moisture variable since it is relatively easy to mea-
sure. It provides information about how near the air is to saturation, but does not
provide information about the mass of water contained in an air sample.

5. Absolute humidity.
This is defined as the density of water vapor component of moist air,
e
ρv = , (5.32)
Rv T
where e is the vapor pressure, Rv is the gas constant of water vapor, and T is temper-
ature.
 5.4 MOIST UNSATURATED ATMOSPHERIC BOUNDARY LAYER 149

Virtual Temperature
Using Dalton’s law of partial pressures (p = pd + e) the equation of state for moist air
can be written as

R R md R mv
p= ρT = T+ T , (5.33)
M Md V Mv V
or,
R Md
p= ρT (1 + ( − 1)Q) , (5.34)
Md Mv
where Md and Mv are the molecular masses of dry air and water vapor, md and mv are
the air and water vapor masses and m is total mass of air contained in the air sample.
Md
Since Mv − 1 =0.608, a useful expression of Eq. 5.33 becomes

p = Rd ρT (1 + 0.608Q) , (5.35)

where Rd is the specific gas constant for dry air.

Example 5.5
Example: Obtaining specific humidity Q and air density from measured air pres-
sure, temperature and relative humidity.

a If in a location measured air pressure is p = 94.5 kPa, T = 2.85 ◦ C and RH= 46%,
find mass specific humidity Q.
Using Eq. 5.22 with T =2.85 ◦ C, es = 0.75 kPa. Since RH=46%, e = 0.46 ×
0.75 =0.345 kPa. Then, using Eq. 5.29, the specific humidity is,

0.345[kPa] [kg]
Q = 0.622 × = 2.3 × 10−3 . (5.36)
94.5[kPa] − 0.345[kPa] [kg]

b Find corresponding mass density of moist air.

From Eq. 5.35,
p kg
ρ= ) = 1.16[ 3 ] (5.37)
Rd T (1 + 0.608Q m
kg
Note that moist air has lower density ρ than dry air, ρd = 1.225 [ m 3 ]. This is

always the case since water vapor molecule has lower molecular mass than dry
air molecules. Therefore, a moist air parcel is ”lighter” than dry air, and will tend
to float upwards.
c Using data above, obtain mass mixing ratio r at STP conditions. Comment on
this result.
150 Chapter 5. THE ATMOSPHERIC BOUNDARY LAYER

The dimensionless term in brackets in Eq. 5.35 multiplies temperature T . We define

its product with T as virtual temperature, TV ,

TV
z }| {
p = Rd ρ T (1 + 0.608Q) . (5.38)

Therefore, TV is
TV = T (1 + 0.608Q) . (5.39)

Equation 5.34 now acquires the same form of the state equation 5.7 with TV playing the
role of temperature, T ,

p = Rd ρTV . (5.40)

Based on this equation of state, a physical interpretation and definition of virtual temper-
ature is the temperature which dry air would have if its pressure and density were equal to those
of moist air. Virtual temperature is larger than ambient temperature, specially in humid
environments.
It is convenient to define virtual potential temperature, θV , using Eq. 5.13 with θ and
T replaced by θV and TV , respectively,
κ κ
p0 100
 
θV = TV = TV . (5.41)
p p

The mixing ratio of water vapor in air is very small (< 0.04), so the specific heat capacity
is not significantly different from that of dry air. Therefore, the difference between θV
and θ, or TV and T is seldom more than 7 K and usually less than 1 K, but height profile
trends can be different, hiding valuable information. A good practice is to use virtual
temperatures, specially in the humid environments such as water surfaces, irrigated fields,
forests and jungles.

Example 5.6
Example: Potential virtual temperature height profiles.

a Obtain potential temperature vertical profile from the environmental temperature

in Table 5.1, but this time assume a moist atmosphere with RH= 50%.
Using Eq. 5.41 with θ0v = T0v , results are displayed in Fig. 5.7
b Same as above but now RH=70%.
At a first glance, the potential temperatures profile in Fig. 5.7 indicates a stable
or subadiabatic atmosphere at the top, something that might be missed while
glancing through the environmental temperatures profile.
c Comment on profile differences for different RH values.
 5.5 SATURATED ATMOSPHERIC BOUNDARY LAYER 151

5.5 Saturated Atmospheric Boundary Layer

In a saturated atmosphere part of the water vapor may condense. In this case latent
heat of condensation is released into the air parcel. If condensed water is trapped in an
upward-moving parcel the corresponding saturated lapse rate must be lower than the
dry adiabatic case: the rate at which the parcel cools as it expands is lower because of
the addition of latent heat of condensation. If there is no exchange of heat between the
parcel and the surrounding environment, the process is called moist adiabatic. In practice
this does not occur since part of the condensed material may fall out, and the process is
not fully adiabatic since there is an exchange of heat with the surroundings. Instead the
process is almost adiabatic or pseudoadiabatic. The lapse rate between the adiabatic and
the pseudoadiabatic cases differ substantially, even though the amount of mass lost might
be small.
The moist adiabatic or saturated lapse rate is given by,

∂T dQs −1
Γs = −( )s = g(cp + Le ) (5.42)
∂z dT
dQs
where dT is he slope of saturation - specific humidity versus temperature, given by the
equation,

es mw Le 1 1
ln = ( − ) (5.43)
6.11 R∗ 273.2 T
in which the saturation vapor pressure es is in millibars. The above equation is obtained
from the Clausius-Clapeyron equation

des mw Le dT
= (5.44)
es R∗ T 2
and the fact that at T = 273.2 es = 6.11 milibar.
As shown in Eq. 5.42, the saturated lapse rate depends heavily on temperature. For
example, at T = 273.2 K , Γs = 0.0069 km−1 and at T = 303 K , Γs = 0.0036 km−1 , with
both cases at p = 1000 milibar.

5.6 Local and Nonlocal Static Atmospheric Stability

The concept of atmospheric stability is central when describing and predetermine pollu-
tion conditions. Stability can be exemplified by considering a frictionless ball in a peak,
a valley, a straight surface, or on an undulating surface as shown in Fig. 5.8. In the first
case we have instability, since when the ball is infinitesimally moved, it is accelerated
152 Chapter 5. THE ATMOSPHERIC BOUNDARY LAYER

Figure 5.8. Stability types represented by a ball resting on a smooth surface.

away from the peak. When initially is placed in the bottom of a valley and is infinites-
imally moved to either side, it tends to come back and stays near to its initial position,
thus exemplifying stability. In the case of the straight surface, when the frictionless ball
is moved to either side, the ball keeps moving at a constant velocity away from its initial
location. This is the neutral stability case. The fourth case is called metastability, which
corresponds to instability if the ball travels a long enough distance on either side.
If in the atmosphere a parcel that is moved adiabatically in the vertical and tends to
come back to its equilibrium point, that portion of the atmosphere is stable. If moved
vertically and it accelerates away, the corresponding layer of the atmosphere is unstable.
If moved and it keeps traveling vertically at constant speed, the corresponding layer is
neutrally stable.
The force in charge of returning, moving away or keeping the parcel moving is the
buoyancy force or acceleration, ab [m s−2 ]. It is given by the Archimedes principle,
!
ρ − ρp
ab = g , (5.45)
ρp
in which ρ is the density of atmosphere and ρp is the density of a parcel. The coordinates
origin is on the surface with positive acceleration upwards. If the parcel’s density is lower
than the surrounding air, the parcel accelerates upward, and viceversa. Think of a balloon
filled with a light gas such as Helium.
Using the equation of state of moist air, Eq. 5.45 can be written as

TV − TV p
 
ab = −g . (5.46)
TV
This means that a parcel accelerates upwards if its density is lowered by means of elevating
its temperature with respect to the surrounding atmosphere. Think of a hot air balloon.
 5.6 LOCAL AND NONLOCAL STATIC ATMOSPHERIC STABILITY 153

5.6.1 Local static stability

By expanding the Taylor series of TV and TV p around an equilibrium value TV e and
substituting into Eq. 5.46 we obtain,

g ∂TV ∂TV p
 
ab ∼
=− − ∆z . (5.47)
TV ∂z ∂z
This equation is an approximation since we got rid of higher order terms from the Taylor
expansion. The term in the parenthesis is the difference between environment and parcels
change of temperature with height z, determining the sign of acceleration ab . A dry air
parcel that is pushed from its equilibrium position always moves along the dry adiabatic
lapse rate Γ. This is sketched in Fig. 5.9. Then,

Figure 5.9. Comparison of parcel and environmental temperature profiles when

test parcel is moved adiabatically. In (a) T < Tp so ρ > ρp , parcel is lighter than the
environment and continues to rise. In (b),T > Tp so ρ < ρp , parcel is heavier than
environment and sinks back to original position.

g ∂TV g ∂ΘV
   
ab ∼
=− + Γ ∆z = − ∆z (5.48)
TV ∂z TV ∂z
If we define stability s [s−2 ] as
g ∂ΘV
 
s= , (5.49)
TV ∂z
local static stability can be divided in the three categories according to the sign of s as
shown in Table 5.2. If s < 0, the atmosphere is unstable and up or down movements are
enhanced since ab > 0. If s > 0, ab < 0 and parcel movements are suppressed denoting
154 Chapter 5. THE ATMOSPHERIC BOUNDARY LAYER

a stable atmosphere. If s = 0, ab = 0 and parcel travels conserving its initial velocity,

indicating a neutral atmosphere. The explanation of local stability and instability are
sketched in Fig. 5.9.

Table 5.2. Static atmospheric stability categories

∂ΘV ∂TV
Unstable s<0 ∂z <0 ∂z < −Γ
∂ΘV ∂TV
Neutral s=0 ∂z =0 ∂z = −Γ
∂ΘV ∂TV
Stable s>0 ∂z >0 ∂z > −Γ

Performing a similar adiabatic comparison between existing or environmental virtual

temperature lapse rate γ of an atmospheric layer and Γ as done in Fig. 5.9, atmospheric
layers can be classified as follows:
1. Superadiabatic, when γ > Γ.
2. Adiabatic, when γ = Γ.
3. Subadiabatic, when 0 < γ < Γ.
∂TV
4. Isothermal, when ∂z = 0.
∂TV
5. Inversion, when ∂z > 0.
These stability categories are shown schematically in Fig. 5.10. So far, our discussion
has been based on that the temperature profiles of the atmosphere are locally linear. In
practice, a broader stability concept is needed, since parcels travel distances such that
locality cannot be considered practical. Also observations of the convective boundary
layer where ∂ΘV ∼
∂z = 0 or slightly positive, instead of behaving like a neutral or slightly
stable layer, has all the attributes of an unstable boundary layer. Therefore s in practice is
not a good measure of stability for parcels that are performing large displacements from
their equilibrium position.

Figure 5.10. Super, sub and adiabatic atmospheric layers depiction.

 5.6 LOCAL AND NONLOCAL STATIC ATMOSPHERIC STABILITY 155

Example 5.7
Example: Local stability analysis using a temperature profile.

a Provide a local stability analysis on the five points shown in the graph of height
vs registered temperatures shown in Fig. 5.11obtained during an experiment per-
formed on a dry atmosphere.
First, dry adiabats are drawn in the graph at each point, as shown in Fig. 5.11.
Based on local stability analysis in each point, see Fig. 5.9 we have that: point 1
is stable, 2 is neutral, 3 unstable, 4 neutral and 5 stable.
b What would be the procedure if the atmosphere cannot be considered dry, and
Q is known?
c How much would the analysis with high moisture content change with respect
the dry atmosphere supposition?

Figure 5.11. Temperature versus height registered temperatures.

5.6.2 Nonlocal static stability

To determine the stability of an atmospheric layer, an air parcel is displaced up and down
adiabatically from possible starting points. The parcel buoyancy at any level is then deter-
156 Chapter 5. THE ATMOSPHERIC BOUNDARY LAYER

mined by the difference in virtual potential temperature of the parcel and the environment
at any level, as done in the local stability analysis. Nevertheless, now the displaced parcel
would continue to move farther away from its starting point and is tracked to the level
where becomes neutrally buoyant ΘV p = ΘV . Once these exercise is performed for all
relevant layers, their stability is determined following the following criteria:
1. Unstable: Regions where parcels enter and transit under their own buoyancy. Parcels
may not traverse the whole unstable region.
2. Stable: Subadiabatic regions that are not unstable.
3. Neutral: Regions where the adiabatic lapse rate are not stable.
4. Unknown: Top or bottom portions apparently stable or neutral, that do not end on a
material surface, such as ground surface, strong inversion, or tropopause: above or
below a layer that may be cold or warm, respectively, that could provide a source of
buoyant air parcels.
To illustrate the procedure refer to Fig. 5.12. Now we use potential temperature ab-
scissa θ, simplifying analysis since the dry adiabatic lapse rate is given by the vertical.
Note that in cases, especially those with unstable regions, there is a difference from the
traditional characterization of stability based on the value of s. The nonlocal charac-
terization is more consistent with empirical evidence. It is clear now that unstable and
convective boundary layers have strong and efficient mixing, while stable boundary layers
have weak mixing.

Figure 5.12. Nonlocal stability characterization for the various hypothetical virtual
potential temperature profiles.
 5.7 PASQUILL STABILITY CLASSES 157

5.7 Pasquill Stability Classes

A practical scheme to determine atmospheric stability was proposed by Frank Pasquill in

1961, [5]. The original purpose was to use it as part of a Gaussian dispersion model, a
subject to be covered in Chapter 7. Nevertheless, this stability criteria has been applied
to other areas such as high rise building design, [6]. As will be shown, the Pasquill sta-
bility classes rely only on qualitative-quantitative radiation observations and wind speed
data, providing a practical and easy to implement procedure. Kahl et al and Chapman
recommend not to use it as a stand alone procedure. Other schemes such as Richardson
and Monin-Obukhov will be presented in Chapter 6, Section 6.4.
The seven Pasquill stability classes are:
Class A Moderately unstable.
Class B Slightly unstable.
Class C Slightly stable.
Class D Neutrally stable.
Class E Moderately stable.
Class F Moderately stable.
Class G Strongly stable.
The neutral or adiabatic condition Γe =-1 K/ 100 m, corresponds to class D. The worst
case pollution scenario takes place when strongly stable G class occurs.
The categorization based on radiation and wind speed of these indexes is explained
in Table 5.3. In ref Wark et al is suggested that these indexes should be used with the
following comments in mind:
1. Strong insolation refers to clear skies, with solar angle greater than 60 degrees above
horizon, typical of a sunny afternoon. Very convective atmosphere.
2. Slight insolation refers to a sunny fall afternoon, or summer day with clear skies and
solar angle only 15 to 35 degrees above horizon.
3. Cloudiness will decrease the incoming solar radiation and should be used in con-
junction with solar angle to determine the solar radiation. For example, strong solar
radiation on a clear day would be reduced to moderate insolation with broken (5/8
to 7/8) cloud cover with middle clouds and to slight insolation with low clouds.
4. Night refers to the period between one hour before sunset and one hour after sunrise.
No stable conditions appear during day.
In general terms, Table 5.3 shows that during daytime, as wind speed decreases and
the solar radiation increases, the Pasquill stability class tends to become unstable. At
nighttime, as wind speed decreases and negative net radiation increases, the Pasquill sta-
158 Chapter 5. THE ATMOSPHERIC BOUNDARY LAYER

Table 5.3. Meteorological conditions defining Pasquill stability classes.

Daytime insolation Night-time conditions

Surface wind speed Strong Moder- Slight Thin overcast or < = 4/8
(m/s) ate > 4/8 low cloud cloudiness
<2 A A–B B E F
2–3 A–B B C E F
3–5 B B–C C D E
5–6 C C–D D D D
>6 C D D D D

Source: Pasquill, 1961.

NOTES:
1. Strong insolation corresponds to sunny midday in midsummer in England; slight insolation to similar condi-
tions in midwinter.
2. Night refers the period from 1 hour before sunset to 1 hour after sunrise.
3. The neutral category D should also be used, regardless of wind speed, for overcast conditions during day or
night and for any sky conditions during the hour preceding or following night as defined above.

bility class tends to become stable. Bias towards neutral stability class D is reported in
[7]: neutral conditions are reported when in fact ”stable” conditions exists. On average,
the Pasquill classes were found to stray from the true stability by one class. These con-
clusions were reached when Pasquill stability was compared with temperature profile
analysis carried out in several locations. Table 5.4 compares Pasquill classes and existing
atmospheric stability Γx .

Table 5.4. Comparison of Pasquill classes and Γx . Based on ref Wark.

Pasquill Classes Γx (C/100m)

A ≤ -1.9
B > -1.9 ≤ -1.7
C > -1.7 ≤ -1.5
D > -1.5 ≤ -0.5
E > -0.5 ≤ 1.5
F > 1.5 ≤ 4.0
G > 4.0
 5.8 POLLUTION, TROPOSPHERIC LAYERS AND INVERSIONS 159

Table 5.5. Key to Pasquill stability classes adapted for Japan. In construction

Solar Radiation [W m− 2] (Daytime) Net radiation[W m− 2] (Nightime)

Mean wind Strong Moder- Slight Thin overcast or < = 4/8
speed (m/s) at ate > 4/8 low cloud cloudiness
10 m
<2 A A–B B E F
2–3 A–B B C E F
3–4 B B–C C D E
4–6 C C–D D D D
>6 C D D D D

Source: Nakano, et al., 2006.

NOTES:
1. Strong insolation corresponds to sunny midday in midsummer in England; slight insolation to similar condi-
tions in midwinter.
2. Night refers the period from 1 hour before sunset to 1 hour after sunrise.
3. The neutral category D should also be used, regardless of wind speed, for overcast conditions during day or
night and for any sky conditions during the hour preceding or following night as defined above.

Another version of Pasquill classes developed for Japan is shown in Table 5.5. The
Pasquill categorization is refined: The day radiation intervals are now four instead of
three, and night intervals are now three instead of two. The most stable class G now
appears in the categorization table. This implementation requires objective radiation and
sensible heat measurements. Dominance of the neutral stability class D is apparent. This
fact is justified with measurements by Nakajima et al in [6]. The Pasquill stability indexes
were obtained mainly for rural and suburban areas, not taking into account the amount
of sensible heat fluxes present in urban areas.

5.8 Pollution, Tropospheric Layers and Inversions

In moderate to extremely unstable conditions normally encountered during midday and

afternoon, the mixing within the ABL is intense enough to cause conservable scalars such
as Θ or ΘV to be uniformly distributed with height. This is the case of a mixed layer. It
is an adiabatic layer overlying a superadiabatic surface layer. They can occur sporadically
or intermittently in the free atmosphere in the presence of turbulence.
To a lesser extent pollutants emitted locally and humidity have also uniform distri-
bution with height in a mixed layer. As a matter of fact a mixed layer can be visualized
using the reflectivity properties of particulate matter (pollution) in a LIDAR, as shown in
160 Chapter 5. THE ATMOSPHERIC BOUNDARY LAYER

Fig. 5.13. The presence of aerosols can be detected by the backscatter (darker layer). As
aerosols are mainly emitted at the surface, their concentration is generally higher in the
mixing layer than in the free troposphere. Therefore the mixing layer height estimation
by LIDAR is based on the detection of the sharp decrease (gradient) in aerosol backscat-
ter at the top of the mixing layer. This decrease marks the interface between the aerosol
containing mixing layer and the relatively clean free troposphere, [8].

Figure 5.13. Average diurnal variation of backscattered signal (2011–2016) in Mex-

ico City. White line indicates mixing layer height according to proposed automatic
algorithm. From [9].

The term inversion is used when the potential virtual temperature ΘV increases with
height. The strength of an inversion is given by the difference across a given depth of the
inversion layer. Inversions are classified according to their location (surface or elevated),
time ( nocturnal), and mechanism of formation (radiation, evaporation, advection, subsi-
dence, etc). Knowledge of an inversion height is very important in atmospheric pollution
since it provides the dilution degree of pollutants as will be demonstrated below.
The residual layer as presented in Section 5.2, is important for air quality. It is formed
when the turbulence of a well mixed layer starts to decay from above before sunset. The
mixing layer leaves behind the residual layer whose initial mean state variables such as
temperature and humidity are similar to the decaying mixing layer. In the residual layer
passive pollutants disperse equally in every direction, if no wind is present. Passive
pollutants may react during the night. When a new mixed layer ”mixes in” during the
next morning, it encounters pollutants not previously emitted from the surface. This ”mix
 5.8 POLLUTION, TROPOSPHERIC LAYERS AND INVERSIONS 161

in” phenomenon is called entrainment. During the day the pollutants formed during the
night may reach the surface by turbulent mixing.
The residual layer never touches the ground, so it is never affected by the surface
directly. Therefore it is not considered strictly as a boundary layer.

5.8.1 Vertical Temp and humidity profiles

Dry Atmospheric Boundary Layer
A typical diurnal sequence of observed radiosonde potential temperature at 3 hour inter-
vals is shown in Fig. 5.14. Before sunrise, when surface temperature is near a minimum,
the vertical profile is characterized by a nocturnal inversion. This is because of radia-
tive cooling. During this time a stable stratified ABL is present, so vertical turbulent
exchanges are suppressed. Under these conditions we have the case of a stable boundary
layer (SBL). Shortly after sunrise, surface heating leads to upward exchange of heat and
subsequent warming of the lowest layer due to heat flux convergence. This process starts
substituting the nocturnal inversion from below, and an unstable convective boundary
layer (CBL) with depth h starts to build up.
The CBL depth h grows with at a maximum rate a few hours after sunrise and slows
considerably in the late afternoon. We can distinguish three layers during the daytime
CBL: In the shallow surface layer, instability leads to turbulence by buoyancy and wind
shear. The layer on top is neutral and if nonlocal stability analysis is carried out, it will
be found that these two layers are mixed. These two layers comprise the bulk of the CBL.
Therefore, after morning hours, the daytime CBL is dominated by buoyancy generated
turbulence.These layers are capped by a stable layer, where turbulence (mixing) decreases
with height. Although we may consider the height zi of the inversion the same as that of
the mixing layer h, the latter may be 10 - 30 percent lower.
Just before sunset, net radiative loss of energy from the surface leads to a surface
inversion. This inversion deepens through the night due to radiative and sensible heat
divergence. The surface inversion height hi may not correspond to mixing layer height
h of the SBL because of significant turbulence activity. The ABL height can be better
estimated from temperature soundings depicted in panel c of Fig. 5.14 . Using advanced
measuring techniques the ABL it can be seen that height h decreases with time during
early evening hours, until it attains a constant value of less than hi .
Regarding humidity variations with height, if there is no significant vegetation and
therefore evapo-transpiration, the specific humidity is nearly constant in the mixed layer.
This is shown in Fig. 5.15. Only in the morning there is more humidity in the surface due
to evaporation of water deposited in the ground and or evapo-transpiration because of
the few vegetation present.
162 Chapter 5. THE ATMOSPHERIC BOUNDARY LAYER

Figure 5.14. Diurnal variations of potential temperature profiles and inversion

heights. In (a) day 33, and in (b) for days 33-34 Wangara experiment. In (c), curve
(A) is the inversion heights for the CBL. Curve (B) is the surface inversion height in
nighttime SBL.

Figure 5.15. Diurnal variations of specific humidity profiles during day 33 Wangara
experiment
 5.9 A SIMPLIFIED POLLUTION BOX MODEL 163

Humid Atmospheric Boundary Layer

If the surface is populated by vegetation, the surface layer humidity will be affected by
plant physiology, especially by the Krebs cycle. During sunrise, when plant respiration
turns to transpiration, there is an important water vapor emission that can last for 10 to 20
minutes. It can severely influence the temperature and humidity profile near the surface
during this time.
In Fig. 5.16 is shown the modeled vertical profiles of virtual potential temperature and
specific humidity during the early morning hours. They correspond to a computational
experiment of an urban area with non-vegetated roofs ( 10- 15 m high) and the same
roof area vegetated with siempre viva plants. During sunrise, when the plants humidity
emissions takes place in the vegetated case, cooling at the height of the roof takes place.
The vertical temperature profile is thus changed with respect the non-vegetated case. In-
versions at lower roof height are induced. Also vegetated roof night temperatures slightly
increase with respect the non-vegetated case before the humidity emission at sunrise takes
place, [10].

5.9 A Simplified Pollution Box Model

With the concepts studied in this chapter, we are in a position to demonstrate the relation
between mixing height, air velocity, emissions and pollution concentration. We follow
in part the corresponding material in [11]. Consider the simplified box pollution model
shown in Fig. 5.17. Assume that the box is placed over a city, where we want to obtain a
preliminary pollution assessment. The model assumes that u[m s−1 ] is a constant wind
speed aligned with x dimension, Q [kg s−1 ] is mass emission rate of a pollutant c, H(t)[m]
is mixing height, and ∆x,∆y and ∆z are the geometrical dimensions of the box in [m].
Consider the case of an inert gas (no chemical reactions) whose concentration is rep-
resented by c[kg m−3 ], and assume no removal mechanisms. Then, a mass balance inside
the box yields:

d
(c∆x∆yH) = Q + uH∆y (c(0) − c) , (5.50)
dt
where c(0) is the initial concentration homogeneously distributed insde the mixing layer.
The terms on the right represent the changes in concentration of c(t) resulting from emis-
sions and advection. The term on the left, assumes immediate and homogeneous mixing.
Dividing by ∆y and ∆z,

d H
(cH) = q + u (c(0) − c) , (5.51)
dt ∆x
164 Chapter 5. THE ATMOSPHERIC BOUNDARY LAYER

Figure 5.16. Measured, (a) virtual potential temperatures and (b) RHs, with solid
line for green and dashed line for conventional roof simulations. From [10]

Figure 5.17. A basic box model for pollution.

 5.9 A SIMPLIFIED POLLUTION BOX MODEL 165

where q [kg m−2 s−1 ]is is emissions per unit area. Assuming that for a certain time frame
H can be considered constant,
dc q u
= + (c(0) − c) . (5.52)
dt H ∆x
The units of the term u
∆x is in s−1 and its inverse represents the residence time of c inside
the box, i.e., the time required by the airshed to clear the box, after the emissions are
turned off. We we call it τ ,
∆x
τ= . (5.53)
u
Therefore,
dc q 1
= + (c(0) − c) . (5.54)
dt H τ
This differential equation has the following analytic solution,

qτ
  
−t/τ
c(t) = c(0)e + + c(0) 1 − e−t/τ . (5.55)
H
This equation describes the evolution of c due to an initial concentration that decays expo-
nentially with characteristic time τ , plus the contribution of emissions and the advection
by wind. We have at our disposal a simple but informative model, where we can objec-
tively test the cause and effect that variables such as mixing height, residence time and
dilution have on air pollution.
166 Chapter 5. THE ATMOSPHERIC BOUNDARY LAYER

Example 5.8
Example: Pollution scenario using box model.

a Obtain the CO concentration in a city. The emission is at a rate q = 200 µg m−2

h−1 . Consider a background concentration of c(0) =1 µg m−3 . Wind speed is
u = 2 m s−1 . The city is 50 × 50 km, and mixing height is 500 m.
Using Eq. 5.55 and using a computational program, the solution is shown in
Fig. 5.18. Make sure you are using congruent units.

Figure 5.18. Box model CO concentrations. Steady state takes about

21 hours to be reached.

dc
b Same as above but obtain the steady state concentration value, i.e., dt = 0.
The steady state equation for c becomes,
qτ
c= + c(0) . (5.56)
H
Then steady state concentration is c = 3.8 µg m−3 . From Fig. 5.18 the steady state
concentration is reached after 21 hours. This concentration is about 4 times the
background concentration.
c Consider steady state solution, but now H = 1 000 m. Comment on the concentra-
tion-dilution effect caused by the change in mixing height. What features would
you add to the analysis so it becomes more realistic?
 5.9 A SIMPLIFIED POLLUTION BOX MODEL 167

The box model sheds light on the the effect of entrainment. Consider the case when
H is an increasing function of time, say linear:

H(t) = at + H0 , (5.57)

where a[m s] is the velocity at which the mixing height gains altitude and H0 is the
initial mixing height. Denote as ca the existing concentrations above the mixing layer to
be entrained, and c(t) the concentration in the mixing layer. Then, H increases its height
to H + ∆H after a time ∆t, while concentration changes to c + ∆c. Performing mass
balance,
(c + ∆c) (H + ∆H) = cH + ca ∆H (5.58)

Carrying out the products in the left hand side, neglecting second-order term ∆c∆H in
the right hand side and identifying terms,

H∆c = (ca − c) ∆H . (5.59)

Rearranging variables, dividing by ∆t and taking limit ∆t → 0,

dc ca − c dH
 
= . (5.60)
dt H dt

Expanding this equation, substituting H(t) by at+H0 , the following first order differential
equation is obtained,
dc c ca
+ = , (5.61)
dt t + h1 t + h1
H0
where h1 = a . Multiplying Eq. 5.61 by the integrating factor (t + h1 ), the following
solution is obtained,
t h1
c(t) = ca + c(0) , (5.62)
t + h1 t + h1
where remember that c(0) is the concentration inside the mixing layer when the experi-
ment begins, i.e., at t =0. Then, initially the concentration in the mixing layer is c(0), and
as time elapses c(t) is asymptotic to ca .
Theretofore, considering the box model suppositions such as immediate and homo-
geneous mixing, if we just observe the evolution of initial conditions above and below
the initial mixing layer, as mixing height grows the value of the concentration left the day
before, eventually dominates the pollution.
168 Chapter 5. THE ATMOSPHERIC BOUNDARY LAYER

Example 5.9
Example: Entrainment scenario using box model.

a Implementing Eq. 5.62 in a computational program, find the evolution of the

concentrations of a non-reactant pollutant when the initial mixing height is 500
m, and its top climbs at a speed of 500 m per hour. At the beginning of the
experiment the concentration in the mixing layer is 2 µgm−3 and above is 8µgm−3 .
Perform the experiment for 3 hours.
Figure 5.19 (a) shows the concentration evolution. When mixing height reaches
2 000 m, concentrations approach 8 µgm−3 , the initial concentrations found in
the entertainment zone. Pollution is increased inside the mixing layer.
b Same as above, but at the beginning of the experiment the concentration in the
mixing layer is 8 µgm−3 and above is 2 µgm−3 .
Figure 5.19 (b) shows the concentration evolution. When mixing height reaches
2 000 m, concentrations approach 2 µgm−3 , the initial concentrations found in
the entertainment zone. Pollution is decreased inside the mixing layer.

Figure 5.19. Entrainment concentrations for linearly developing mixing height

after 3 hours. In (a) when ca =8 and cb =2. In (b) whenca =2 and cb =8.

c Repeat above examples considering different values of a constant pollutant

source Q.
 EXERCISES 169

Exercises

5.1 Equation Example

Show that also

ρ T 1
= ( ) (γ−1) (5.63)
ρ0 T0
(Answer on Page 171)

5.2 Mathematical Proof

∂Θ
Proof mathematically that ∂z = 0.
(Answer on Page 171)

5.3 Math Samples

Using values in table below, obtain:
a) Virtual temperature at each level and its difference form actual environment tem-
perature.
b) Compare gradients of potential temperature and virtual temperature between
adjacent measurement levels.
c) The local static stability parameter, s and the percentage contribution of specific
humidity gradient to the same at different height levels.

Table 5.6. Measurements on a research vessel

Height above sea-surface: 2.1 3.6 6.4 11.4

Air Temperature (C) 27.4 27.72 27.69 27.62
Specific humidity (g kg −1 ) 18.02 17.83 17.54 17.29

(Answer on Page 171)

5.4 Data from radiosonde

Using data below obtained using a radiosonde, obtain and plot

170 Chapter 5. THE ATMOSPHERIC BOUNDARY LAYER

Table 5.7. Radiosonde measurements

Height (m) Pressure (mb) Temperature (C) Specific Humidity

(g kg −1 )
1.5 (Surface) 1013 11.8 6.5
50 1007 10.9 6.1
100 1001 10.1 5.8
150 995 9.6 5.6
200 989 9.2 5.5
300 977 8.4 5.3
400 965 7.4 5.1
500 954 6.4 4.9
550 948 6.0 4.9
600 942 5.6 4.8
650 936 5.1 4.7
700 931 4.9 4.6
750 925 4.9 4.5
800 919 4.9 4.5
850 914 5.0 4.5
900 908 5.0 4.5
1000 897 4.8 4.4

a) Virtual temperature as a function of height.

b) Virtual potential temperature as a function of height .
c) Characterize various layers on the basis of local and non-local stability
d) estimate height of inversion layer and ABL height.
(Answer on Page 171)
 EXERCISES 171

Answers to Exercises

Answer to Exercise 5.1 (page 169)

Answer for exercise 5.1.

Answer to Exercise 5.2 (page 169)

Answer for exercise 5.2.

Answer to Exercise 5.3 (page 169)

Answer for exercise 5.3.

Answer to Exercise 5.4 (page 169)

Answer for exercise 5.4.
172 Chapter Bibliography

Chapter Bibliography

[1] Roland S. Stull An Introduction to Boundary Layer Meteorology. Kluwer Academic Publishers,
2000. ISBN 90-277-2768-6
[2] J.C. Kaimal and J.J. Finnigan Atmospheric Boundary Layer Flows: Their Structure and Measure-
ment. Oxford University Press, 1994 ISBN 0-19-506239-6
[3] L. T. Molina, S. Madronich, J. S. Gaffney, E. Apel, B. de Foy, J. Fast, R. Ferrare, S. Herndon, J.
L. Jimenez,B. Lamb, A. R. Osornio-Vargas, P. Russell, J. J. Schauer, P. S. Stevens, R. Volkamer,
and M. Zavala. An overview of the MILAGRO 2006 Campaign: Mexico City emissions and
their transport and transformation. Atmos. Chem. Phys., 10, 8697-8760, (2010). www.atmos-
chem-phys.net/10/8697/2010/ doi:10.5194/acp-10-8697-2010
[4] Holstlag, A.A.M. and A.P. van Ulden. A simple scheme for daytime estimates for the surface
fluxes from routine weather data. J. Climate Appl. Meteor. 22, 517-529, (1983).
[5] Pasquill, F., 1961. The estimation of dispersion of windborne material. Meteorol. Mag. 90,33-49.
(1961).
[6] Keigo Nakajima, Toru Yamanaka, Ryozo Ooka, Hideki Kikumoto, Hirofumi Sugawara, Ob-
servational assessment of applicability of Pasquill stability class in urban areas for detec-
tion of neutrally stratified wind profiles, Journal of Wind Engineering and Industrial Aerody-
namics,Volume 206,2020. 104337,ISSN 0167-6105. https://doi.org/10.1016/j.jweia.2020.104337.
https://www.sciencedirect.com/science/article/pii/S0167610520302476)
[7] Kahl, Jonathan, Chapman, Hillary Atmospheric stability characterization using the Pasquill
method: A critical evaluation Atmospheric Environment 187, (2018). DOI 10.1016/j.at-
mosenv.2018.05.058
[8] Emeis, S., C. Munkel, S. Vogt, W.J. Muller, and K. Schafer. Atmospheric boundarylayer struc-
ture from simultaneous SODAR, RASS, and ceilometer measurements. Atmos. Environ., 38,
pp. 273-286. (2004).
[9] J. L. Garcia-Franco, W. Stremme, A. Bezanilla, A. Ruiz-Angulo, M. Grutter. Vari-
ability of the Mixed-Layer Height Over Mexico City Boundary-Layer Meteorol, 2018.
https://doi.org/10.1007/s10546-018-0334-x
[10] Morales, Williams,Jazcilevich, Aron, Garcia-Reynoso, A.,Caetano, Ernesto, Gomez,
Gabriela,Bornstein, Robert. Influence of Green Roofs on Early Morning Mixing Layer Depths
in Mexico City. Journal of Solar Energy Engineering,138 DOI:10.1115/1.4034807
[11] John H. Seinfeld and Spyros N. Pandis. Atmospheric Chemistry and Physics. John Wiley and
Sons, Inc.,Wiley Interscience, 1998. ISBN 0-471-17815-2
Chapter 6

Buoyancy and viscosity in the

Atmospheric Boundary Layer

6.1 A viscous atmosphere 175

6.2 Interaction between buoyancy and mean wind 176

6.3 Wind velocity profiles in near-neutral boundary layers 177

6.4 Wind profile for non-neutral atmospheres: Monin-Obukhov 183

6.5 The Obukhov length and the gradient Richardson number 185

6.6 Prediction of the mixing height 189

Exercises 193

Answers to Exercises 195

Chapter Bibliography 196

In Chapter 5, the effect of buoyancy and thermodynamics in the Atmospheric Boundary

Layer were introduced. These concepts enabled us to define different types of commonly
used stability definitions. Another important component of boundary layer theory is
the introduction of shear stress in the wind flow caused by the presence of a surface.
This shear effect interacts with buoyancy, providing additional mechanisms that must be
taken into account to define more precise stability and mixing potential criteria. This is
the subject of this chapter. Figure 6.1 provides a dramatic illustration of fluid phenomena
brought by the presence of shear forces.

– 173 –
174 Chapter 6. BUOYANCY AND VISCOSITY IN THE ATMOSPHERIC BOUNDARY LAYER

Figure 6.1. Scene in Panama City, Florida, February 5th, 2012. Cool air offshore
was nearly at the saturation point. The drop in temperature caused by a lift of
about 50 meters, saturated the wind parcels above the buildings forming the clouds.
On the back side, the air sinks back down and warms, thus becoming unsatu-
rated and the cloud disappears. The cloud formation provides a view of turbu-
lent flow on top of the buildings generated by shear forces. Credit to: Helicopter
pilot JR Hott. More photographs are displayed in https://blogs.agu.org/wildwild-
science/2012/02/06/condo-wave-clouds/
 6.1 A VISCOUS ATMOSPHERE 175

6.1 A viscous atmosphere

Consider the situation depicted in Fig. 6.2(a). This fluid phenomena shows that the
horizontal velocity increases from zero as we move upwards, ( ∂u
∂z > 0), until a velocity U
parallel to the surface is asymptotically reached. This phenomenon in the atmosphere is
observed experimentally. As explained in texts of fluid mechanics, this is a manifestation
of the interaction between the surface and the fluid moving at a velocity U parallel to
the surface. The subjacent physical mechanism is the presence of shear stress, τ . The
space between observed velocities until U is reached, is used to measure the width of the
boundary layer as defined in Chapter 5.

Figure 6.2. In (a), wind-speed profile and shear stress τ . In (b), stretching along prin-
cipal axis of an element in the fluid made up of ideal ”rectangle molecules”. Lengths
of sides and height of rectangle remain unchanged. Think of a deck of cards tossed
on a table.

In [1], an intuitive explanation of shear stress is presented. Consider an element of

the fluid as shown in Fig. 6.2(b). Focus on one “rectangle element”, S. Then, ”faster
elements from above S will diffuse across S and impart momentum to the fluid below,
and, likewise, slower elements above S will diffuse across S to slow down the fluid above
S”. Since diffusive forces between elements are present, kinetic energy is dissipated into
heat.
This explains the role played by shear stress, and therefore the shape acquired by the
wind-speed profile in Fig. 6.2(a). This elements interaction is called viscosity. Besides
deformation, the S elements rotate, giving rise to vorticity. These effects provide the
conditions for turbulence. Therefore, turbulent mixing is in operation in this type of
flow. A visualization of turbulent flow is shown in Fig. 6.3, where size of large eddies is l
and boundary layer thickness is Lt . Outer velocity is U and inner velocities are u. More
on turbulence will be discussed in Chapter 7.
176 Chapter 6. BUOYANCY AND VISCOSITY IN THE ATMOSPHERIC BOUNDARY LAYER

Figure 6.3. Eddies in a turbulent boundary layer. Flow above turbulent layer has
outside velocity U and the inner average velocity is u. The eddies have a length
scale l, about the size of largest eddies. The boundary thickness is Lt . The interface
between boundary inner layer and outer layer is sharp. Based on Tennekes, p12.

If there is no momentum interaction between the rectangle elements as just described,

then the fluid is not-viscous. We could imagine this situation if “ideal skates” are placed
between layers of the rectangles in Fig. 6.2, so they perfectly slide and no momentum is
transferred. This is the case of laminar flow. In this case, instead of logarithmic profile
shape as shown in Fig. 6.3, the velocity is linear with z near the surface. In this case,
no turbulent mixing operates, and a smooth flow is expected. Since no viscous diffusive
forces are present, no transformation of kinetic energy into heat takes place.
The micrometeorology taking place near the Earth’s surface must be modeled con-
sidering a viscous fluid. Corresponding changes in wind speed, temperature, humidity
are studied in this chapter. The specific case for point sources emissions atmospheric
transport, will be studied in Chapter 7.

6.2 Interaction between buoyancy and mean wind

As described in Chapter 5, during the day the surface is warmer than air above in response
to solar heating. This gives rise to convective flows (vertical) such as surface plumes, up-
drafts and downdrafts. These transfer momentum and heat in the vertical direction, and
are responsible for convective or buoyancy-generated turbulence. The resulting vigor-
ous mixing of momentum leads to considerable weakening and, sometimes, elimination
of mean wind in the Atmospheric Boundary Layer, ABL. This is the reason why in the
 6.3 WIND VELOCITY PROFILES IN NEAR-NEUTRAL BOUNDARY LAYERS 177

convective boundary layer, or, CBL, it is common to find near uniform humidity and tem-
perature. For same reason, far enough from a source, pollutant concentrations tend to be
homogeneously distributed. Strong wind shears are found only near the surface and near
the transition layer, located in the base of an inversion, that often caps the mixed layer.
During the night the surface cools down in response to longwave radiation. Near sur-
face inversions develop, vertical momentum exchanges are inhibited, and a stable bound-
ary layer, SBL, is formed. Although not as strong as during daytime, nighttime surface
shears are developed, and low level jets appear.
The above description explains why, as shown in Fig. 6.4, after sunrise the wind speed
of upper and lower layers of the ABL approach. Nevertheless, the wind speed always
increase with height.

Figure 6.4. Diurnal variation of 40-day-averaged wind speeds at various heights in

the ABL during the Wangara Experiment Observed vertical profiles of mean wind
components and potential temperatures and calculated Ri in nocturnal ABL under
moderately stable conditions.

6.3 Wind velocity profiles in near-neutral boundary layers

Although precise neutral conditions are seldom encountered in the atmosphere, under
certain conditions such as overcast skies with strong winds, a quasi-neutral case may
be considered. As a matter of fact, quasi-neutral conditions are regularly attained. In
Chapter 5 when the Pasquill stability classes were introduced, class D prominently ap-
peared in the classification table. If this is the case, simple theoretical and semi-empirical
178 Chapter 6. BUOYANCY AND VISCOSITY IN THE ATMOSPHERIC BOUNDARY LAYER

approaches developed in fluid dynamics can be applied to micrometeorology. These

∂u
schemes characterize the vertical profile of u, ∂z , near the surface. They are used to
obtain the wind field, necessary to describe pollution, humidity and temperature trans-
port. Moreover, this near-neutral case will provide a basis to develop schemes when the
atmosphere cannot be considered neutral or near neutral, as will be shown in Section 6.5.

6.3.1 Power-law profile

Adapting work by the aerodynamicist Ludwig Prandtl for flat-plates, an empirical version
for the vertical profile of u in micrometeorology has been proposed:
m
u z

= (6.1)
ur zr
where ur is the wind speed at a reference height zr . Usually ur is chosen at zr = 10 m, a
height smaller or equal to the height of maximum wind velocity.
The power-law profile does not have a sound theoretical basis but, as shown in Fig. 6.5,
it provides reasonable fit with observed velocity profiles in the lower part of the ABL.
The exponent m depends on surface roughness and stability. Its values range form 0.1 for
smooth water, snow and ice surfaces, to about 0.4 for developed urban areas. Figure 6.5
also shows dependence of m with z0 , the roughness length parameter, which will be
defined soon.

Figure 6.5. Comparison of observed wind profiles and the power-law at different
sites, under stability conditions. Izumi and Caughey (1976)
 6.3 WIND VELOCITY PROFILES IN NEAR-NEUTRAL BOUNDARY LAYERS 179

6.3.2 Logarithmic-law profile

Consider the flow near a smooth surface and consider U its velocity outside the boundary
layer, as shown in Fig. 6.3. Near the surface, the velocity u is expected to depend on the
fluid density, ρ, the horizontal shear stress on the surface, τ , the viscosity of the fluid, ν,
τ
and height, z. Define u∗ = ρ as the friction velocity. Then,

u = u(ν, u∗ , z) . (6.2)

By recasting this equation in a dimensionless form we get,

u zu∗
 
=f , (6.3)
u∗ ν
where f is a function with a dimensionless argument. Therefore, we are in the liberty to
postulate a logarithmic function to play the role of f . The solution near the surface, or
inner layer, must overlap smoothly with the solution outside the boundary layer, or the
outer layer. For this to happen,
df (z + ) 1
z+ = , (6.4)
dz + κ
where z + = zuν ∗ , and κ=0.4 is the Von Karman constant. The choice of this constant
stems from experimental experience. Rearranging terms,
df (z + ) 1 1
= . (6.5)
dz + κ z+
Integrating this expression by separation of variables,
1
ln(z + ) + B0 ,
f (z + ) = (6.6)
κ
where B0 is an integration constant. Then so far,
u 1
= ln(z + ) + B0 . (6.7)
u∗ k
The appearance of natural logarithm function, ln, in this expression, agrees with the
the requirement that u merges smoothly with the outer flow velocity U . Furthermore, it
renders the expected shape of the velocity profile shown in Fig. 6.3. The procedure shown
here is an example of similarity analysis discussed in the fluid dynamics literature as in
[2]. This experience will become handy when developing the physics of velocity profiles
in non-neutral atmospheres in next section.
To obtain a boundary layer velocity profile useful in micrometeorology, a proper
z
choice for the dimensionless variable z + must be found. A convenient choice is z + = z0 .
The parameter z0 is called roughness length. Then, the logarithmic-law profile used in
micrometeorology is
u∗ z
 
u= ln +B . (6.8)
k z0
180 Chapter 6. BUOYANCY AND VISCOSITY IN THE ATMOSPHERIC BOUNDARY LAYER

Example 6.1
Example: Finding the wind speed profile from field data.

a Obtain the wind speed profile for the following wind speed data collected in the
field. Wind data table from [3], p.196. .

z(m) 0.5 1 2 4 8 16
u(ms−1 ) 7.82 8.66 9.54 10.33 11.22 12.01

The wind profile is shown in Fig. 6.6.

Figure 6.6. In (a), wind velocity profiles of data and fitted logarithmic-law. In
(b), profile and its extrapolation to show z0 .

b Obtain the corresponding fitted data using the logarithmic-law.

The data points in graph height versus velocity are fitted by ln function using a
computational program, as shown in Fig. 6.6(a). The fitted function is

z
 
u = 1.213 × ln + 8.67 . (6.9)
z0
−8.67
The extrapolated value at the abscissa for u = 0 is z0 = e 1.213 = 7.9 × 10−4 m.
Using κ = 0.4, u∗ = 0.485 ms−1 . This is shown Fig. 6.6(b).
Based on the values obtained for z0 , the data was collected on a short grass leveled
surface.
κ
c Draw the wind speed profile in ln graph. Reinterpret the term u∗ and z0 in this
plane. Hint: ln(z) = au−ln(z0 ). The logarithmic rule becomes a line with slope a.
 6.3 WIND VELOCITY PROFILES IN NEAR-NEUTRAL BOUNDARY LAYERS 181

It is important to consider the integration constant B to obtain correct results. Math-

z
ematically, z0 scales z in the expression z + = z0 . It provides control where to place the
origin of the velocity profile in the abscissa. Physically, the choice of z0 varies the shape
of the velocity profile and is the height at which the wind speed profile extrapolates to
zero under neutral conditions. It allows the application of the logarithmic-law profile for
different surfaces shown in in the table in Fig. 6.7. The previous example illustrated these
concepts.

Figure 6.7. Typical values of z0 for different types of terrain. From [3], p.199.
182 Chapter 6. BUOYANCY AND VISCOSITY IN THE ATMOSPHERIC BOUNDARY LAYER

For very rough or undulating surfaces the logarithm-law is displaced vertically, as

shown in Fig. 6.8. This called displacement heigth. Inside the canopy a second wind
maximum is found, topped by a an exponential region just before the canopy height.

Figure 6.8. Displacement height representation due to vegetation canopy. The flow
over forest canopy wind speed, u, as a function of height, z. The canopy of the trees
act as a surface displaced a distance d0 , above the surface. The roughness length,
z0 , lies within the tree tops height.

The adapted formula has the form:

u∗ z − d0
 
u= ln +B , (6.10)
k z0
where d0 is called the zero plane displacement or displacement height. The values that
d0 can take, can be from 0 to the mixing height, h0 . A relationship between zero-plane
and average vegetation is shown in Fig. 6.9
 6.4 WIND PROFILE FOR NON-NEUTRAL ATMOSPHERES: MONIN-OBUKHOV 183

6.4 Wind profile for non-neutral atmospheres: Monin-Obukhov

The preceding wind profile formulas presuppose neutral conditions: buoyant forces are in
equilibrium and therefore there is no vertical momentum due to convection. Parcels are in
thermal equilibrium and temperature profile coincides with Γ, the vertical adiabatic lapse
rate. Since many times neutral or near-neutral conditions are used as a useful approxi-
mation, they play an important role. Nevertheless, in the physical world, this situation is
sustained only for brief periods of time. There are stable or unstable conditions that can-
not be considered near-neutral. These are of great importance for pollution studies and
other civil engineering applications, and must be studied to predict valid meteorological
information.
As we did in the near-neutral case,let us start by putting forward which parameters
our solution must be based on. The Russians Monin and Obukhov, M-O, in 1954, had
the physical insight of proposing the additional dimensionless arguments that must be
taken into account: surface temperature flux H0 /ρcp and the buoyancy parameter g/θ0 ,
[5]. Then,
u = u(l, u∗ , q, c0 , g/θ0 ) (6.11)

As we did in the logarithmic-law case, we can form an equality relating groups of

dimensionless variables via a function φm to be determined,

κz ∂u z
 
= φm , (6.12)
u∗ ∂z L
−3
where z
L is the independent variable in this formulation, and L = − uκgQ
∗ θ0
0
, a variable
with a length dimension, is called the Obukhov length. Remember that κ = 0.4 is the
von Karman constant. The subscript m in φm , denotes the momentum case. This step
is formally obtained by using the Buckingham Pi Theorem, see [6], p220. In the case
of the logarithmic-law, the nature of the function satisfying the equality was obtained
using von Karman constant, κ, to merge the inner and outer solutions. In this more
complex case where laboratory experiments are not possible, function φm is determined
via observations.
Accepted M-O φm forms for the stable and unstable cases, are:

 1.0 + 4.8 Lz for z
L ≥ 0(stable)
φm = z

−1 z
(6.13)
 1.0 − 19.3 5 for < 0(unstable)
L L

In Fig. 6.10, the fit of φm to data under very stable conditions is shown.
184 Chapter 6. BUOYANCY AND VISCOSITY IN THE ATMOSPHERIC BOUNDARY LAYER

Figure 6.10. The M-O φm function for mean wind shear and its fit to data is shown.
Data is from the Kansas experiment, [7], under very stable conditions.

Using same procedure, the M-O similarity analysis for mean potential temperature pro-
duces,
kzu∗ ∂Θ z
 
− = φh , (6.14)
H0 ∂z L
and for mean vapor mixing ratio,

kzu∗ ∂C z
 
− = φc . (6.15)
C0 ∂z L

The corresponding accepted M-O forms, [8], for temperature profile are:

 1.0 + 7.8 Lz for z
L ≥ 0 (stable)
φh =
−1 (6.16)
 1.0 − 12 z 2 for z
< 0 (unstable)
L L

These φm and φh M-O functions are universal for locally homogeneous, steady surface
layers.
 6.5 THE OBUKHOV LENGTH AND THE GRADIENT RICHARDSON NUMBER 185

The Obukhov length L is negative when H0 > 0 or unstable conditions, positive in

stable conditions when H0 < 0, and infinite when H0 = 0 or neutral conditions. Therefore,
z z z
the range of L is −∞ < L < ∞. Conditions for near-neutral conditions are when L is
small, therefore when z  |L|, i.e., near the surface.
Once that we know the form of the velocity, temperature and humidity gradients, the
proposed formulas have to be integrated to obtain respective fields. We follow Wyngaard,
z
[6], p222. Consider first stable conditions, where the gradients are linear with L, and
therefore integration is straightforward. Then,
u∗ z z
 
u(z) = ln + 4.8 , (6.17)
κ z0 L
and

T∗ z z
 
Θ(z) = Θ(zr ) + ln + 7.8 , (6.18)
κ zr L
Q0
where T∗ = u∗ , and zr is a reference surface temperature.
For the unstable case, where integration is more difficult, the following expressions
are proposed by ref Wilson (1960). Consider,
 −1
z 2

2
φm,h = 1 + γ| | 3 , (6.19)
L
for both φm and φh . Then, the mean profiles are,
UNDER CONSTRUCTION
The following constants are suggested:κ = 0.4, Pt = 0.95, γm = 3.6, γh = 7.9.

6.5 The Obukhov length and the gradient Richardson number

g ∂Θv
In Chapter 5, the local stability index s = Tv ∂z was defined. It only considers tem-
perature and temperature gradients. Another stability index that also considers the ABL
dynamics in lieu of wind shear is necessary. The Obukhov length L provides such an
index. Writing again the expression for L,

−u3∗
L= ,
κ (g/Θ0 ) (H0 /ρcp )

we see that it is a ratio of a term containing the shear information in the numerator, and
a term containing the buoyancy information in the denominator. We briefly follow the
physical interpretation of L given in Arya, [3], p214. By definition, L can acquire values
that go from −∞ to +∞. The extreme values correspond to the heat flux approaching
186 Chapter 6. BUOYANCY AND VISCOSITY IN THE ATMOSPHERIC BOUNDARY LAYER

Figure 6.11. Obukhov’s length L as a of friction velocity, u∗ and surface flux H0 .

From [3], p215.

zero form the positive (unstable) and the negative (stable) side, respectively. A practical
range of |L|, corresponding to a wide range of values of u∗ and |H0 |, is shown in Fig. 6.11.
κg
Here, it is assumed that T0 = 0.013 m s−2 K−1 . The magnitude of L, |L|, represents the
thickness of the layer influenced by the Earth’s surface in which shear or friction effects
are important. Near the surface,z  |L|, wind shear effects dominate over buoyancy
effects. When z  |L|, buoyancy dominates over shear-generated turbulence. Therefore,
z
as mentioned before, the independent variable L measures the relative importance of
buoyancy versus shear effects in a stratified surface layer.
A similar and related parameter called the gradient Richardson number, Ri , is now
introduced. Its definition is,
g ∂Θ∂z
v
Ri =  2 . (6.20)
Tv ∂u
∂z

where u is the wind speed. Ri is dimensionless and same sign as the static stability
parameter s. Note that Ri is a ratio of the buoyancy force/mass provided by s, and
changes in velocity gradients/mass. The larger the gradients relative to buoyancy, the
smaller Ri and viceversa. A large positive value of Ri > 0.25 indicates weak, decaying
or non-existent shear (or turbulence) in environment. It is a better measure of mixing
 6.5 THE OBUKHOV LENGTH AND THE GRADIENT RICHARDSON NUMBER 187

intensity and provides a simple criterion for the existence or not existence of a stably
stratified environment.
A vertical distribution of Ri can be used to determine the extent of stratified layers
within the ABL. An example of wind speed, potential temperature and Ri vertical profile
is shown in Fig. 6.12.

Figure 6.12. Observed vertical profiles of mean wind components and potential
temperatures and calculated Ri in nocturnal ABL under moderately stable condi-
tions. From [3], p104.
188 Chapter 6. BUOYANCY AND VISCOSITY IN THE ATMOSPHERIC BOUNDARY LAYER

∂θ
It corresponds to moderately nocturnal stable conditions, since ∂z > 0. With θ information
alone, is difficult to localize the mixing layer height. The jump in the values of Ri > 0.4
gives away the location of the mixing layer height of the SBL at h ≈350 m. The mixing
layer is therefore objectively identified by Ri . Below h, dominance of shear (Ri < 0.25)
is found near the surface, and therefore turbulent mixing is clearly localized there. The
wind velocity magnitude, dominated by component u, follows that of Fig. 6.2.
With the Ri data provided, a description of this scene is as follows: Under moderately
stable conditions, vigorous turbulent mixing takes place near the surface up to height of
≈ 100 m. As we ascend, the turbulent mixing subsides, until the mixing layer height at
h ≈350 m is reached. Here, a laminar layer caps the mixing layer.

The gradient method

A useful technique called the gradient method provides valuable surface micrometeo-
rology information and is a good example of how Ri and M-O theory are related. The
linear finite-difference and logarithmic finite-difference approximations in Appendix B
are used. In [3], is shown that for surface layer meteorological applications, the logarith-
mic approximation is superior than the liner approximation, with acceptable errors of no
more than 8%.
Applying the logarithmic approximation to the velocity and potential temperature
gradients at a geometric mean height zm in Eq. 6.20, an approximation to Ri becomes,

g ∆Θzm z2
 
Ri = 2 ln . (6.21)
Tv (∆u) z1

Based on empirical considerations, the M-O forms are modified such that

−1
φh = φ2m = 1 − 15 Lz 2 , for z
L < 0(unstable)
(6.22)
φh = φm = 1 + 15 Lz , for z
L ≥ 0(stable) .

These are slight modifications to relations in Eqs. 6.13 and 6.16 that have the advantage
z
of explicitly relating L and Ri via,

z
L = Ri , for Ri < 0
z Ri
(6.23)
L = 1−5Ri , for 0 ≤ Ri ≤ 0.2 .

The integration with respect to height z to find vertical profiles of u and φ using these
M-O version of similarity functions yields the ubiquitous expressions,
h i
u∗
u= κ ln( zz0 ) − Ψm ( Lz )
θ∗
h i (6.24)
(θ − θ0 ) = κ ln( zz0 ) − Ψh ( Lz )
 6.6 PREDICTION OF THE MIXING HEIGHT 189

in which θ0 is the extrapolated temperature at z = z0 . Functions Ψm and Ψh are given by,

z
z dζ
Z  
L
Ψm ( zz0 ) = 1 − φm ( ) ,
z0
L
L ζ
(6.25)
z
z dζ
Z  
L
Ψh ( zz0 ) = 1 − φh ( ) .
z0
L
L ζ

Using the M-O version given in Eqs. 6.22, and considering that for moderately rough
z0
surfaces L ≈ 0, then the integrals in Eqs. 6.25 are,

Ψm = Ψh = −5 Lz , for z
L ≥0
 2
 2 
Ψm = ln 1+x
2
1+x
2 − 2 tan−1 x + π
2 , for z
L <0 (6.26)
 2

1+x z
Ψh = 2 ln 2 , for L <0

1
where x = 1 − 15 Lz 4 . This version of M-O similarity is the most widely used.

6.6 Prediction of the mixing height

As shown in Chapter 5 when the box model example was demonstrated in Section 5.9,
the mixing height hi plays a central role in air pollution studies, since it determines the
dilution of pollutant emissions in a given locality. Here we present a set of empirical
procedures to obtain hi from heat fluxes and other M-O parameters. In this section, we
mainly follow Cimorelli, et al, [9].
The general approach is to consider that hi depends on the combination of two factors:
the mechanical or shear turbulence combined with buoyant accelerations in a Convective
Boundary Layer, CBL, and the mechanical or shear turbulence in a Stable Boundary Layer,
SBL. Mechanical or shear turbulence is provided by the flow and surface interaction due
to viscosity as described in Section 6.1, at the beginning of this chapter. The interactions
between buoyancy and mean wind were described in Section 6.2.
We will denote by zim the mixing height due to mechanical or shear turbulence, and
by zic the mixing height due to convective or buoyant processes. The same expression for
calculating zim is used in both the CBL and the SBL.
190 Chapter 6. BUOYANCY AND VISCOSITY IN THE ATMOSPHERIC BOUNDARY LAYER

Example 6.2
Example: Application of the Gradient Method
The following hourly averaged measurements were obtained in an homogeneous
rural site, ([3], p. 231):

z(m) 2 8
u(ms−1 ) 3.34 3.98
T (◦ C) 28.04 28.10

a Obtain the velocity and potential temperature gradients at 4m.

The height at which data is going to be placed is at the geometric mean, i.e., at
√
zm = z1 z2 = 4m. The corresponding logarithmic finite-difference approxima-
tions for the velocity, temperature are:

∂u 1 ∆u 0.64
=   = = 0.115 s−1 ,
∂z zm ln 2
z 4 ln 4
z1

∂T 1 ∆u 0.94
=   =− = −0.170 K m−1 .
∂z zm ln 2
z 4 ln 4
z1

Therefore, the gradient of the potential temperature at zm =4 m is,

∂θ ∂T
≈ + Γ = −0.170 + 0.01 = −0.16 K m−1 .
∂z ∂z
b Obtain at 4m the gradient Richardson number, Ri , and the Obukhov length, L.
At zm =4 m,
g ∂Θ∂z
v
9.81 × 0.16
Ri =  2 = − = −0.387 ,
Tv ∂u 302.2 × 0.0132
∂z

Since
zm
= Ri = −0.387 ,
L
then
4
L=− Ri = −10.33 [m] .
0.387
c Obtain friction velocity, u∗ , temperature scale, θ∗ , and surface heat flux, H0 .
Hint: use Eq. 6.22 to obtain φm,h . To obtain u∗ and θ∗ first solve exercise xx.
Remember H0 = −ρcp u∗ θ∗ [W m−2 ].
 6.6 PREDICTION OF THE MIXING HEIGHT 191

6.6.1 Convective mixing height zic

The mixing height due to convective or buoyant processes zic is obtained using the fol-
lowing energy balance,

zic H(t0 ) 0 t
Z Z
zic θ(zic ) − θ(z)dz = (1 + 2A)dt , (6.27)
0 0 ρcp
where θ is potential temperature, H is the surface sensible heat flux as a function of time
beginning at sunrise when H becomes positive, A = 0.2 according to Deardorf, [12],
and t is the hour after sunrise. Weil and Brower, [13], found good agreement between
predictions and observations of zic , using this approach.

6.6.2 Mechanical mixing height zim

During night the CBL depth zim is mainly controlled by mechanical turbulence. It can be
calculated by assuming that it approaches the equilibrium height zie given by Zilitinkevich
in [?] as

u∗ L
 
zie = 0.4 , (6.28)
f
where f is the Coriolis parameter, i.e., f = 2Ω sin(φ), where Ω = 7.2921 × 10−5 rad s−1 ,
and φ is latitude. In the mid-latitudes, f = 10−4 rad s−1 .
This solution is not smooth in time. To avoid sudden and unrealistic changes in zim ,
the CBL depth, the solution is smoothed using Venkatram, [11], approach via

dzim zie − zim

= , (6.29)
dt τ
where the time scale τ governs the rate of change in height of the layer: As τ is smaller,
the faster the smoothing velocity and vice-versa. Also, as zim approaches the equilibrium
height zie , the change in height with time goes to zero, thus reaching a smooth in time
solution.
In order to select a satisfactory value for τ , it is chosen proportional to mixed layer
height and surface friction velocity,
zim
τ= , (6.30)
βτ u∗
where usually βτ = 0.2. Check that when u∗ is high, τ is low, increasing the velocity to
reach equilibrium height zie , and vice-versa.
The smoothed evolution value of zim is obtained by numerically integrating Eq.6.30
such that,
−∆t −∆t
 
zim (t + ∆t) = zim (t)e τ + zie (t + ∆t) 1 − e τ , (6.31)
192 Chapter 6. BUOYANCY AND VISCOSITY IN THE ATMOSPHERIC BOUNDARY LAYER

where zim (t) is the previous hour smoothed value. For computing the time scale τ in
Eq.6.30, zim is taken from the previous hour estimate and u∗ from the current hour. In
this way, the time scale (and thus relaxation time) will be short if the equilibrium mixing
height grows rapidly but will be long if it decreases rapidly.
In mid-latitudes, Eq. 6.28 can be empirically represented, as shown by Venkatram,
[14], by
3
zie = 2300u∗2 , (6.32)

where zie is the unsmoothed mechanical mixed layer height.

In practice, during the beginning of the day when the Stable Boundary Layer, SBL,
transitions to the Convective Boundary Layer, CBL, Eqs. 6.32 and 6.31, are used even
though they are primarily designed for the SBL.
 EXERCISES 193

Exercises

6.1 Math Exercise

Show that also

ρ T 1
= ( ) (γ−1) (6.33)
ρ0 T0
(Answer on Page 195)

6.2 Another math proof needed

∂Θ
Proof mathematically that ∂z = 0.
(Answer on Page 195)

6.3 Use table to obtain results

Using values in table below, obtain:
a) Virtual temperature at each level and its difference form actual environment tem-
perature.
b) Compare gradients of potential temperature and virtual temperature between
adjacent measurement levels.
c) The local static stability parameter, s and the percentage contribution of specific
humidity gradient to the same at different height levels.

Table 6.1. Measurements on a research vessel

Height above sea-surface: 2.1 3.6 6.4 11.4

Air Temperature (C) 27.4 27.72 27.69 27.62
Specific humidity (g kg −1 ) 18.02 17.83 17.54 17.29

(Answer on Page 195)

6.4 Data from Radiosonde

Using data below obtained using a radiosonde, obtain and plot
a) Virtual temperature as a function of height.
b) Virtual potential temperature as a function of height .
194 Chapter 6. BUOYANCY AND VISCOSITY IN THE ATMOSPHERIC BOUNDARY LAYER

c) Characterize various layers on the basis of local and non-local stability

d) estimate height of inversion layer and ABL height.

Table 6.2. Radiosonde measurements

Height (m) Pressure (mb) Temperature (C) Specific Humidity

(g kg −1 )
1.5 (Surface) 1013 11.8 6.5
50 1007 10.9 6.1
100 1001 10.1 5.8
150 995 9.6 5.6
200 989 9.2 5.5
300 977 8.4 5.3
400 965 7.4 5.1
500 954 6.4 4.9
550 948 6.0 4.9
600 942 5.6 4.8
650 936 5.1 4.7
700 931 4.9 4.6
750 925 4.9 4.5
800 919 4.9 4.5
850 914 5.0 4.5
900 908 5.0 4.5
1000 897 4.8 4.4

(Answer on Page 195)

6.5 Logarithmic approximation

Obtain finite logarithmic approximation to friction velocity u∗ and temperature scale θ∗
from Eqs. 6.12, 6.14, respectively.
(Answer on Page 195)
 EXERCISES 195

Answers to Exercises

Answer to Exercise 6.1 (page 193)

Answer for exercise 6.1

Answer to Exercise 6.2 (page 193)

Answer for exercise 6.2

Answer to Exercise 6.3 (page 193)

Answer to exercise 6.3

Answer to Exercise 6.4 (page 193)

Answer to Exercise 6.4

Answer to Exercise 6.5 (page 194)

Answer to Exercise 6.5
196 Chapter Bibliography

Chapter Bibliography

[1] Alexandre J Chorin, Jerrold E Marsden. A Mathematical Introduction to Fluid Mechanics, 3rd
edition. Springer, 2013. ISBN: 1461269342.
[2] Pijush K. Kundu, Ira M. Cohen, David R Dowling Fluid Mechanics,6th Edition Academic
Press,2016. ISBN-13: 978-0124059351
[3] S. Pal Arya Introduction to Micrometeorology, 2nd Edition Academic Press,2001. ISBN-13:
978-0-12-059354-5
[4] Stanhill, G. A simple instrument for field measurement of tutbulent diffusion flux. J. Climate
Appl. Meteor. 8, 509-513, (1969).
[5] Monin, A. S., and Obukhov, A. M. Basic laws of turbulent mixing in the surface layer of the
atmosphere, Tr. Akad. Nauk SSSR Geofiz. Inst., 24, 163-187, (1954). English translation by John
Miller, 1959.
[6] John C. Wyngaard Turbulence in the Atmosphere. Cambridge University Press, (2010). ISBN
978-0-521-88769-4
[7] Izumi, Y. Kansas 1958 Field Program data report. Environmental Research Paper, No. 379,
(1971). Air Force Cambridge Research Laboratories, Bedford, MA.
[8] Hogstrom U. Non-dimensional wind and temperature profiles in the atmopsheric surface
layer: a re-evaluation. Boundary-Layer Meteorol., 42,55-78, (1988).
[9] Alan J. Cimorelli, Steven G. Perry, Akula Venkatram, Jeffrey C. Weil, Robert J. Paine, Robert
B. Wilson, Russell F. Lee, Warren D. Peters, and Roger W. Brode. AERMOD: A Dispersion
Model for Industrial Source Applications. Part I: General Model Formulation and Boundary
Layer Characterization. Journal of Applied Meteorology and Climatology, 44, 5,p682-693,(2005).
DOI: https://doi.org/10.1175/JAM2227.1
[10] Zilitinkevich, S.,S. On the determination of the height of the Ekman boundary layer. Bound.
Layer Meteor..,3, 141-145.
[11] Venkatram, A. A semi-empirical method to compute concentration associated with surface
relases in the stable boundary layer. Atmos. Environ.,26A, 947-949, (1982).
[12] Deardorff,J.,W., Progress in Understanding Entraunment at the top of a Mixed Layer. Work-
shop on the Planetary Boundary, American Meteorological Society, (1980).
[13] Weill, J.C. and R. P. Brower. Estimating convective boundary layer parameters for diffusion
applications. PPSP-MD-48, Maryland Power Plant Siting Program, Maryland Department of
Natural Resources, (1983).
[14] Venkatram, A. Estimating the Monin-Obukhov length in the stable boundary layer for dis-
persion calculations. Bound. Layer Meteor.,19, 481-485, (1980).
Chapter 7

Transport Processes and

Atmospheric Turbulence

7.1 Mass conservation, the continuum assumption and 198

incompressibility
7.2 The Advection Diffusion Equation (ADE) 208

7.3 Solution of the Advection-Diffusion Equation 211

7.4 The nature of the solution of the Turbulent Advection Diffusion 214
Equation
7.5 Pointwise source Gaussian transport 215

7.6 Continuous source with deposition and settling: The Ermak 223
solution
7.7 Effective stack height 224

Chapter Bibliography 230

An equation able to describe the concentration values of a non-reactive substance as a

function of time and position will be obtained. For such purpose Partial Differential
Equations (PDE’s)are used. They describe a large set of physical phenomena and in par-
ticular, the linear second order PDE’s Advection-Diffusion Equation (ADE) describes
transport of mass and heat in the atmosphere. The ADE predicts the transport of a chem-
ical specie or a scalar variable in the atmosphere due to advective (wind speed) and
turbulent eddy-diffusivity mechanisms. A solution in the form of a Gaussian distribution

– 197 –
198 Chapter 7. TRANSPORT PROCESSES AND ATMOSPHERIC TURBULENCE

will be obtained based on mathematical technique named separation of variables. Based

on emissions rates, wind variables and turbulent parameters, it will predict non-reacting
concentration values as a function of position and time. An example of an emission whose
concentrations we will be able to predict is shown in the painting by Camille Pissarro
shown in Fig. 7.1. He was interested in depicting the changing landscape in Europe at
the end of the 19th century.

Figure 7.1. Perhaps naively, Camille Pissarro captures the changing landscape in
France during the 1870’s due to industrialization. In this chapter, the emission of
chimneys as depicted in this painting, will be mathematically described allowing the
estimation of downwind average pollutant concentrations.

7.1 Mass conservation, the continuum assumption and incompressibility

To describe mass transport in the atmosphere, we start by invoking the conservation of

mass principle: mass is neither created nor destroyed.
Also, we assume that for each time t the atmosphere behaves as a fluid with a well-
defined mass density c(x, t) measured using concentration units. This means that if there
is a subregion in space D, the mass m of fluid in D can be obtained by integrating c in
7.1 MASS CONSERVATION, THE CONTINUUM ASSUMPTION AND INCOMPRESSIBILITY 199

the volume V occupied by D,

Z
m(D, t) = c(x, t)dV . (7.1)
D

This is the continuum assumption. To perform this integral c must be continuous, some-
thing that doesn’t hold for molecular structure scale but that is extremely accurate for the
macroscopic case.
Another assumption is that air is considered incompressible since we will be con-
cerned with phenomena with wind speeds much less than 300 km h−1 . This means that
atmospheric air compression is solely due to gravity thus forming a layered fluid with
pressure (and density) variation with height, as described in Chapters 2 and 5. The main
point here is that the considered wind velocities are not enough to compress air. This is
expressed mathematically by the dot vector product,

∇ · u = 0, (7.2)

in

which u= (u, v, w) is wind speed vector, and ∇, or nabla, is the differential operator
∂ ∂ ∂
∂x , ∂y , ∂z . A 1-D schematic of compressibility and incompressibility due to x wind
component u is given in Fig. 7.2.

Figure 7.2. Wind velocity gradient and compressibility. In (a) zero velocity gradient
implies incompressibility. In (b) and (c), the velocity gradient sign determines zones
with higher or lower pressure. These become apparent in the atmosphere only for
air flows with gradients in the hundreds of km h−1 . In (d), gravity g determines that
zones near surface have higher pressures and elevated zones have lower pressure.
200 Chapter 7. TRANSPORT PROCESSES AND ATMOSPHERIC TURBULENCE

7.1.1 Continuity Equation.

In this subsection we mainly follow Chorin text, [1]. Consider a subregion D fixed in
time but of arbitrary size. We want to know the change in time of mass in D. Then,

d d dc(x, t)
Z Z
m(D, t) = c(x, t)dV = dV . (7.3)
dt dt D D dt
The derivative operator was safely introduced inside the integral because c is contin-
uous. Let ∂D be the boundary of D; let n be the unit outward normal at points of D; and
let dA the area element of ∂D as shown in Fig. 7.3 . The flow per unit area across ∂D is
obtained by u · n and the mass flow by cu · n.

Figure 7.3. A regular domain D with boundary ∂D. Unit outward normal vector n
on the surface. Vector velocity u, exiting domain

Then, by the principle of mass conservation,

d
Z Z
c(x, t)dV = − cu · ndA , (7.4)
dt D ∂D
which means that the change in time of mass inside the volume V occupied by D is equal
to the mass flow leaving D through a unit area dA of ∂D. This is the integral form of
the law of conservation of mass.
The law of conservation of mass expressed in Eq. 7.4 equates a volume with a surface
integral. More useful information is extracted if we obtain a volume integral on both
7.1 MASS CONSERVATION, THE CONTINUUM ASSUMPTION AND INCOMPRESSIBILITY 201

sides. The convergence theorem states that the surface integral of the normal component of a
vector taken over a closed surface is equal to the integral of the divergence of this vector taken over
a volume enclosed by the surface. Then, the area integral on the right hand side of Eq. 7.4
can be expressed as a volume integral:
Z Z
cu · ndA = ∇ · (cu)dV . (7.5)
∂D D

Therefore, Eq. 7.4 becomes

d
Z  
c(x, t) + ∇ · (cu) dV = 0 . (7.6)
D dt
Since the integration domain D is arbitrary and integrand is continuous then the
integrand is identical to zero,

d
c(x, t) + ∇ · (cu) = 0 . (7.7)
dt
Z π
At this point the reader my object. For example, sin xdx = 0 but the integrand is
−π
not identical to zero for all x! In this case, the integral limits are exactly one period of
sin x and therefore total area under the function is zero. But note that we stated that D
in Eqs. 7.4 and 7.6 is arbitrary. This means that if we are free to choose as many integral
limits as we want and the integral result is always zero, the integrand is identical to zero.
This results is known as the Du Bois-Reymond Lemma.
Equation 7.7 is the differential form of the law of conservation of mass. It is also
known as the continuity equation. Unlike other texts, we were able to arrive to this
expression without invoking conservation of mass in a infinitesimal cube, where neither
the continuum assumption nor the macroscopic laws of physics hold.

7.1.2 Advection
Carrying through the differentiation procedure ∇ · (cu) in Eq. 7.7,

d d
c(x, t) + ∇ · (cu) = c(x, t) + u · ∇c + c∇ · u . (7.8)
dt dt
Since we consider atmospheric velocities such that air is incompressible, c∇ · u = 0.
Then Eq. 7.8 becomes,

d d ∂ ∂ ∂
c(x, t) + u · ∇c = c(x, t) + u c(x, t) + v c(x, t) + w c(x, t) = 0 (7.9)
dt dt ∂x ∂y ∂z
| {z }
Advection
202 Chapter 7. TRANSPORT PROCESSES AND ATMOSPHERIC TURBULENCE

where the operator ∇ is expanded to explicitly show the advection component in Eq. 7.9.
This term represents the action of the wind field u = (u, v, w) transporting mass con-
centration c. Equation 7.9 states that that the change in density is due to velocity field
u = (u, v, w): just consider u = 0 in Eq. 7.9 and c(x, t) becomes a constant. A windfield u
different to zero must exists to obtain a change in concentration values in time and space.

7.1.3 Advection and turbulence

Atmospheric variables such as wind are highly irregular and random. We first encoun-
tered turbulence in Chapter 6, when the interaction of a surface and a viscous fluid was
present. As shown in Fig. 7.4 not only wind speed, but humidity and temperature are
also irregular. In fluid dynamics this situation is called turbulence. This phenomenon
can only be characterized, since there is no a precise definition.

Figure 7.4. Observed time series of velocity, temperature, and specific humidity
fluctuations in the atmospheric surface layer at suburban Vancouver, Canada, dur-
ing moderately unstable conditions, from Roth [2].
7.1 MASS CONSERVATION, THE CONTINUUM ASSUMPTION AND INCOMPRESSIBILITY 203

According to Tennekes, [3], its characteristics are:

Unsustainability. Turbulence cannot maintain itself but depends on its environment
to obtain energy. If energy is not provided, the phenomenon tends to decay.
Irregularity. Turbulence is characterized by irregular or randomness. A deterministic
approach is impossible and therefore statistical methods are used.
Diffusivity. Diffusivity of turbulence causes rapid mixing and therefore increased
rates of momentum, heat and mass transfer. These rates are several orders of magni-
tude larger than rates due to molecular diffusion.
Instability. Turbulence originates as an instability of laminar flows as velocity in-
creases, or more, accurately as Reynolds number increases.
Three dimensionality. Turbulent diffusion is rotational and therefore three dimen-
sional.
Turbulence is a flow property. Turbulence is not a feature of the fluid but of fluid
flows. It can occur in a gas or a liquid.
These characteristics make impossible to physically reproduce a turbulent phenomenon.
The only possibility will be to obtain an estimation of mean variables. For this purpose,
the following techniques are used to incorporate turbulence into the Advection Equation
in Eq. 7.9.

Fluctuations, mean and Reynolds decomposition

In order to deal with highly irregular turbulence flow it is necessary to use a statistical
description and find a way to deal only with averaged or mean variables. Average statis-
tics, such as means, variances, and higher moments are the mathematical tools used to
describe turbulent flow to obtain equations of mean motion.
In a turbulent flow in the Planetary Boundary Layer, PBL, velocity, temperature, mass
concentrations and other variables vary irregularly in space. To deal with this problem,
variables are decomposed in the mean and fluctuation part:

ũ = u + u0 (7.10)
ṽ = v + v 0 (7.11)
w̃ = w + w0 (7.12)
0
θ̃ = θ + θ (7.13)
0
c̃ = c + c (7.14)

The left hand side represents the instantaneous value (denoted by a tilde). On the
right hand side, mean (denoted by upper bar) and fluctuating parts (denoted by primed
204 Chapter 7. TRANSPORT PROCESSES AND ATMOSPHERIC TURBULENCE

letters) are contained. This is called Reynolds decomposition. The mean value, is the
value around which the ”needle” vibrates or fluctuates, when we placed an anemometer
in Fig. 6.3. The instant fluctuating values of the velocity are ũ.
The time mean f of variable f˜ is defined as

1 T ˜
Z
f= f (t)dt , (7.15)
T 0
where T is the sampling time over which the mean is desired. It should be long enough
to incorporate the effects of significant large eddies of the flow, and not too long to mask
trends such as diurnal variations of the flow. In the Planetary Boundary Layer (PBL), T
is in the order of 103 or 104 s.
Since in Reynolds decomposition a fluctuation is the deviation of an instantaneous
variable from it mean, by definition the mean of any fluctuating variable f 0 is zero, i.e. f 0 =
RT
1
T 0 f 0 (t)dt = 0. This means that, on the average, there ere are compensating negative
fluctuations for positive fluctuations.
An example of this situation is shown in Figure 7.5. It shows how a Gaussian proba-
bility distribution resulted by counting the number of appearances of values of a turbu-
lent variable. The Gaussian is a typical probability distribution for randomly distributed
atmospheric turbulent variable, such as velocity, temperature and concentrations in mod-
erate and strong winds. As such 99.6 percent of the area of this probability distribution
falls between -3 and +3 standard deviations σ. The mean of the fluctuations is equal to
zero.

Figure 7.5. Schematic of determining a Gaussian probability density of a stationary

turbulence variable.

In general, in the surface layer the average magnitudes of vertical fluctuations are
much larger than the mean velocity, whereas the magnitude of horizontal velocity fluc-
tuations are of the same order or less than the mean horizontal velocity. The magnitudes
of fluctuations of thermodynamic variables are at least two order of magnitude smaller
7.1 MASS CONSERVATION, THE CONTINUUM ASSUMPTION AND INCOMPRESSIBILITY 205

than their mean values. The relative magnitudes of turbulent fluctuations decrease with
increasing stability and height in the PBL. Therefore, σ contains the information about
”how turbulent” a flow is: smaller values of σ imply less turbulence, and viceversa.
As shown in Fig. 7.4 , measurements taken at height of 27.4 m under unstable condi-
tions, show that there is a marked difference between vertical and horizontal velocities,
temperature and absolute humidity fluctuations. It can be seen that under unstable and
convective conditions, buoyant plumes and thermals cause strong asymmetry between
positive and negative fluctuations, especially in vertical velocity, temperature and humid-
ity. This is explained in Fig. 7.6, where the probability density is positively skewed and
the turbulent variable is more likely to take on large positive values than large negative
values. Under near-neutral conditions and strong winds, the fluctuations of the variables
tend to be more similar and symmetrical with the mean, as shown in Fig. 7.5.

Figure 7.6. Schematic of determining a positively skewed probability density of a

stationary turbulence variable.

Ensemble-averaged turbulence models

In this section the Reynolds decomposition of wind velocity will be inserted into the
advection equation Eq. 7.9. For that, the Reynolds-averaging conditions are needed.If f
and g are two independent variables with mean values f and g, and a is a constant, then:

a = a (7.16)
f = f (7.17)
f +g = f +g (7.18)
af = af ; (7.19)
206 Chapter 7. TRANSPORT PROCESSES AND ATMOSPHERIC TURBULENCE

∂f ∂f
= (7.20)
∂s ∂s
where s = x, y, z, or t. These equations conform what is known as Reynolds algebra.
They express the distributive property of summation, and commutativity of averaging on
differentiation and integration. Also, constants are not affected by averaging and averages
behave like constants.
An important element missing in this algebra is what happens when the average of
the product of two fluctuating variables is taken ,i.e., f 0 g 0 . This product is not defined and
when this case appears special action and compromises must be taken, as will be shown.
The Reynolds-averaged equations consist on substituting in the advection equation
(or in Navier-Stokes equations) the variables ũ = u + u0 , ṽ = v + v 0 , c̃ = c + c0 , ...,
and assume the consequences. When Reynolds decomposed variables are substituted in
Eq. 7.9 the following equation is obtained:
d d d
(c + c0 ) + (u + u0 ) · ∇(c + c0 ) = c + c0 + u · ∇c + u0 · ∇c + u · ∇c0 + u0 ∇ · c0 = 0 . (7.21)
dt dt dt
By taking the Reynolds average of this equation and applying Reynolds algebra:
d d
c + c0 + u · ∇c + u0 · ∇c + u0 · ∇c + u · ∇c0 + u0 ∇ · c0 = 0 . (7.22)
dt dt
By setting to zero all the occurrences where the average of a fluctuating variable appears:
d
c + +u · ∇c + u0 ∇ · c0 = 0 . (7.23)
dt
The last term was not cancelled since it is the Reynolds average of the product of two
fluctuating variables. For the time being, we can only put this term in a more amenable
form by considering again, incompressibility. Consider that,

∇ · u0 c0 = u0 ∇ · c0 + c0 ∇ · u0 . (7.24)

By incompressibility last term is zero, obtaining

∇·u0 c0 = u0 ∇ · c0 . (7.25)

This into Eq. 7.23 gives,

d
c + u · ∇c + ∇ · u0 c0 = 0 . (7.26)
dt
For sake of economy, the overline from averaged quantities is dropped. Then, an expanded
version of last equation becomes,

∂c ∂c ∂c ∂c ∂u0 c0 ∂v 0 c0 ∂w0 c0
+u +v +w + + + = 0. (7.27)
∂t ∂x ∂y ∂z ∂x ∂y ∂z
| {z } | {z }
Advection Diffusion terms
7.1 MASS CONSERVATION, THE CONTINUUM ASSUMPTION AND INCOMPRESSIBILITY 207

Note that we started with a pure advection equation for c and by introducing fluctuations
u0 and c0 a set of new variables called diffusion terms appeared: the turbulent fluctuations
u0 c0 , v 0 c0 , w0 c0 . These terms are covariances or turbulent fluxes. These are spatial gradients
(divergence) of turbulent transport and represent a set of new unknowns: besides c we
need to solve for these turbulent fluxes too. How do we generate a complete a set of
equations and unknowns? This situation is called the turbulence closure problem.

7.1.4 Solving the turbulence closure problem

A scheme named K-theory is used to overcome the appearance of new variables. The
problem is ”solved” or ”closed” by considering that turbulent diffusion is analogous to
molecular diffusion or Fick’s Law. The reasoning is that turbulent transport is proportional
to the gradient in the direction of bulk velocity:
∂c
u0 c0 = −Kx ( ) (7.28)
∂x
∂c
v0 c0 = −Ky ( ) (7.29)
∂y
∂c
w0 c0 = −Kz ( ) (7.30)
∂z
(7.31)

where the proportionality constants Kx , Kx and Kz are called turbulent eddy diffusivi-
ties. We considered here horizontal isotropy, i.e., horizontal turbulent diffusivity is same
regardless of direction:Kx = Ky . As a matter of fact, vertical diffusivity is much larger
than horizontal turbulent diffusivity.
If temperature is the scalar parameter to be transported, we might as well use as
analogy Fourier’s law of heat conduction:
∂θ
u0 θ0 = −Kh ( ) (7.32)
∂x
∂θ
v0 θ0 = −Kh ( ) (7.33)
∂y
∂θ
w0 θ0 = −Kz ( ) (7.34)
∂z
where horizontal isotropy is supposed, and Kh = Kx = Ky . Then Kh and Kz are the
horizontal and vertical eddy-diffusion coefficients respectively.
There is no sound physical basis for turbulent K-theory, but if we substitute Eqs. 7.31
or Eqs. 7.34 in Eq. 7.27 and consider the diffusion coefficients constant in space and time,
we get
∂c ∂c ∂c ∂c ∂2c ∂2c ∂2c
+u +v +w − (Kh 2 + Kh 2 + Kz 2 ) = 0 , (7.35)
∂t ∂x ∂y ∂z ∂x ∂y ∂z
208 Chapter 7. TRANSPORT PROCESSES AND ATMOSPHERIC TURBULENCE

which is an advection-diffusion equation (ADE) on mean variables, namely of c under

the influence of mean flow u = (u, v, w) and eddy diffusion coefficients Kh , Kz . Since
it is a linear partial differential equation (PDE), an analytic solution for a non-reactive
contaminant c can be found. Initial and boundary conditions for c must be provided.
The fact that we arrived to this equation is a great accomplishment since we are in the
verge of obtaining a closed form formula for c. Nevertheless, an important hurdle must
be cleared: How to obtain the eddy diffusion coefficients Kh , Kz . For this topic we will
devote Section 7.5.
In what follows, the analytic solution of Eq. 7.35 will be obtained and, in the process,
important information on its characteristic behaviour will be discovered.

7.2 The Advection Diffusion Equation (ADE)

As shown, the ADE is a convenient prototype for a mathematical model of transport pro-
cesses, and therefore is used in fluid dynamics when mass or another scalar variable such
as temperature is transported due to the combined action of a wind field and diffusive
turbulent processes. Consider a four dimensional variable c(t, x, y, z).The transport of c
is given by
∂c ∂c ∂c ∂c ∂2c ∂2c ∂2c
+u +v +w −(Kx 2 + Ky 2 + Kz 2 ) = 0 (7.36)
∂t ∂x ∂y ∂x ∂x ∂y ∂z
| {z } | {z }
Advection Turbulent diffusion

As shown in Eq. 7.36 an ADE describes transport of c by two phenomena: advection

where transport is due to the action of a wind field u = (u, v, w), and transport by diffu-
sive processes with the corresponding diffusion coefficients Kx , Ky , Kz . If the units used
for u are m s−1 then from Eq. 7.36 the units of Kx , Ky , Kz must be m2 s−1 . The ADE
is broken into its advective and diffusive parts for its analysis. We take advantage of a
mathematical technique that uses tests functions to obtain pure wave solutions or nor-
mal modes. These will reveal the physical nature of advection and diffusion, an important
objective of this chapter.

7.2.1 Behaviour in time of the advection term

For sake of simplicity consider a one dimensional advection problem
∂c ∂c
+u =0. (7.37)
∂t ∂x
This equation describes the change of c while being carried by a wind with speed u. Insert
the test function c(x, t) = a(k)eikx+λ(k)t into Eq. 7.37 and consider a(k) as a constant and
 7.2 THE ADVECTION DIFFUSION EQUATION (ADE) 209

√
i = −1. The parameter k is assumed to be real and λ(k) must be chosen such that
Eq. 7.37 is satisfied. The solution c(x, t) found for each k is called normal mode. This
procedure yields,
λ = −iuk . (7.38)

The normal mode is then,

c(x, t) = a(k)eik(x−ut) . (7.39)

Note that the normal mode has the form of a wave travelling with a speed u. Also note
that at t = 0 the magnitude is a(k), and that this magnitude is preserved for all t. It can be
concluded that this equation describes a phenomenon where the magnitude and shape
of the initial conditions is perfectly preserved: There are are no built in attenuation mech-
anisms. A schematic of this is shown in Fig. 7.7a, where series of snapshots are shown
representing an initial conditions travelling through space do to wind action, without any
change in shape.

7.2.2 Behaviour in time of the diffusion term

For simplicity consider a one dimensional diffusion process described by

∂c ∂2c
= Kx 2 (7.40)
∂t ∂x

We try again the test function method and insert c(x, t) = a(k)eikx+λ(k)t into Eq. 7.40 to
obtain the normal modes. This time we get

λ = −Kx k 2 (7.41)

The corresponding normal mode is

2t
c(x, t) = a(k)eikx−Kx k (7.42)

In this occasion we have a decaying with time solution since |c(x, t)| = a(k)e−Kx t . The
rate of decadence will be dictated by the magnitude of Kx ; the larger its magnitude the
faster the rate of decay with time. This equation is describing a phenomenon where there
is a built in attenuation mechanism. A schematic of this behaviour is shown in Fig. 7.7 b,
where snapshots are shown of an initial condition and its development through time .
The diffusion processes we are interested are not of molecular type (micro scale),
but due to turbulent mechanisms or eddy diffusion (macro scale). Nevertheless, we will
impose a similar mathematical description, such as Fick Law presented in Subsection 7.1.4
, and specific schemes for pollution problems in Section 7.5.
210 Chapter 7. TRANSPORT PROCESSES AND ATMOSPHERIC TURBULENCE

7.2.3 Behavior in time of the advection and diffusion terms

Consider now a one dimensional problem of advection and diffusion described by

∂c ∂c ∂2c
+u = Kx 2 (7.43)
∂t ∂x ∂x
In this case the concentration c(x, t) is being transported by the wind with velocity
u and by eddy diffusion Kx in the x spacial dimension. Once more insert c(x, t) =
a(k)eikx+λ(k)t into Eq. 7.43 to analyze the normal modes. We get ,

λ = −Kx k 2 − iuk (7.44)

The corresponding normal mode is

2t
c(x, t) = a(k)eik(x−ut) e−Kx k (7.45)

The normal mode shows a traveling wave (the advection part) multiplied or modulated
by a decaying term (the diffusion part). The solution is then a wave traveling at speed u
decaying at a rate given by magnitude of the diffusion coefficient K as it travels in space.
This is schematically shown in Fig. 7.7c.

Figure 7.7. In (a), snapshots at time t1 , t2 and t3 of an initial condition advected by

wind component u. In (b), snapshots of a diffused initial condition. In (c), snapshots
of an advected and diffused initial condition. At all time and in all cases, mass is
conserved, i.e., area under the curve is always the same.
 7.3 SOLUTION OF THE ADVECTION-DIFFUSION EQUATION 211

7.3 Solution of the Advection-Diffusion Equation

In the previous section a solution in terms of normal modes for the ADE was obtained.
We learned that its solution is composed of a wave traveling at speed u, exponentially
decaying at a rate given by the magnitude of diffusion coefficient K. Although this is
valuable information, we must construct a ”physical solution”, i.e., a solution related to
a specific set of initial and/or boundary conditions describing a particular problem of
interest. The particular solutions obtained below can be found in texts such as [4].

7.3.1 Instantaneous source

Consider a point source (smoke stack) that is releasing instantaneously a puff qp units
of mass per unit time of a non-reactant chemical specie whose concentration is given by
c(t, x, y, z). Place the origin of a cartesian system of reference on this source and align a
horizontal axis in the direction of a mean constant wind in such a way that u = (u, 0, 0).
Under such conditions the transport of a non-reactant chemical specie from a point source
due to the action of wind speed u and turbulent diffusion is described by the advection
diffusion equation
∂c ∂c ∂2c ∂2c ∂2c
+u = (Kh 2 + Kh 2 + Kz 2 ) (7.46)
∂t ∂x ∂x ∂y ∂z
with boundary conditions

c(x, y, z, 0) = qp δ(x)δ(y)δ(z) (7.47)

c(x, y, z, t) = 0, x, y, z → ±∞ (7.48)

where (
1 if x = 0
δ(x) = (7.49)
0 otherwise

The function δ(x) is called the Dirac delta. The parameters Kh (= Kx = Ky ) and Kz
are the horizontal and vertical eddy diffusion coefficients. To solve Eq. 7.46 the prob-
lem is recasted by expressing the concentration of the transported specie as c(t, x, y, z) =
c1 (t, x) c2 (t, y) c3 (t, z). Here is assumed that the solution can be expressed as the mul-
tiplication of functions that depend on t and each one depends exclusively on x, y or z.
This scheme is called separation of variables. Its main justification is that it works.
Subsisting c as separated variables, Eq. 7.46 becomes

∂c1 ∂c1 ∂ 2 c1
+u = Kh (7.50)
∂t ∂x ∂x2
∂c2 ∂ 2 c2
= Kh 2 (7.51)
∂t ∂y
212 Chapter 7. TRANSPORT PROCESSES AND ATMOSPHERIC TURBULENCE

∂c3 ∂ 2 c3
= Kz 2 (7.52)
∂t ∂z
The normal modes for these equations were found above and can be used to obtain an
analytical solution for Eq. 7.50. For this we have to invoque superposition, a property of
linear PDE’s: If we have two normal modes for a linear PDE, their sum will also satisfy
this equation. Then if,

c1 (t, x; k1 + k2 ) = c1 (t, x; k1 ) + c1 (t, x; k2 ) (7.53)

c1 (t, x; k1 + k2 ) also satisfies the corresponding equation Eq. 7.50.

From the beginning we specified that the k’s are real numbers. Instead of Eq. 7.50, if
we form a solution by performing an infinitesimal sum over all k’s and using the normal
modes obtained in the preceding section, we get a solution
Z +∞ 2K
c1 (x, t) = a(k) e−(k x t−ik(x−ut))
dk (7.54)
−∞
1
R +∞ qp3
Using the boundary condition at t = 0 and the fact that −∞ eikx dk = 2π, a(k) = 2π .
Then, a solution for the ADE is
1
qp3
Z +∞ 2 K t−ik(x−ut))
c1 (x, t) = e−(k x
dk (7.55)
2π −∞

In order to evaluate the integral in Eq. 7.55 the square is completed in the exponent
(x − ut)2 (x − ut)2 1 i(x − ut)2 1 2 i(x − ut)2
k 2 Kx t − ik(x − ut)) − + = (k(Kx ) 2 − )2 ) −
4Kx t 4Kx t 2(Kx t 4Kx t
(7.56)
1 i(x−ut)2 1 1
Let η = k(Kx ) 2 − 2(Kx t )
2 and dη = (Kx t) 2 dk, then
1 1
qp3 (x−ut)2 2
Z +∞ 2
−
c1 (x, t) = 1 e 4Kx t e−η dη (7.57)
2π(Kx t) 2 −∞
R +∞ 2 1
Using the fact that −∞ e−η dη = π 2 ,
1
qp3 (x−ut)2
c1 (x, t) = 1 e− 4Kx t (7.58)
2(πKx t) 2

This procedure is carried out for the rest of the equations and the solution for 7.50 is
obtained,
1
2
qp3 y
− 4K
c2 (t, y) = 1 e ht (7.59)
2(πKh t) 2
1
qp3 z2
c3 (t, z) = 1 e− 4Kz t (7.60)
2(πKz t) 2
 7.3 SOLUTION OF THE ADVECTION-DIFFUSION EQUATION 213

The overall solution of 7.46 is then

(x−ut)2 y2 z2
qp −( + 4K + 4K )
c(t, x, y, z) = 3 1 e
4Kh t ht zt (7.61)
8(πt) 2 (Kh Kh Kz ) 2

7.3.2 Continuous source without reflection

Consider now a continuous source of strength q in units of mass per time . This means
that the source started emitting at t = 0 and continuos as t → ∞. It is expected that
the concentration reaches a steady concentration. Then the advection diffusion problem
becomes
∂c ∂2c ∂2c ∂2c
u = (Kh 2 + Kh 2 + Kz 2 ) (7.62)
∂x ∂x ∂y ∂z
As before the source q is placed in the origin. At x = 0 we must have that the material
flux that crosses plane at that point must equal the source strength,
∂c
u( )|x=0 = qδ(x)δ(y)δ(z) (7.63)
∂x
or Z
1
Z +∞
dc = δ(y)δ(z) qδ(x)dx (7.64)
u −∞
R +∞
Since by definition of Dirac delta −∞ δ(x)dx = 1, then
q
c(0, y, z) = δ(y)δ(z) (7.65)
u
Similar as before the other boundary is

c(x, y, z) = 0, x, y, z → ±∞ (7.66)

Following similar procedure as the case of instantaneous source the final solution is
2 2 2 1
q c
−( 2K x
(K y
+K z
+K ) 2 −x)
c(t, x, y, z) = 1 e x x y z (7.67)
4π(Ky Kz x2 + Kx Kz y 2 + Kx Ky z 2 ) 2

In principle we have an idealized but adequate mathematical model to describe the trans-
port in the atmosphere of a chemical specie or concentration. It can be simplified further
by aligning the center of the major emission with the x axis, and by considering that
∂c ∂ c 2
advection is dominant over turbulent diffusion in his axis, i.e., u ∂x >> Kh ∂x2.

As a result we get,
∂c ∂2c ∂2c
u = Kh 2 + Kz 2 (7.68)
∂x ∂y ∂z
With the further condition,
(7.69)
214 Chapter 7. TRANSPORT PROCESSES AND ATMOSPHERIC TURBULENCE

the following solution is obtained,

2 2
q u y
− 4x z
(K +K )
c(x, y, z) = 1 e y z (7.70)
4πr(Ky Kz ) 2

where r2 = x2 + y 2 + x2 . Serious limitations of this type of formula are present. For

example, if center line surface concentrations are considered, Eq. 7.68 reduces to
q
c(x, 0, 0) = 1 (7.71)
4πr(Ky Kz ) 2

This solution indicates that ground level concentrations do not depend on wind speed U
and is inversely proportional to x. Observations indicate that c is inversely proportional
to ux1.76

7.4 The nature of the solution of the Turbulent Advection Diffusion

Equation

We managed to obtain a relatively simple expression such as Eq. 7.70 to predict concentra-
tions. If we know the emission rate q, wind speed u, and turbulent diffusion coefficients
Kh , Kz , a concentration c(x, y, z) can be obtained. Besides the limitations noted as we
built Eq. 7.70, what does the calculated concentration values mean in reality?
First of all, Eq. 7.70 is the result of simplifications made to define the turbulent diffu-
sion coefficients K using Fick´s Law. More important, the equation uses averaged values
of turbulent wind speed u to obtain average concentrations c(x, y, z). In Fig. 7.8 is shown
the difference between an instant and averaged plume. Due to average smearing, the max-
imum concentrations may vary by a factor of 4 or more. Therefore, if maximum values
are above dangerous levels, their effects are missed when using averaged solutions.
 7.5 POINTWISE SOURCE GAUSSIAN TRANSPORT 215

Figure 7.8. Plume profiles and concentration distribution at different averaging

times. Note the peak concentrations reduction due to average smearing.

7.5 Pointwise source Gaussian transport

In order to use all the practical infrastructure developed in different areas of applied
mathematics and engineering regarding Gaussian distribution, the turbulent coefficients
Kh and Kz are converted to the more familiar standard deviation notation by using the
change of variables σh2 = 2 Kuh x and σz2 = 2 Kuz x in Eqs. 7.55 and 7.67. So the equations for
a instantaneous puff becomes
−y 2 −z 2
q 2
c(x, y, z) = e 2σh e 2σz2 (7.72)
2πuσh σz

where usually σ is in [m], u in [m s−1 ], q in [g s−1 ] and c in [ g m−3 ]. The values for
σ are referred to as the turbulent diffusion or turbulent dispersion parameter. It is the
standard deviation of the Gaussian normal to the plume axis. As such, its values are
directly proportional to the thickness of the plume. Since σs are proportional to x1/2 and
to u−1/2 , the plume grows monotonically with distance downwind and, becomes slender
with increasing wind speed.
For a continuous source we have,
−y 2 −z 2
q 2
c(x, y, z) = e 2σh e 2σz2 . (7.73)
2πuσh σz
These equations are useful only for the concentration in the down wind direction up to
the point in the x-direction where the concentration hits ground level (z = 0) . If there is
216 Chapter 7. TRANSPORT PROCESSES AND ATMOSPHERIC TURBULENCE

appreciable ”reflection” of pollutants from the ground in the downwind direction after
that point, their contribution will not be taken into account.
To consider the height H[m] of the stack, the change of variable z = z − H is per-
formed. Then, for a continuous source at height H we have,
−y 2 −(z−H)2
q 2
c(x, y, z) = e 2σh e 2σz2 . (7.74)
2πuσh σz
This equation has the form of a cone whose sections are ellipses. On the main axis,
which is the horizontal, a Gaussian distribution of the contaminant c with a standard
deviation σh is placed. On the minor axis, which is the vertical, a Gaussian distribution of
the contaminant c with a standard deviation σh is placed. These Gaussian distributions
describe the concentration profile. This is shown in Fig. 7.9.

Figure 7.9. Gaussian plume emanating form a stack with geometric height h. Actual
elevation of centerline is H= h+ ∆h. Coordinates are aligned with wind component u.
 7.5 POINTWISE SOURCE GAUSSIAN TRANSPORT 217

This form of the Gaussian solution for the ADE is used in the atmospheric transport
of pollutants literature. Practical procedures to obtain the σ’s are presented next.

7.5.1 The Pasquill-Gifford curves

The most accepted set of charts used to obtain the values for σh and σz are the Pasquill-
Gifford curves shown in Figs. 7.10 and 7.11. They are used under the following consid-
erations:
1. The concentrations obtained correspond to sampling times of 10 minutes.
2. The solid curves correspond to open country terrain and are refereed to as rural
curves, whereas the dashed curves (McElroy-Pooler curves) represent urban terrain.
3. The estimated concentrations represent more nearly the lowest several hundred meters
of the atmosphere.
If use of tables is not convenient, the next equations are used instead:

σh = c xd and σz = c xb (7.75)
218 Chapter 7. TRANSPORT PROCESSES AND ATMOSPHERIC TURBULENCE

Figure 7.10. Rural and urban horizontal dispersion coefficients σx as a function of

stability category. (From [4]).
 7.5 POINTWISE SOURCE GAUSSIAN TRANSPORT 219

Figure 7.11. Rural and urban vertical dispersion coefficients σz as a function of

stability category. (From [4]).
220 Chapter 7. TRANSPORT PROCESSES AND ATMOSPHERIC TURBULENCE

The distance x is in [km] and the σ’s in [m]. The parameters c, d and b are given in
following tables. Note that they are heavily dependent on the Pasquill stability categories.

Table 7.1. Parameters used to calculate Pasquill-Guifford σh . (From [4]).

Pasquill Stability
Category c d
A 24.1670 1.5334
B 18.3330 1.8096
C 12.5000 1.0857
D 8.3330 0.72382
E 6.2500 0.54287
F 4.1667 0.36191

For urban environments, the following formulas for σh and σz by McElroy and Pooler
are used.
Usually the estimation of σz is more in doubt than σh . This is especially true for
distances of more than 1 km. If conditions are neutral to moderately unstable, at distances
of few kilometers, the center line ground level concentrations using above scheme should
be within a factor of 2 to 3 of actual values.

7.5.2 Continuous source with reflection

To consider the situation where the pollutants are reflected to the atmosphere after hitting
the ground, a hypothetical mirror image emission source is created at height −H, as
shown in Fig. 7.12.
As a result we have,
−y 2 −(z−H) 2 −(z+H) 2
q 2
c(x, y, z) = e 2σh {αp e 2σz2 + (1 − αp ) e 2σz2 } (7.76)
πuσh σz
where αp can have values from 0 to 1, and is a pollutant reflection parameter. If αp = 0
there is no reflection (complete deposition) and if equal to 1, complete reflection. In-
termediate values of αp allow some degree of reflection or deposition. When αp =0.5,
Eq. 7.76 assumes the usually used form for the the Gaussian case with reflection.
 7.5 POINTWISE SOURCE GAUSSIAN TRANSPORT 221

Table 7.2. Parameters used to calculate Pasquill-Guifford σz . (From [4]).

Pasquill Stability
Category x (km) a b
A∗ < .10 122.800 0.94470
0.10 – 0.15 158.080 1.05420
0.16 – 0.20 170.220 1.09320
0.21 – 0.25 179.520 1.12620
0.26 – 0.30 217.410 1.26440
0.31 – 0.40 258.890 1.40940
0.41 – 0.50 346.750 1.72830
0.51 – 3.11 453.850 2.11660
> 3.11 ** **
∗
B <.20 90.673 0.93198
0.21–0.40 98.483 0.98332
> 0.40 109.300 1.09710
C∗ All 61.141 0.91465
D < .30 34.459 0.86974
0.31 – 1.00 32.093 0.81066
1.01 – 3.00 32.093 0.64403
3.01 – 10.00 33.504 0.60486
10.01 – 30.00 36.650 0.56589
> 30.00 44.053 0.51179
E < 10 24.260 0.83660
0.10 – 0.30 23.331 0.81956
0.31 – 1.00 21.628 0.75660
1.01 – 2.00 21.628 0.63077
2.01 – 4.00 22.534 0.57154
4.01 – 10.00 24.703 0.50527
10.01 – 20.00 26.970 0.46713
20.01 – 40.00 35.420 0.37615
> 40.00 47.618 0.29592
F < .20 15.209 0.81558
0.21 – 0.70 14.457 0.78407
0.71 – 1.00 13.953 0.68465
1.01 – 2.00 13.953 0.63227
2.01 – 3.00 14.823 0.54503
3.01 – 7.00 16.187 0.46490
7.01 – 15.00 17.836 0.41507
15.01 – 30.00 22.651 0.32681
30.01 – 60.00 27.074 0.27436
> 60.00 34.219 0.21716
∗
If the calculated value of σz exceeds 5000 m, σ is set to 5000 m
∗∗
σz is set to 5000
222 Chapter 7. TRANSPORT PROCESSES AND ATMOSPHERIC TURBULENCE

Table 7.3. Formulas used by Briggs to obtain σh and σz for urban cases. (From [4]).

Briggs Formulas Used to Calculate McElroy-Pooler σy

Pasquill Stability
Category σy (meters)*
A 0.32 × (1.0 + 0.0004 x)−1/2
B 0.32 × (1.0 + 0.0004 x)−1/2
C 0.22 × (1.0 + 0.0004 x)−1/2
D 0.16 × (1.0 + 0.0004 x)−1/2
E 0.11 × (1.0 + 0.0004 x)−1/2
F 0.11 × (1.0 + 0.0004 x)−1/2
∗
Where x is in meters

Briggs Formulas Used to Calculate McElroy-Pooler σz

Pasquill Stability
Category σz (meters)*
A 0.24 × (1.0 + 0.001 x)−1/2
B 0.24 × (1.0 + 0.001 x)−1/2
C 0.20×
D 0.14 × (1.0 + 0.003 x)−1/2
E 0.08 × (1.0 + 0.015 x)−1/2
F 0.08 × (1.0 + 0.015 x)−1/2
∗
Where x is in meters

Figure 7.12. An imaginary source is introduced to derive mathematically gaseous

reflection at surface of Earth. Effective stack height is H = h + ∆h and the region
with reflection starts where the reflected imaginary sources effluents enter positive
values of z.
 7.6 CONTINUOUS SOURCE WITH DEPOSITION AND SETTLING: THE ERMAK SOLUTION 223

Example 7.1
Example: Obtaining concentrations using Gaussian plume equation

Sulfur dioxide is emitted at a rate of 160 g s−1 from a stack with an effective height
H of 60 m in a rural setting. The wind speed at the height of the stack is 3 ms−1 ,
under atmospheric neutral conditions.
a Obtain the ground-level average concentration along the center-line at a distance
of 500 from the stack.
Substituting values in Eq. 7.76 with σh = 36m and σz =18.5m,x =500m,
2 2
160 × 106 1 −(−60) 1 −(60)
c(500, 0, 0) = e 2 2(18.5) + e 2(18.5)
π(6)(36)(18.5) 2
3
= 12.7 × 10 (5.25 × 10−3 )
= 66.0µm3 of SO2 . (7.77)

b Obtain the the ground-level average concentration along the center-line at a dis-
tance of 500 from the stack and at a crosswind distance of 50 m.
c Using a computer program, graph surface concentrations and locate place of
maximum concentration.

7.6 Continuous source with deposition and settling: The Ermak solution

This section follows closely [5]. When the emitted particles are more massive than air,
they tend to settle out of the atmosphere at a defined rate known as the settling velocity ,
wset [m s−1 ]. For particles than can be considered spherical of uniform radius, the settling
2
velocity can be approximated by Stoke’s law, wset = 2 cgR
9µ , where c is the particle density
[kg m−3 ] , R is the particle radius, µ is the dynamic viscosity of air [kg (ms)−1 ] and g is
the gravitational acceleration [m s−1 ].
To include the settling effect, wset is incorporated as a vertical velocity in the advection
part of Eq. 7.62 as follows,

∂c ∂c ∂2c ∂2c ∂2c

u − wset = (Kh 2 + Kh 2 + Kz 2 ) (7.78)
∂x ∂z ∂x ∂y ∂z

In addition to vertical settling, there might be partial deposition by ground absorption

or adsorption of pollutant particles takes place. This phenomenon is characterized by a
deposition velocity, wdep [m s−1 ]. To take settling and deposition into account the surface
224 Chapter 7. TRANSPORT PROCESSES AND ATMOSPHERIC TURBULENCE

boundary condition is defined as,

∂c
(K + wset c)|z=0 = wdep c|z=0 . (7.79)
∂z
Here, it is supposed that the vertical flux of contaminant particles at the surface is pro-
portional to surface concentration. Applying Laplace Transform methods Ermak found
that Gaussian plume with deposition and settling is
−wset (z−H)
q −y 2 2
c(x, y, z) = e 4r e 2σz
(7.80)
4πur

7.7 Effective stack height

In this section the ”real” height H of a stack is estimated. This height is the effective
stack height, which is the actual height which the rising plume coming out from the stack
reaches before being advected by wind action. This plume rise is because of momentum and
buoyancy acting on the out-coming effluents. As shown in Fig. 7.12, H is comprised of
h, the geometric height of the stack, and ∆h, the plume rise above h. Then,

H = h + ∆h. (7.81)

Also, the effective horizontal distance xf from the stack will be estimated. We follow
portions of Chapter 4 in [4].
If the velocity of the effluents coming out from the stack is less than 1.5 of the horizon-
tal wind velocity, there is a stack-tip down wash. Then h must be supplanted by another
stack height h0 obtained by,
vs
h0 = h + 2ds ( − 1.5) (7.82)
us
where h is in [m], ds [m] is the stack diameter, vs [m s−1 ] is the stack velocity and us
[m s−1 ] is the horizontal wind speed at height H. If this is the case, h0 must be used for
subsequent calculations.

7.7.1 Briggs calculations

The effective height is estimated for unstable, or neutral, and stable conditions. Different
set of formulas are used for these conditions as described in [6] .
 7.7 EFFECTIVE STACK HEIGHT 225

Unstable or Neutral Atmospheric Conditions

A buoyancy flux, Fb [m4 s−3 ] is determined using,

Ts − Ta
Fb = gVs d2s (7.83)
4Ts
where Ts is stack gas temperature [K], Ta is ambient temperature [K] and g is gravity
acceleration (9.8 [m2 s−1 ] ). Depending on the value of Fb , a crossover temperature (∆T )c
is calculated, and compared with the difference between stack and ambient temperature
∆T to decide wether momentum or buoyancy dominates the effluent flux. If ∆T < (∆T )c ,
momentum dominates, and if ∆T ≥ (∆T )c the flux is buoyancy dominated.
If Fb < 55 [m4 s−3 ]
1
Vs3
(∆T )c = 0.0297 Ts 2 (7.84)
ds3
and if Fb ≥ 55 [m4 s−3 ]
2
Vs3
(∆T )c = 0.00575 Ts 1 (7.85)
ds3
If Fb < 55 [m4 s−3 ] and ∆T < (∆T )c , then

Vs
∆h = 3ds . (7.86)
u
If Fb < 55 [m4 s−3 ]and ∆T ≥ (∆T )c , then
3
F4
∆h = 21.425ds b (7.87)
u
If Fb ≥ 55 [m4 s−3 ] and ∆T < (∆T )c , then

Vs
∆h = 3ds . (7.88)
u
If Fb ≥ 55 [m4 s−3 ] and ∆T ≥ (∆T )c , then
3
F5
∆h = 38.71ds b . (7.89)
u

Stable Atmospheric Conditions

In this case (∆T )c is calculated as,

√
(∆T )c = 0.019582Ts vs s (7.90)
226 Chapter 7. TRANSPORT PROCESSES AND ATMOSPHERIC TURBULENCE

where the parameter s is calculated form,

∂θ
s = g ∂z (7.91)
Ta

For stability class E, the potential temperature gradient ∂θ

∂z is taken equal to 0.02 km−1 ,
and for class F, ∂θ
∂z is equal to 0.035 km−1 .
If ∆T < (∆T )c the flux is momentum dominated and from the following the lowest
value for ∆h is chosen,
Vs
∆h = 3ds (7.92)
u
or
Fm 1
∆h = 1.5( √ ) 3 (7.93)
u s
where the momentum flux Fm [m4 s−3 ] is calculated using

Ta
Fm = Vs2 d2s (7.94)
4Ts
If ∆T ≥ (∆T )c the flux is buoyancy dominated and,

Fb 1
∆h = 2.6( )3 (7.95)
us
The procedure is shown in a condensed form in Fig. 7.13.

Figure 7.13. Briggs procedure to estimate effective height. From [4].

 7.7 EFFECTIVE STACK HEIGHT 227

Horizontal distance of effective height

To obtain the horizontal distance from the stack where the plume reaches its effective
height, xf [m], the following rules are followed:
For unstable or neutral condition,
If Fb < 55,
5
xf = 49Fb8 (7.96)

If Fb ≥ 55,
2
xf = 119Fb5 (7.97)

For stable atmosphere,

us
xf = 2.0715 √ (7.98)
s

7.7.2 Other effective height formulas

More equations to predict plume rise are found in [7].

Carson and Moses scheme

Based on 711 observations, regardless of stability conditions, Carson and Moses, [8] pro-
posed the following formula

1/2
Vs ds Q
∆h = −0.029ds + 2.62( h ) (7.99)
us us
where ∆h is in [m] , Vs is the gas exit velocity in [m s−1 ], ds is the stack diameter in [m],
us is the wind velocity at stack exit in [m s−1 ] and Qh is the heat emission rate in [kJ s−1 ].
To obtain Qh ,

Qh = ṁcp (Ts − Ta ) (7.100)

where ṁ is the stack gas mass flow rate in [kg s−1 ], i.e. , ṁ = πd2s Vs P4RT
Mw
s
, cp is the
constant pressure specific heat of stack gas, Ts is the stack gas temperature in [K], Mw is
the molecular weight of stack gas.
228 Chapter 7. TRANSPORT PROCESSES AND ATMOSPHERIC TURBULENCE

Holland formula
Thomas, Carpenter and Colbaugh employed and compared 10 different formulas to esti-
mate ∆h, concluding that the Holland formula showed fairly good agreement with mea-
surements in electric generations stations,

Vs ds (Ts − Ta )
∆h = (1.5 + 2.68 × 10−3 P d( )) (7.101)
us Ts
where the symbols and units are as listed above, with P expressed in milibars. There is
a slight tendency to underestimate the plume height. The last term may be replaced by
Qh
0.0096 Vs d .

Concawe formulas
On the basis of the analysis of Moses and Kraimer on 17 equations and 615 observations
involving 26 different stacks, the following Concawe formulas showed good agreement
with observations,

1/2
Qh
∆h = 2.71( 3/4
) (7.102)
us
or the optimized form,

Q0.444
h
∆h = 4.71( ) (7.103)
u0.694
s
where the symbols and units are as listed above.
 7.7 EFFECTIVE STACK HEIGHT 229

Example 7.2
Example: Obtaining the effective height H and horizontal distance xf .

The wind speed at the top of a stack is 3 m s−1 and the effluent gas velocity is 6 m
s−1 with temperature of 440 K. The stack diameter is 2 m, and geometrical height is
40 m. The atmospheric stability conditions are neutral, with a temperature of 300 K.
The effluent gas molecular mass is 64 g mol−1 (SO2 ).
a Estimate the effective horizontal distance and height using Briggs formulas. Find
concentration 200 m downwind.
b Estimate the effective height using optimized Concawe formula. Find concentra-
tion 200 m downwind.
c A similar stack operating under same conditions is placed 50 m on y coordinate.
Both plumes are parallel and have same downwind direction. Find concentration
200 m downwind aligned with either stack.
Hint: Use superposition principle.
230 Chapter Bibliography

Chapter Bibliography

[1] Alexandre J. Chorin and Jerrold E. Marsden. A Mathematical Introduction to Fluid Mechanics,
Third Edition. Springer, Texts in Applied Mathematics. ISBN 978-1-4612-6934-2.
[2] Roth, M. Turbulent transfer chracteristics over suburban surface . Ph.D. Thesis. University of
British Columbia, Vancouver, Canada.
[3] Henk Tennekes A First Course in Turbulence, 1st edition. The MIT Press, 1972. ISBN 978-
0262200196.
[4] Kenneth Wark, Cecil Warner, Wayne Davis Air Pollution: Its Origin and Control, 3rd Edition.
Addison-Wesley, (1997). ISBN-13: 978-0673994165.
[5] Stockie, John M., The Mathematics of Atmospheric Dispersion Modeling, SIAM Review, 53,
2, pp. 349-372, (2011). doi 10.1137/10080991X.
[6] G. A. Briggs, Plume Rise Predictions, In Lectures on Air Pollution and Environmental Impact
Analyis. American Meteorology Society, Boston, Massachussetts, 1975.
[7] User’s Guide for the Industrial Source Complex (ISC3) Dispersion Modls, Vol II-Desciption of Model
Algorithms. EPA-454/B-95-003b, USEPA OAQPS. Research Traingle Park, NC, September
(1995).
[8] J.E. Carson and H. Moses. The validity of Several Plume Rise Formulas, J. Air Pollu. Control
Assoc. 26, 11, (1976).
Appendix A

Dimensional analysis

Examples are used to illustrate the dimensional analysis method, amply used in fluid
dynamics and atmospheric flows such as the Monin-Obukhov similarity equations. It is
based on the Buckingham Pi Theorem and is an optimal and systematic approach to
determining governing dependent variables in a physical problem. It is based in consid-
ering only the fundamental dimensions such as mass, length or time of the governing
variables. To obtain these, it requires knowledge of the physical problem and intuition.

A.0.1 The dimensional matrix

Consider a fluid problem involving the variables velocity V, length L, force F, density ρ,
dynamic molecular viscosity µ and gravity acceleration g. These variables are expressed
in terms of its fundamental dimensions: M for mass; L for length and T for time. A
convenient way the express this is using the dimensional matrix:

Table A.1. Dimensional matrix: Columns contain the values of the fundamental
variables exponents for each of the corresponding dependent variables

V L F ρ µ g
M 0 0 1 1 1 0
L 1 1 1 -3 -1 1
T -1 0 -2 0 -1 -2

By inspecting the entries of the dimensional matrix for the force F column, its di-
mensions can be obtained: M1 L1 T−2 . Dimensions for the other variables are obtained
analogously.
Any product Π of variables V, L, F, ρ, µ and g has the form

Π = V k1 Lk2 F k3 ρk4 µk5 g k6 . (A.1)

– 231 –
232 Appendix A. DIMENSIONAL ANALYSIS

The dimension of Π is given by

[Π] = [M 0 L1 T −1 ]k1 [M 1 L1 T −2 ]k2 [M 0 L1 T −1 ]k3 [M 0 L1 T −1 ]k4 [M 0 L1 T −1 ]k5 [M 0 L1 T −1 ]k6

= [M k3 +k4 +k5 ][Lk1 +k2 +k3 −3k4 −k5 +k6 ][T −k1 −2k3 −k5 −2k6 ] . (A.2)

Since the Π groups must be dimensionless, we must demand that

k3 + k4 + k5 = 0, k1 + k2 + k3 − 3k4 − k5 + k6 = 0, −k1 − 2k3 − k5 − 2k6 = 0 .

Note that the coefficients for the ki ’s, including zeroes, in each equation are the row entries
of the dimensional matrix. This set of three equations and six unknowns is undetermined.
It is necessary to assign values to three of the ki ’s and solve for the remaining ki ’s to obtain
a solution. Suppose the set (k1 , k2 , k3 ) is known.
Appendix B

State variables, temperature,

heat, sensible, and latent heat
flux

The surface upon which solar radiation is reflected, is complex and varied since it’s shaped
by vegetation, animal and human life. Deserts, lakes, jungles, forests, tundras and oceans,
compose Earth’s surface.Furthermore, these surfaces change with meteorological condi-
tions such as rain, or snow, and year’s stations. Construction of urban centers, dams,
irrigation systems and agriculture, complicate this picture even further since each type
of surface has its own thermodynamic properties, having different radiative responses.
Thermodynamic principles provide necessary knowledge to understand, and predict the
behavior of such complicated systems. Specific applications of these principles are applied
to bare, vegetation and urban surfaces.
State variables. In thermodynamics the macroscopic state of a system at equilibrium
is determined by a small set of state variables. For example, if we know the values
of independent state variables such as temperature T , volume V , and pressure P of
a sample, all other thermodynamic properties such as work, and internal energy of
the fluid are determined. Two remarkable characteristics of state variables like P , V
or T , are that they do not depend on the history of the system, and that they follow
the rules of calculus. All that is needed is their present values, regardless of how they
got there, and are subject of integration or differentiation.
Temperature. As stated in Chapter x, the atmosphere can be considered as an ideal
gas. As such, it follows the Universal Law of Gases

P V = nRT, (B.1)

– 233 –
234 Appendix B. STATE VARIABLES, TEMPERATURE, HEAT, SENSIBLE, AND LATENT HEAT FLUX

or
P V̄ = RT, (B.2)
V
where V̄ = n, n is the gas number of moles, and R is the ideal gases constant. Using
above identities, the ideal gas temperature, T , is defined as

P V̄
T = . (B.3)
R
This definition gives a macroscopic definition of temperature. All we need to know
T , is the pressure P of the fluid and the volume V it occupies.
An important fact of Eq. B.3 is that the limit

P V̄
T = lim , (B.4)
P →0 R

for all gases tend linearly to the same point, as shown in Fig. . This provides a common
origin, and therefore the possibility of establishing a temperature scale, known as
the absolute temperature scale or Kelvin scale. Its units are Kelvins, K. The origin
corresponds to -273.15 ◦ C. The units of T depend on the units chosen for the ideal gas
constant R. In the international system R =. In chemistry, is convenient to use R =.
In both cases T is in Kelvins, K.
In the microscopic realm, the Kinetic Model of Gases provides a description of tem-
perature, and other state variables. This model is based on the following assumptions:
1. A gas consists of a collection of molecules in continuous random motion.
2. Gas molecules are infinitesimally small points.
3. Molecules move in straight lines until they collide.
4. Molecules do not influence one another except during collisions.
Note that no attractive or repulsive forces between ideal gas molecules are considered
here. Considering only these assumptions, and the momentum definition provided
by Newton’s second law of motion, is possible to arrive to

1 2
P V = nM vrms , (B.5)
3
where n is the amount in moles of gas molecules, M is the corresponding molar mass,
2
and vrms is the root mean square speed of the molecules, i.e.,
!2
2 v12 + v22 + v32 + ... + vN
2
vrms = ,
N

where N are the number of molecules in the sample, and their speeds are v1 , v2 , v3 , ..., vN .
2
The appearance of vrms in Eq. B.5 is because pressure P can be obtained by the
 235

force per area exerted by the momentum carried by molecules velocities. Compar-
ing Eqs. B.1, and B.5 we get,
1 2
nRT = nM vrms . (B.6)
3
This identity relates the macro and molecular approach. From Eq. B.6,

2
1 V M vrms
T = . (B.7)
3 R
This equation states that temperature T is a function of the square of molecules ve-
locity. This means that T is a measure of the molecular kinetic energy.
For the same token,
−2
3RT

vrms = . (B.8)
M
This results allows us to calculate vrms of any gas at any temperature by substituting
corresponding M and T , as shown in Fig p147 atkins.
Enthalpy change, or sensible heat flux. The pressure, volume or temperature of a
system such as an air parcel, can be changed by exerting or extracting work, w, and/or
by adding or removing heat, q. For example, in the Joule experiment work can be
added to a fluid by introducing a rotating paddle wheel. This results in temperature
rise. On the other hand, heat can be added or removed by placing a heat reservoir
(a large warmer or colder body) in contact with the system. A measurable property
change of the system resulting from these experiments is called internal energy, U ,
as expressed by,

∆U = q + w. (B.9)

This equation is an important result of the First Law of Thermodynamics. U has

units of work or heat, J.
Consider the case of an experiment taking place in the atmosphere at constant pres-
sure, i.e, an isobaric process. Under this conditions, Eq. B.9 becomes,

∆U = qp − P ∆V, (B.10)

where qp is heat supplied or retrieved at constant pressure, and P ∆V is the work

performed by volume change. The minus sign means that internal energy U is reduced
by expansion, and increased by contraction of the air parcel. Equation B.10 can be
written as,
U2 − U1 = qp − P (V2 − V1 ), (B.11)
236 Appendix B. STATE VARIABLES, TEMPERATURE, HEAT, SENSIBLE, AND LATENT HEAT FLUX

or,
qp = (U2 + P V2 ) − (U1 + P V1 ), (B.12)

where the subscripts 1 and 2 indicate initial and final state of the system, respectively.
The two parenthesis in the right hand side contain an ubiquitous property used in
thermodynamics, and thermochemistry: enthalpy, H. Since for the definition of H
only state variables U , P , and V intervene, H is itself a state variable. From Eq. B.12,

qp = H2 − H1 = ∆H. (B.13)

Equation B.13 means that, at constant pressure, removal or addition of heat is syn-
onymous to enthalpy change in a parcel. Therefore, for an infinitesimal change at
constant pressure
δqp = dH. (B.14)

Note that the differential of a property such as qp , and a state variable such as H is
distinguished by using symbol δ and d, respectively. This is because a state variable
such as H can be differentiated following the Laws of Calculus, while heat, and in
particular qp , not.Explain..
Coming back to Eq. B.14, since at constant pressure H is only a function of temperature
T the chain rule provides,
dH
 
dH = dT. (B.15)
dT P
Therefore,

dH
 
δqp = dT. (B.16)
dT P
 
dH
The quantity dT P is called heat capacity at constant pressure, cp . Then,

dH = cp dT. (B.17)

The heat capacity cp can be determined experimentally for many substances, including
air. Using a constant pressure calorimeter, its nominal value at 300 K is determined
at 1.01 kJ kg−1 K−1 . This means that, when one kg of air is at 300 K, one kJ is needed
to increase its temperature by one degree Kelvin.
When an air parcel is transported, and comes in contact with a warmer or colder
surface, energy is interchanged. Therefore, changes in temperature occur. The energy
exchanged between the two bodies at different temperatures, is equal to the change
in enthalpy of the air parcel. In the context of meteorology, and atmospheric sciences,
enthalpy change dH is called sensible heat. In the atmosphere sensible heat is thus
transported by the temperature of air parcels.
 237

Latent heat. In a humid air parcel, under conditions to be studied below, water may
change of phase: liquid to gas, evaporation; liquid to solid, freezing; solid to liquid,
melting; and gas to liquid, condensation. Direct change form solid to gas is called
sublimation. In all these cases the energy of the surrounding air is changed, not
because of contact between two bodies at different temperature as was the case of
sensible heat, but at the expense of energy used to modify the molecular arrangement
of water while changing phase. Consider separately these cases:
– Evaporation. In liquid water, molecules are loosely in contact with each other as
a result of interacting electrostatic forces. Molecules travel in the liquid by forcing
their way through intermolecular passages in a similar way as a passengers in the
subway move to reach the door. To change the phase of a liquid to gas, enough
energy must be provided for the molecules to overcome liquid’s intermolecular
electrostatic forces, thus expanding intramolecular distances. To accomplish this,
energy must flow from the environment to the fluid. When a human goes out
from a swimming pool, water on the skin evaporates by soaking energy from the
human body overpowering intermolecular attraction separating the molecules,
thus forming vapor. The resulting lower energy in the human body translates to
reduced molecular kinetic energy, and thus lower body temperature. While the
body gets colder, by conservation of energy, the surrounding atmosphere gets
warmer.
The energy per mass needed to accomplish evaporation is called the specific en-
thalpy of evaporation, or specific latent heat of evaporation, LE . Representative
values for water are,

Table B.1. Condensation/Evaporation Specific Latent Heat for Water

T (◦ C) L(J kg−1 × 106 )

-40 2.60
0 2.50
40 2.40
100 2.26

For atmospheric applications LE ≈ 2.5Jkg −1 × 106 is generally assumed.

– Condensation. As energy is retrieved from vapor, intramolecular distances shorten,
so intermolecular forces assume its bonding role. This is the exact reverse case
of evaporation. In this case, heat flows from the fluid to the environment. Same
enthalpy values shown in Table B.1 are used, but with reversed sign.
238 Appendix B. STATE VARIABLES, TEMPERATURE, HEAT, SENSIBLE, AND LATENT HEAT FLUX

– Freezing. Water in liquid state, with a loose arrangement of molecules as de-

scribed above, becomes an ordered crystal-like lattice. Molecules wiggle around a
fixed position. Intramolecular distances become closer, and intermolecular forces
become so strong that water becomes solid. To get to this state, the necessary en-
thalpy to extract from the fluid is given by the latent heat of fusion or freezing,
Lf . AT 0 ◦ C, Lf ≈ 3.3 × 105 J kg−1 .
– Melting. If enough enthalpy is provided to water in solid state, the rigid lattice
gives way to a loose molecular arrangement. This is the reverse process as in
freezing, and same values for Lf are considered, but with reversed sign.
– Sublimation. In this case, solid state goes directly to vapor. To obtain the neces-
sary enthalpy to accomplish this, Ls , must be the sum of enthalpies that go from
solid to liquid, and from liquid to gas,

Ls = Lf + L.

In the atmosphere, latent heat is transported by the water vapor content in an air
parcel. Latent heat of a parcel in a dry environment, such as in a dessert, will be
considered negligible, but it is of great importance in a humid environment such as
on the ocean surface, a forest, or a jungle.
Latent, and sensible heat fluxes. For atmospheric applications the above latent, and
sensible heat definitions are translated to heat fluxes by obtaining the amount of
energy flow normal to a surface in a given amount of time. Therefore, latent heat
flux,HL , and sensible heat flux, HS , in Js−1 m−2 , or Wm−2 will be considered herein.